From mccay at redestb.es Sun Nov 1 14:39:52 1998 From: mccay at redestb.es (Alan R. King) Date: Sun, 1 Nov 1998 09:39:52 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- Yet another instance of s > r occurs dialectally in Galician (also Galego, Gallego; Romance, nearest relation Portuguese). It occurs with phonological regularity as a (substandard) geographically localized dialect feature (for isogloss details see Fernandez Rey), and is mentioned as such in several modern manuals of Galician linguistics, according to which: etymological (and standard Galician) /s/ (generally alveolar in Galician, with regional variants) is realized in the varieties in question as: [r], more exactly a "voiced simple alveolar liquid vibrant" (see Freixeiro Mato), in the following context: "in word internal position, or as a sandhi phenomenon, PRECEDING A VOICED CONSONANT, and to a lesser extent (Fernandez Rey: "exceptionally"), before the voiceless consonants /f/ and /th/ ["voiceless interdental fricative"], and even before /ch/ ["voiceless prepalatal affricate"]" (translated-paraphrased from Freixeiro Mato). Examples from Freixeiro and from Fernandez Rey [format: "standard spelling" + /broad transcription of realization, reinterpreted by me avoiding phonetic symbols not available/ + 'meaning']: "desde" /derde/ 'since, from' "as vacas" /arbakas/ 'the cows' "lesma" /lerma/ 'kind of mollusk' "as mans" /armans/ 'the hands' "escindir" /erthindir/ 'separate, split' "os zapatos" /orthapatos/ 'the shoes' "as flores" /arflores/ 'the flowers' "todos xuntos" /todor shuntos/ 'all together' Just for the record: Both sources also mention another, more "sporadic" tendency in some dialects of Galician to aspirate /s/ when "implosive" (i.e. when followed by a consonant), e.g. "desde" /dehde/ 'since, from' "disgusto" /dihhusto/ 'displeasure' (in many parts of Galicia, /g/ in most positions is regularly realised as [h]) "espantallo" /ehpantalyo/ 'scarecrow' "a escoba" /ahkoba/ 'the broom' Finally, I should no doubt mention that in Galician, /s/ is normally voiceless, but is regularly voiced when it precedes a voiced consonant, so in standard pronunciation "desde" is [dezde], "as vacas" is [azbakas], etc. Sources: Fernández Rei, Francisco: Dialectoloxía da lingua galega [second edition]. Vigo: Xerais, 1991. (see page 57) Freixeiro Mato, Xosé Ramón: Gramática da lingua galega. I: Fonética e fonoloxía. Vigo; A Nosa Terra. (see page 161) Alan Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From manaster at umich.edu Sun Nov 1 14:40:29 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Sun, 1 Nov 1998 09:40:29 EST Subject: h- in Turkic In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Sasha raises important points to which I respond briefly below. I do not strongly disagree with anything he says, but I do feel that the subject should not be closed. On Sat, 31 Oct 1998, Alexander Vovin wrote: > ----------------------------Original message---------------------------- > Alexis, > > I think your message regarding initial h- raises one important > methodological issue. Namely, can we allow a reconstruction of a segment > for a proto-language that is preserved in a single language of otherwise > big language family, and for which there is no second independent > evidence? It seems that you would answer that question in the affirmative > in this particular case, My answer is: It depends. Some syllable-final consonants in Uto-Aztecan survive only in Tubatulabal. Some IE laryngeals survive only in Hittite. > although I remember that once you yourself were > bashing (quite justifiable, in my opinion, a person X from Moscow > Nostratic school for search of IE and Nostratic accent distinctions > uniquely preserved in Bengali). But that is not the only or main reason I was bashing them! > I would hate to disagree with you on > Khaladj h-, but I think I have to. But I have not taken a strong enough position on Khalaj h- for anyone to be able to disagree with. > I would answer in the negative to the > question I posed above, although I think that some exceptions could be > allowed when a language that unikely preserves segment X, is on the top of > the branching. In all other cases it is much safer to reconstruct > something, especially something radical, like PT *h- on the basis of two > independent pieces of evidence. Khaladj is probably *close* to the root of > Turkic tree, but it does not represent primary branching, I think. It will > be dangerous enough to reconstruct PT *h- on its sole evidence (although I > think that this might eventually turn out to be true -- let us see), but > looking for the traces of something Nostratic in khaladj *only*, does not > seem to be very realistic. I did not say that this IS so, only that it is a hypothesis I would like to investigate. You should anyway be happy because you are the one who made me see that it could not just be *p-. But anyway it may simply be that Khalaj, in addition to what I call Ataturkish (haha), is the only Turkic language I am fairly comfortable with, but it sure FEELS to me like the h- is a real well-established feature here, and I would hate to have to assume that it is purely one of those mysterious that so many language make ex nihilo. Anyway, you are right that all by itself the whole thing would be somewhat improbable, but I do have some other ideas re Altaic *w-. Moreover, while no living Turkic language has anything that looks like a regular reflex of Altaic *p- or *w-, I am not yet ready to dismiss some of Doerfer's claims for older stages of the Turkic. In sum, if I choose to waste my time pursuing the possibility that the Khalaj h- is a reflex of something old, that should not worry you too much(:-). Of course, I have cagily enough NOT been doing so. Rather I keep trying to get someone else to waster THEIR time checking this out(:-). Alexis From bwald at HUMnet.UCLA.EDU Sun Nov 1 14:42:38 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sun, 1 Nov 1998 09:42:38 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Jim Rader's message on "rolled" /r/ (spelled "rr" intervocalically in Spanish orthography) in Latin American dialectology prodded my memory on some Bolivian Spanish I've heard where intervocalic "rr" indeed sounded a lot like "z" to me. In many areas of Latin America either intervocalic rolled (NOT tapped) or final "r" has a varying amount of local friction accompanying the liquid articulation producing a sibilant-like sound. When I first heard this Bolivian z-like intervocalic "rr", it reminded me of the Mexico City final "r", which also (most) often has a lot of local friction, but more "palatal"-sounding, like "zh" (even often devoiced finally to a "sh"-like sound). My general thought was that since Spanish historically lost voiced lingual fricatives (by devoicing and further shifts, no "z", "zh"), the liquid articulation was encroaching on phonetic voiced lingual fricative territory with impunity (since there were no previous voiced fricatives for it to threaten with merger). The interesting thing to me was that there doesn't seem to be any extra-segmental conditioning involved. That is, where palatalised "r" > "zh/sh" (e.g., in Polish or Czech) is heavily contextually conditioned (by an immediately preceding stop and/or following a historically palatal vowel), the development of local friction with rolled "r" among Latin American Spanish dialects does not show such coarticulation motivation, and, thus, seems to be part of a paradigmatic ("unconditioned") shift among consonants. In contrast, the change exemplified in (Parisian) French "chair(e)" to "chaise" does cause merger with previous -z-, but I am unclear on how regular this change actually was, and to what extent sociolinguistic affectation distorted its progress toward a bona fide neogrammarian sound change. Conclusion: r > z may be possible as an unconditioned change depending on aspects of the larger consonantal system in which the original "r" was positioned. Degree of aperture may vary for voiced continuants (phonetic fricatives or liquids) when it does not support a phonemic contrast. (z > r may be less dependent on position in a consonantal system, contributing to Campbell's impression that it is a more "natural" change. This relates to the general issue of directionality of degree of aperture in consonantal sound changes, "weakening", i.e., "opening", apparently being more commonly observed than "strengthening", i.e., "closing", but with enormous qualifications. The qualifications may relate more to stop/(af)fricative changes than to changes within continuants.) In any case, such considerations of consonantal system type (e.g., what are the phonetically contrastive dimensions?) have obvious implications for reconstruction and plausibility of reconstructed systems. (Oblique comment: the Chagga dialects of Northeast Bantu, Kilimanjaro area, have had varied unconditioned outcomes of *t, among them /d/, /r/, /zh/ and /R/ (i.e., velar /r/). The /zh/ variant, which varies with a palatalised /r/ in some dialects, e.g., Moshi, Vunjo, seems to show the closing process from an earlier and more widespread /r/ reflex of *t, where *t > r surrounds the /zh/ reflex,, and stems from the post-alveolar articulation of the original *t -- as preserved in most Bantu languages. The *t > r process also occurs in various other East Bantu areas as part of a lenition process that variously affects the voiceless stops *p, *t or *k -- but most often *p. In Chagga, the /R/, Machame dialect, is also clearly a further evolved form of /r/, and in turn evolves into /h/, cf. Brazilian Portuguese, in the Siha dialect. The *t > d change is most eastern, e.g., Rombo dialect, and not clearly related to the *t > r change, but is shared with the distinct language to its east, Dabida (NB often called Taita in the literature). The unconditioned change of a post-alveolar into a palatal /zh/ rather than a /z/ is paralleled independently by the Bajuni change of *t > ch by a variety of well-motivated steps which reflect the same progression as Chagga *t > zh, but without loss of the voiceless closure component. *t > r starts out as a tap, but then gets rolled and eventually, in Chagga, develops local friction on its way to zh. The central dialects demonstrate the progression. Most likely the r > zh change has nothing to do with the origin of r in *t, and is unusual, but in a less direct way the overall directions of the consonantal shifts in East Bantu may be involved, making r > zh more likely in this larger context than in general -- but cf. Mexico City final r > "zh") P.S. Trask's admonition on probability of chance resemblances reminds me of the classic probability problem: how many people do you need in a room before there is a more than chance (p > .5) probability that TWO will have the same birthday? I forgot the answer, something around 30 (cntr. 366 possible birthdays). The answer says NOTHING about which date this will be. The probability for that remains 1 in 365 and a 1/4 (p < .003). Koestler's fallacy is alive and well in popular American culture in the "strange-but-true" folklore about how many "famous" Americans were either born or died on July 4. There are many other variants of such folklore, e.g., Cabalistic-like algorithms about probability of deaths in office for US presidents and such, depending on year, their ordinal rank as presidents, the number of letters in their names, etc etc. I guess we all know some of the "other" great people who were born the same day as us. Conclusion: most people hate mathematics but love numbers. From mcv at wxs.nl Mon Nov 2 01:40:04 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Sun, 1 Nov 1998 20:40:04 EST Subject: r and s: Galician In-Reply-To: <3.0.3.32.19981031235447.006dbc18@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Yet another instance of s > r occurs dialectally in Galician (also Galego, >Gallego; Romance, nearest relation Portuguese). And in dialectal Mallorqui' too. Same rules as in Galego and Sardo (es bisbe "the bishop" > er birbe). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From dyen at hawaii.edu Mon Nov 2 21:57:12 1998 From: dyen at hawaii.edu (Isidore Dyen) Date: Mon, 2 Nov 1998 16:57:12 EST Subject: Cladistic language concepts In-Reply-To: Message-ID: ----------------------------Original message---------------------------- The social conditions are temporary because they are replaced by other conditions. The effects on the language can be either temporary or permanent. The effects are often permanent. As for the spread of socially dominant languages, like it or not, that is going on at a great pace and there is little that one can see in local resistance that is a real obstacle to it. On Thu, 15 Oct 1998, bwald wrote: > ----------------------------Original message---------------------------- > Roger Wright quotes from my last message. > > >Benji Wald says: > > > >>On the contrary, up > >>to the present, history indicates the continual long-term fragmentation of > >>languages and destruction of mutual intelligibility. > > >Yes; nicely put. "Up to the present" -- > >There's a good case for saying this may never happen again. > > I'm glad that Roger appreciated the qualification I put on what I was > saying. I deleted a further paragraph on how I invite such views as he is > proposing about a sharp discontinuity between the past and the future, > since I see little reason to suppose that such a discontinuity has come > about in the 20th c. or will in the foreseeable future, despite impressive > advances in communication technology (at least impressive to us current > beings) and increasing sharing of various kinds of literacies. Meanwhile, > the same old problems of miscommunication that have always existed (and > have occasionally been reported in the past) persist. (Would it be > surprising if "human nature" exists not only in our linguistic devices but > in how we use them?) > > I think the main thing that would interest me among Roger's proposals is > what changes in the nature of human society (as a whole or in its various > various parts) he envisages or suggests to have relatively recently arrived > which will override the steady and unrelenting effects of social localism > that in the past (presumably) have been the major cause of linguistic > fragmentation and loss of mutual intelligibility. I am skeptical, and > suspect he is underrating the long-term cumulative effects of localism, and > overrating the stability of centralised power, but I am open to hearing > interesting proposals about the "changing" relation of social change to > linguistic change. > > With regard to Isidore's latest message, I share his appreciation for > Hubey's comments, but paused at the following passage: > > >The point is, as I see it, that linguistic change > is built into the way the community interacts with its language, whereas > some aspects of linguistic change are conditioned by the social changes > that are going on in the community. The latter type of change, since it is > local and temporary I thought could be excluded from being regarded as > a 'main factor', but I suppose it gets to be a matter of defintion. > > I'm not sure I understand the intent of "local and temporary". The > "temporary" part seems to suggest that the local changes are eventually > undone, as if afterwards they seem to have never occurred (i.e., no > *lasting* harm done to mutual intelligibility). If that is not what is > meant, then they have had their effect in changing the local language AWAY > from other local varieties. This seems more than a matter of definition > (of "main factor"?) to me, but of the cumulative consequences of local and > temporary (temporally bounded?) changes. > > Of course, most changes do spread beyond the temporary local interest group > that intiates them, so their effects don't go away but continue, and > establish the typical mosaic patterns we commonly find in dialect > geography. I cannot address all fronts at once, so when I emphasised > fragmentation and loss of mutual intelligibility as the traditional and > still usual focus of historical linguistics (at least on the elementary > level), I did not complicate that traditional picture with diffusion and > convergence which takes place across languages as well as in them, due to > bilingualism, bidialectalism, register/stylistic complexity, or whatever > level of analysis is appropriate at the time. I am continually struck by > cases in which it is not clear which family or (even more commonly *in the > literature*) branch of a family some language (group) or other belongs to > because of convergence and sharing of features. Such things most > strikingly occur at the margin of isogloss bundles. Isidore (and no doubt > Roger) is certainly right that channels and interests promoting > communication ("mutual" intelligibility) *across* local groups is also a > factor in change. > > The issue remains: will languages continue to fragment and produce mutual > unintelligbility, or will the whole world eventually abandon its local > variety in favor of some "homogeneous commercial or totalitarian English" > or whatever (no doubt incomprehensible to us current beings). I think not. > From nbvint at nessie.mcc.ac.uk Mon Nov 2 17:23:32 1998 From: nbvint at nessie.mcc.ac.uk (nigel vincent) Date: Mon, 2 Nov 1998 12:23:32 EST Subject: s > r in Sardinian Message-ID: ----------------------------Original message---------------------------- Presumably the development of s > r in Sardinian is a rather different phenomenon from that in Latin which started this discussion off since the Latin environment is intervocalic while the Sardinian one is preconsonantal. Moreover, Michael Jones notes that in Logudorese-Nuorese there is an interesting complementarity, with r > s before voiceless consonants (thus 'batos kanes' '4 dogs' beside 'bator gatos' '4 cats' and s > r before voiced consonants (see his chapter on Sardinian in M. Harris & N. Vincent (eds) The Romance Languages, 1988, p. 323). For a lot more (fascinating) detail, see Michel Contini 'Etude de geographie phonetique et de phonetique instrumentale du sarde' Alessandria, Edizioni dell'Orso, 2 vols, 1987. [NB on a terminological note people in the Anglophone world these days usually call the language 'Sardinian' not 'Sard'. Posner's usage here is decidedly 'arcaizzante'.] Nigel Vincent Tel: +44-(0)161-275 3194 Department of Linguistics Fax: +44-(0)161-275 3187 University of Manchester e-mail: nigel.vincent at man.ac.uk Manchester M13 9PL http://lings.ln.man.ac.uk/Html/NBV/ UK Visit our web-page: http://lings.ln.man.ac.uk/ From Roger.Wright at liverpool.ac.uk Mon Nov 2 16:26:53 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Mon, 2 Nov 1998 11:26:53 EST Subject: s > r In-Reply-To: <19981029.173548.12590.0.TonyBreed@juno.com> Message-ID: ----------------------------Original message---------------------------- And in Spanish, including the most pompous and high-class registers in which dropping final /s/ would in general be seen as unacceptable, a word-final -/s/ before word-initial rolled /r/- just isn't pronounced at all: "los reyes" as [lor'ejes]. (This is also normal in the rare cases of what looks from the spelling like word-internal /sr/; thus "Israel" as [ira'el]). In the light of this continuing discussion, I'm now wondering if this /s/ has assimilated rather than dropping; before any other voiced consonant it would indeed assimilate, by voicing, "los malos" as [lozm'alos] (but [lozr'ejes] with [z] isn't normal). RW >I don't know if this transformation exists throughout Sard or in limited >areas. In that section, Posner talks about the general weakness of /s/, >transforming, variously, to /j/, /h/, /S/, and /x/ via /S/, as well as >dropping completely in certain cases (in French), in addition to the Sard >example. From lsa at lsadc.org Mon Nov 2 16:18:47 1998 From: lsa at lsadc.org (LSA) Date: Mon, 2 Nov 1998 11:18:47 EST Subject: October Bulletin Message-ID: ----------------------------Original message---------------------------- The October 1998 LSA Bulletin is now available on the LSA web site: www.lsadc.org From mccay at redestb.es Tue Nov 3 13:20:12 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 3 Nov 1998 08:20:12 EST Subject: s > r (Spanish) Message-ID: ----------------------------Original message---------------------------- Roger Wright wrote: >----------------------------Original message---------------------------- > > >And in Spanish, including the most pompous and high-class registers in >which dropping final /s/ would in general be seen as unacceptable, a >word-final -/s/ before word-initial rolled /r/- just isn't pronounced at >all: "los reyes" as [lor'ejes]. (This is also normal in the rare cases >of what looks from the spelling like word-internal /sr/; thus "Israel" >as [ira'el]). In the light of this continuing discussion, I'm now >wondering if this /s/ has assimilated rather than dropping; before >any other voiced consonant it would indeed assimilate, by voicing, "los >malos" as [lozm'alos] (but [lozr'ejes] with [z] isn't normal). > RW I don't have any systematic argument to back this up, but as a "second-language" Spanish speaker of over twenty years' standing (and a linguist, though not a phonetician, residing in a largely Castilian-dominated country), it had never occurred to me before to "hear" these pronunciations as anything other than assimilations or /s/ to the following /r/. And while I only speak from intuition and impressions, I am almost certain that the pronunciation of such /sr/ phoneme sequences as those mentioned by Roger Wright is something different from (and phonetically more complex than) a single trilled /r/; the phonetic representations [lor'ejes], [ira'el] etc. feel wrong (to me): something's missing in them. Or is this merely because I know how they are spelt?? Would I be saying something totally absurd if I suggested that the assimilated /s/ in these contexts is more like a "rolled [z]" than a normal Spanish /r/? Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From cwilhelm at ucla.edu Tue Nov 3 13:20:29 1998 From: cwilhelm at ucla.edu (Chris Wilhelm) Date: Tue, 3 Nov 1998 08:20:29 EST Subject: XI UCLA Indo-European Conference Message-ID: ----------------------------Original message---------------------------- I would like to announce the Eleventh UCLA Indo-Eurpoean Conference, to be held 4-5 June, 1999, on the UCLA Campus. The Abstract deadline is 1 March, 1999. Contact: Christopher Wilhelm, Coordinator 100 Dodd Hall, UCLA 405 Hilgard Ave. Los Angeles, CA 90095 Tel: (310) 825-4171 (daytimes) (310) 473-4223 (eves. & weekeends) Fax: (310) 206-1903 e-mail: iesa at ucla.edu cwilhelm at ucla.edu From larryt at cogs.susx.ac.uk Tue Nov 3 20:38:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 3 Nov 1998 15:38:09 EST Subject: Q: non-IE terms Message-ID: ----------------------------Original message---------------------------- As many of you know, I'm compiling a dictionary of historical and comparative linguistics, which I have to submit in a couple of months. At present, I have pretty good coverage of general terms from HL, comparative linguistics, philology and language change. And I also have coverage of IE terms which is about as nearly exhaustive as I think I can justify. But I don't have terms from language families other than IE. I've been avoiding those, because of worries about space. However, I've just had a meeting with my editor, and she agrees that we should go for coverage of these as well. So I'm posting a request for assistance. What I'm looking for is named "laws" and processes from families other than IE -- Semitic, Bantu, Japanese, Algonquian or whatever. That is, I'm looking for equivalents of IE terms like `Verner's Law', `First (Slavic) Palatalization' and `ablaut'. If you can suggest any such terms, it would be very helpful if you could provide as much as possible of the following information: * the term * its domain * a definition * an example or two * a *complete* reference to the first published use of the term I'm thinking of things like `Meussen's Law' in Bantu and `rendaku' in Japanese. All assistance will be gratefully received. But please note that terms must be explicitly historical in nature: non-historical terms cannot be included. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 3 19:23:52 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 3 Nov 1998 14:23:52 EST Subject: rhotacism from Ray Hickey In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Sun, 1 Nov 1998, bwald wrote: > P.S. Trask's admonition on probability of chance resemblances reminds me > of the classic probability problem: how many people do you need in a room > before there is a more than chance > (p > .5) probability that TWO will have the same birthday? I forgot the > answer, something around 30 (cntr. 366 possible birthdays). The answer > says NOTHING about which date this will be. The probability for that > remains 1 in 365 and a 1/4 (p < .003). Koestler's fallacy is alive and > well in popular American culture in the "strange-but-true" folklore about > how many "famous" Americans were either born or died on July 4. There are > many other variants of such folklore, e.g., Cabalistic-like algorithms > about probability of deaths in office for US presidents and such, depending > on year, their ordinal rank as presidents, the number of letters in their > names, etc etc. The answer is in fact 23. Put 23 arbitrary people in a room, and the probability that two of them will share a birthday date (not an actual date of birth) exceeds 50%. Put 40 people in the room, and the probability exceeds 90%. Most people won't believe this, and you can win a few bets this way. I've done so, with first-year groups of around 45 students. I've never lost, though I must lose eventually. One year, when I announced my bet, there was widespread giggling. It turned out that, unknown to me, the class contained a pair of twins. So I magnanimously agreed to count the twins as one person, and I won anyway. As Benji points out, human beings are woefully bad at estimating probabilities. Mostly, I think, we tend to interpret `random distribution' as `disperse distribution', meaning that we tend to assume that independent events have a tendency to avoid one another. They don't, or they wouldn't be random. The linguistic consequences of this failing are all too obvious. Ancient Greek for `honey' was , and Hawaiian for `honey' is . Wow! I can hear Arthur Koestler telling us that Something Deeply Significant is going on here. But, of course, neither the Greeks nor the Hawaiians had any interest in ensuring that their words for `honey' were different, nor any means of doing so. Collect enough languages, and enough words, and you're going to be drowning in such coincidences. Standard Italian `two' is a lot more similar to Malay `two' than it is to Neapolitan Italian `two'. In fact, looking at that list of number names on the Web ( http://www.tezcat.com/~markrose/numbers.shtml ) can be quite an illuminating experience. English /tu:/ and German /tsvai/ do not look to be closely related, nor do the Kashmiri dialect variants /zi/ and /do:/, nor do the Pashto variants /bu/, /lu/ and /do:v/, nor does Armenian /erku/ look like anything else IE -- yet all are cognate. Historical linguistics 15, miscellaneous resemblances 0. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Tue Nov 3 15:37:26 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Tue, 3 Nov 1998 10:37:26 EST Subject: s > r (Spanish) In-Reply-To: <3.0.3.32.19981103011411.00729d68@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >I am >almost certain that the pronunciation of such /sr/ phoneme sequences as >those mentioned by Roger Wright is something different from (and >phonetically more complex than) a single trilled /r/; the phonetic >representations [lor'ejes], [ira'el] etc. feel wrong (to me): something's ([lo'rejes] of course) >missing in them. Or is this merely because I know how they are spelt?? > >Would I be saying something totally absurd if I suggested that the >assimilated /s/ in these contexts is more like a "rolled [z]" than a normal >Spanish /r/? I'm afraid so... Now I do agree that there may be someting more than just /rr/ in /irrael/ . I'm not absolutely sure, because my native language is Castilian with some interference from Catalan (and Dutch), and if I say /irrael/ now it's the result of a conscious effort to mend my native spelling-pronunciation ways (I suppose the Castilian spoken in the Basque country may be similarly affected). But it's my impression that the initial part of [rr] in and similar words is voiceless, making it [r.rr] or [hrr], reflecting the voicelessness of the [s] that used to be there. Spanish /s/ is very very reluctant indeed to become voiced [it rather becomes [h] or [r] than [z]], and a word like for me is always [asno], never [azno]. The same phenomenon, who am I telling, is even stronger in Basque, where (voiceless!) and assimilate forwards, ezta? (=ez da). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mott at lingua.fil.ub.es Tue Nov 3 14:29:09 1998 From: mott at lingua.fil.ub.es (Brian Mott) Date: Tue, 3 Nov 1998 09:29:09 EST Subject: r and s: Galician In-Reply-To: <363dda7a.112697997@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- The change of s to r before another consonanant is widespread in southern Spain and also in Madrid, so that los dedos may be articulated as lor dedoh and buenos dias as buenor diah. On Sun, 1 Nov 1998, Miguel Carrasquer Vidal wrote: > ----------------------------Original message---------------------------- > "Alan R. King" wrote: > > >Yet another instance of s > r occurs dialectally in Galician (also Galego, > >Gallego; Romance, nearest relation Portuguese). > > And in dialectal Mallorqui' too. Same rules as in Galego and Sardo > (es bisbe "the bishop" > er birbe). > > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam > From martinez at eucmos.sim.ucm.es Wed Nov 4 16:59:31 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 4 Nov 1998 11:59:31 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- > > The change of s to r before another consonanant is widespread in southern > Spain and also in Madrid, so that los dedos may be articulated as lor > dedoh and buenos dias as buenor diah. it is nice to read it, because I never heard a change like "buenor diah" in Madrid, nor in Toledo, Galicia nor in Southern Spain. You say "widespread", but I never, *really never*, met it. You can hear following (sub)standard variants: in Madrid 1) buenos dias 2) bueno dia in the South (from Toledo toward andalucia) 3) buenoeh diah 4) buenoh diah a Galician should use the standard variation, also "buenos dias", but "bos dias" (in Galician). Have you ever heard "buenor diah" ? Where do you get such informations? I would like to read about it. j.m. From mccay at redestb.es Wed Nov 4 16:28:35 1998 From: mccay at redestb.es (Alan R. King) Date: Wed, 4 Nov 1998 11:28:35 EST Subject: s > r: Iberian miscellanea Message-ID: ----------------------------Original message---------------------------- Referring to my subject line "r and s: Galician", Brian Mott wrote: >----------------------------Original message---------------------------- >The change of s to r before another consonanant is widespread in southern >Spain and also in Madrid, so that los dedos may be articulated as lor >dedoh and buenos dias as buenor diah. True, and quite a well-known fact (to me and presumably to some others in this discussion), but the language spoken in southern Spain and also in Madrid is not Galician, but Spanish (Castilian). I thought I was broadening the discussion. This is not to deny (indeed, it's surely all part of the fun) the potential relevance of the fact that both Galician and Castilian, besides sharing a common inheritance from Latin, are undoubtedly co-members of a linguistic area, that of the Iberian Peninsula, which arguably even includes Basque, a totally unrelated language, as Larry Trask will testify. But in an "areal" context it is all the more to the point to label the particular languages (or even dialects) we're talking about. The point of my posting was that (a part of) GALICIAN has rhotacism TOO. Also in reply to my posting, Miguel Carrasquer wrote: >And in dialectal Mallorqui' too. Same rules as in Galego and Sardo >(es bisbe "the bishop" > er birbe). Agreed. What strikes me as saliently different between the Galician and Mallorcan cases is not the similar rhotacism "rule" per se, but the respective overall phonologies that are the contexts, and which contrast fairly radically. Je m'explique. In Galician words can normally only end in a vowel, /s/, /N/ (velar except under assimilation), /l/, /r/ or /z/ (voiceless interdental fricative or voiceless sibilant, depending on dialect); for word-internal syllables the situation is more complex. But my point is that the range of CC sequences is somewhat limited, especially across word boundaries. In Catalan (of which Mallorcan is a dialect), on the contrary, word-final consonants are notoriously frequent (for a Romance language), resulting in a wide range of CC sequences across word boundaries. Now within Catalan, one of the salient phonological hallmarks of Mallorcan is the strong tendency to assimilate, delete or otherwise modify the first C in just such CC sequences, to a much greater extent than in peninsular Catalan (and inevitably still greater than in non-Catalan Ibero-Romance, where most such sequences are not found - at least across word boundaries - in the first place). Thus the s > r rule in Mallorcan is merely one of a long list of such sandhi rules, a mere leaf on the tree, so to speak. In Galician the same s > r process occurs, but attracts more attention, if you will, because there's no corresponding "tree". Relevant or not to the present discussion, I'm not sure; but maybe it's better to have too much information than to risk missing a significant insight because of incomplete data. Miguel also responded to my more recent posting, this one concerning Castilian and itself in response to Roger Wright's observation about what we could call, without prejudice for the discussion and our possibly different analyses, the sr > rr issue (in "Israel", "los reyes" etc.). Firstly, Miguel, the transcription [lor'ejes] is Roger Wright's (see his message), I just copied it, non mea culpa. Secondly, my intention was to support Wright's hypothesis (that here we may have, not dropping of /s/, but assimilation, complete or partial, of /s/ to following /r/); and I even coincide with his strategy in arguing for this, namely to build on the fact that in such forms [s] is not pronounced even in those varieties (and registers) of Spanish where wholesale loss or modification of syllable- or word-final sibilants is NOT the general norm in the first place. Apart from agreeing, I attempted to up the ante by expressing my surprise that anyone could suggest otherwise, since in my (non-specialist, impressionistic and spontaneous) perception the /s/ segment in question, in the Spanish varieties in question, is not dropped, it's just realized as something other than [s]. I repeat that I can't swear by this, nor prove it, it's just a subjective feeling. If you like I can restate my feeling by saying that I intue that in pronunciation, "Israel" still contains a consonant cluster of some sort. So the /s/ has partially assimilated, in my opinion; it isn't simply lost. I agree. If the textbooks say it's lost here, I wasn't aware (I do not specialize in Castilian, it's just a language I speak every day). The Basque Country is one area (partly) in the Iberian Peninsula where, in contrast to large regions of the peninsula, we can generalize (at least in synchronic terms) that final or preconsonantal consonants in general and sibilants in particular do NOT drop, nor do they get aspirated, rhotacised, or changed into anything other than sibilants. This statements applies to both the indigenous language, Basque, and to the Castilian spoken in the part of the Basque Country that is also part of the Iberian Peninsula, but for the present discussion I am focussing on the latter. I shall not go into whatever qualifications might be in order for Basque, which could include comment on Miguel's point about ezta = ez da, because it would only lead us up a side track. However I must qualify the Spanish side of my statement in a way that may be pertinent: in the Castilian spoken here, and as far as I know in most other places where Castilian is spoken and sibilants generally maintained, and for that matter in the Basque spoken on the southern side of the Pyrenees too, _pace Carrasquer_, sibilants are generally VOICED preceding any voiced consonant. I also believe this to be true of the other Romance languages in the Peninsula. Consequently I am surprised, and indeed puzzled, by Miguel Carrasquer's statement: >Spanish /s/ is very >very reluctant indeed to become voiced [it rather becomes [h] or [r] >than [z]], and a word like for me is always [asno], never >[azno]. In this respect, what is no doubt special about Castilian (together with Basque) vis-à-vis perhaps most other Romance languages in or out of the Iberian Peninsula (exceptions are Galician and Romanian) is that these languages have no PHONEME /z/ (voiced sibilant), hence Castilian speakers do not PERCEIVE voiced preconsonantal /s/ as a "z"; it is a mere conditioned allophone which, in my experience, normal untrained native speakers are absolutely unable to perceive (just as they can't perceive the difference between fricative and plosive allophones of their voiced stop phonemes). They say, indeed, [aZno] (with an "apical" Z in some varieties, naturally) for /asno/ "donkey", just as southern Basque speakers will normally say [eZne] for /eSne/ "milk". (On the other hand I have heard [eSne] from northern Basque speakers, providing a very neat "control group" as far as the Basque data is concerned.) So, back to rhotacism. Whatever may happen in other parts of Iberia (particularly in regions where syllable-final /s/ is in general unstable), in "Basque Country Spanish" where sibilants are stable (and, I should point out to many of our readers, the native dialect of a couple of million souls), at least, we should initially expect the sequence /sr/ to be realized phonetically as [Zr], with voicing of the sibilant preceding the voiced consonant, /r/, although still phonemically /sr/ (given that there is no /z/ phoneme and the voicing contrast is non-distinctive and predictable). The big point is that we DON'T find [Zr], we find something else which some of us in this discussion have transcribed [r], others [rr], and I for my part must for now transcribe as [?r] where I cannot state the nature of the segment represented by [?], but I feel there is something there, i.e. I am explicitly rejecting the suggestion that "Israel" is (in this kind of Spanish) phonetically and phonologically [irael] - /irael/ (where [r] and /r/ are the rolled or trilled r as in Spanish "irreal" - /ireal/ - [ireal]. Actually, I'm saying two things. That "Israel" is not pronounced, here, as if it were "Irrael". And that, in the phonology of (this form of) Spanish, "s" is never dropped, not even in "Israel". In which case, it is probably rhotacised. But only when an r follows. With apologies for a lengthy message, which I hope is useful. Alan From hubeyh at montclair.edu Wed Nov 4 16:26:42 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 4 Nov 1998 11:26:42 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > As Benji points out, human beings are woefully bad at estimating > probabilities. Mostly, I think, we tend to interpret `random Most people are quite bad at estimating numbers. > distribution' as `disperse distribution', meaning that we tend to assume > that independent events have a tendency to avoid one another. They > don't, or they wouldn't be random. Tendency to avoid one another, in the limit is called "mutual exclusivity", and such events are highly dependent. > The linguistic consequences of this failing are all too obvious. > Ancient Greek for `honey' was , and Hawaiian for `honey' is > . Wow! I can hear Arthur Koestler telling us that Something > Deeply Significant is going on here. But, of course, neither the Greeks > nor the Hawaiians had any interest in ensuring that their words for > `honey' were different, nor any means of doing so. Here you unfortunately are falling for the proof by example. Even induction is not valid in logic in the physical sciences and the social sciences. It is also not impossible for this food which is savored even by bears let alone humans to be from a very old word that belongs to protoworld. What is often forgotten is that diffusion processes which give rise to the GAussian density also have the property that if we divided up the density into discrete intervals and tested the number at various levels, the highest is always at 0 which would correspond in the linguistics case to "no change", in the same way that a drunkard who takes steps at random into any direction will most often be found where he started. There's no law that says that (1) linguistic change is 100% regular and (2) that if a sound X changes to Y it cannot change back to X again. If we are arguing by example, then add this; Turkic for honey is "bal" and also means "mud" and it probably does belong to protoworld. > Collect enough languages, and enough words, and you're going to be > drowning in such coincidences. Standard Italian `two' is a lot > more similar to Malay `two' than it is to Neapolitan Italian > `two'. In fact, looking at that list of number names on the Web > ( http://www.tezcat.com/~markrose/numbers.shtml ) can be quite an Mark Rosenfelder, nice guy that he is, produced it to reproduce some of my results. He got his p=0.001 to do that. But he forgot that what he has done is (1) for independent processes and (2) the 25 matches he got for English uses 100,000 or more words from English, not the Swadesh list. I posted before that there is a list in which quantitative reasoning is not off-limits in linguistics. It is called "language" and you can join it by sending email to majordomo at csam.montclair.edu. I am sure many people on that list will be more than happy to be illuminated by more of your examples. See ya' there. > illuminating experience. English /tu:/ and German /tsvai/ do not look > to be closely related, nor do the Kashmiri dialect variants /zi/ and > /do:/, nor do the Pashto variants /bu/, /lu/ and /do:v/, nor does > Armenian /erku/ look like anything else IE -- yet all are cognate. Don't forget that Armenia sat in an area inundated with Turkic speakers and it is "eki" or "iki" in that language, and unless that word can be found in Armenian circa 1,000 BC or earlier, there cannot be any proof that it was not due to borrowing. There is report of Kashogs (Kazak?) north of the Caucasus many centuries before the common era. Let us not also forget that Sumerian for two is "imma" which is one of the 165 cognates between Sumerian and Turkic, which is "ikki". The next edition of Dr. Tuna's book promises to have even more words. In fact, it is too easy. Besides, perhaps someone should compute the probability that if N alleged cognages are found that they will all display a different sound change. After all, if N presumed cognates are found, and it is highly probable that at least N/2 are repeated, then where does the hallowed "regular sound change" of intuitive historical linguistics go? How about if 300 "matches" can be found by "accident"? What is the probability that none of them are repeated? What if 150 are repeated? Is this "regular sound change" or not? This is what happens if badly made calculations run against other calculations. someone should try calculating these probabilities. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From DISTERH at UNIVSCVM.SC.EDU Thu Nov 5 16:07:21 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Thu, 5 Nov 1998 11:07:21 EST Subject: HISTLING on holiday Message-ID: Dear Colleagues, I will be away from my computer Friday, Nov. 6 through Tuesday Nov. 10. This means that HISTLING will also be on holiday, unfortunately. If you have any announcements that you would like to have posted before I leave tommorrow (Friday, Nov. 6) at 8 a.m. EST, please send them to me before that time. Of course you may send messages to HISTLING anytime while I'm gone, but they won't be distributed until late Tuesday night. My apologies for any conconvenience that this may cause you. Dorothy Disterheft Moderator, HISTLING From sally at isp.pitt.edu Thu Nov 5 16:05:08 1998 From: sally at isp.pitt.edu (Sarah G. Thomason) Date: Thu, 5 Nov 1998 11:05:08 EST Subject: rhotacism from Ray Hickey In-Reply-To: Your message of "Wed, 04 Nov 1998 11:26:42 EST." <363FD1A9.255A5469@montclair.edu> Message-ID: ----------------------------Original message---------------------------- H.M. Hubey's claim that "there cannot be any proof" that Armenian /erku/ is not a loanword from a Turkic language is mistaken: there are at least two, and perhaps more than two, other cognate sets between Armenian and other IE languages (e.g. Ancient Greek) with initial correspondences Armenian erk- : other IE dw-. One is a word for "long", the other a word for "fear/terrible" -- sorry, I don't have the actual forms in my head or by my desk. So this is a genuine recurring correspondence, reflecting PIE *dw- > Armenian (e)rk-, the end result of a series of regular sound changes in Armenian. It's well known; in fact, it's just about everybody's favorite example of a phonetically odd regular correspondence -- the example we tend to trot out to demonstrate that genuine correspondences need not be phonetically similar at all. (There are lots of other such examples; this is just the most familiar one, the one that probably everyone who has taken an introductory historical linguistics course has heard about.) Of course, Sumerian "imma" doesn't look very similar to Turkic "ikki", either; but in the absence of other pairs of words showing -mm- in Sumerian and -kk- in Turkic, no historical linguist would accept it as a promising cognate set, especially in the absence of *systematic* evidence of cognacy elsewhere in the lexicon (systematic, i.e. with recurring correspondences, as opposed to scattered similarities of the sort Trask was warning against). I must have missed part of this thread: has Hubey given a definition of "intuitive historical linguistics"? Is it his view that all historical linguistics that isn't supported by statistics is "intuitive"? -- Sally From mcv at wxs.nl Thu Nov 5 15:59:56 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 5 Nov 1998 10:59:56 EST Subject: s > r: Iberian miscellanea In-Reply-To: <3.0.3.32.19981104120907.00713060@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Consequently I am >surprised, and indeed puzzled, by Miguel Carrasquer's statement: > >>Spanish /s/ is very >>very reluctant indeed to become voiced [it rather becomes [h] or [r] >>than [z]], and a word like for me is always [asno], never >>[azno]. > >In this respect, what is no doubt special about Castilian (together with >Basque) vis-`-vis perhaps most other Romance languages in or out of the >Iberian Peninsula (exceptions are Galician and Romanian) is that these Do you mean Roman (or C/S.Italian in general)? Romanian certainly has /z/ ( < *dj, usually). >languages have no PHONEME /z/ (voiced sibilant), hence Castilian speakers >do not PERCEIVE voiced preconsonantal /s/ as a "z"; it is a mere >conditioned allophone which, in my experience, normal untrained native >speakers are absolutely unable to perceive (just as they can't perceive the >difference between fricative and plosive allophones of their voiced stop >phonemes). They say, indeed, [aZno] (with an "apical" Z in some varieties, >naturally) for /asno/ "donkey", just as southern Basque speakers will >normally say [eZne] for /eSne/ "milk". (On the other hand I have heard >[eSne] from northern Basque speakers, providing a very neat "control group" >as far as the Basque data is concerned.) Now I am pussled/nonpluzzed. As an abnormal, trained native speaker, I certainly *can* distinguish [s] and [z] (or apical [S] and [Z])... Let's see what Navarro-Toma's (my 1926 edition, but usually valid enough) has to say about the subject: S SONORA -- Alveolar fricativa sonora: ort. , fon. [z]. [...] La sonora aparece u'nicamente, en nuestra lengua en posicio'n final de si'laba, precediendo inmediatamente a otra consonante sonora; en cualquier otra posicio'n su presencia es anormal y espora'dica. Es siempre, asimismo, una articulacio'n breve y suave; la pronunciacio'n lenta o fuerte impide su sonorizacio'n, reapareciendo en su lugar la sorda. [...] {There's my error, of course. /asno/ is indeed [asno] pronounced in isolation, as a dictionary entry or object of introspection. It becomes [azno] only in connected speech, "when I'm not watching".} En el grupo (, , ), la se sonoriza como en los casos precedentes; pero la punta de la lengua, arrastrada por la ene'rgica articulacio'n de la [r_] [r stroke-above in the original, trilled r --mcv] siguiente, abandona la forma caracteri'stica de la estrechez redondeada que la punta de la lengua forma en la , haciendo perder a e'sta su timbre sibilante y articulando propiamente, en vez de [z] ordinaria y regular, una [R] [inverted r in the original, fricative r --mcv], o sea una fricativa, $114: [IRr_a at li'tA], [lORr_E'y at s], [dO'Rr_ at ales]; otras veces, en pronunciacio'n relativamente fuerte, la se pierde por completo, aumenta'ndose, a manera de compensacio'n, las vibraciones de la [r_] siguiente. {Rhotacism indeed. I must have picked up my non-native pronunciation [hr] ([initially] voiceless trill) from the speech of S/C parts, where -s > -h in connected speech, leading to voiceless [hB], [hG], [hm], [hn], [hl] and pre-aspirated [hp], [ht], [hk] in cases like , , , , , , , , as discussed by Toma's Navarro in subsequent paragraphs.} ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Thu Nov 5 15:57:59 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 5 Nov 1998 10:57:59 EST Subject: rhotacism from Ray Hickey In-Reply-To: <363FD1A9.255A5469@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: [Armenian erku < *dwo:] >Don't forget that Armenia sat in an area inundated with Turkic speakers >and it is "eki" or "iki" in that language, and unless that word can be >found in Armenian circa 1,000 BC or earlier, there cannot be any proof >that it was not due to borrowing. There's more than sufficient proof that Arm. erk- comes from *dw-. We have "birth pains" (cf. Greek "pain", Alb. "pain", from PIE *dau-, d at u-, du:-), Arm. "long time", "long" < PIE *deu-, *dwa:- (cf. Grk. d(w)a:n "long time"), Arm. "I fear" < PIE *dwei- (Grk. d(w)e(i)os "fear"), and of course Arm. "two" < PIE *dwo: (Grk. du(w)o:) There is also a mountain of evidence that PIE *o: > u in Armenian, so the hypothesis that * became by regular sound change (*dw- > erk-, *o: > u), is vastly superior at interpreting the facts than the hypothesis that the word was borrowed from Turkic, a hypothesis that fails to explain the /r/, the /k/ (Turkic /k/ is aspirated, so should have given Arm. /k`/), and the /u/, and explains the initial /e/ (phonetically [je] in Armenian) only by invoking a much less widespread variant of Turkic *iki. >Let us not forget that Sumerian for two is "imma" The Sumerian for "2" is actually or . ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From DRC at antnov1.auckland.ac.nz Thu Nov 5 15:57:35 1998 From: DRC at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 5 Nov 1998 10:57:35 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > ----------------------------Original message------------------------ ---- > Larry Trask wrote: > > > > As Benji points out, human beings are woefully bad at estimating > > probabilities. Mostly, I think, we tend to interpret `random > > Most people are quite bad at estimating numbers. > > > distribution' as `disperse distribution', meaning that we tend to assume > > that independent events have a tendency to avoid one another. They > > don't, or they wouldn't be random. > > Tendency to avoid one another, in the limit is called "mutual > exclusivity", > and such events are highly dependent. > > > The linguistic consequences of this failing are all too obvious. > > Ancient Greek for `honey' was , and Hawaiian for `honey' is > > . Wow! I can hear Arthur Koestler telling us that Something > > Deeply Significant is going on here. But, of course, neither the Greeks > > nor the Hawaiians had any interest in ensuring that their words for > > `honey' were different, nor any means of doing so. > > > Here you unfortunately are falling for the proof by example. Even > induction > is not valid in logic in the physical sciences and the social sciences. > It is also not impossible for this food which is savored even by bears > let alone humans to be from a very old word that belongs to protoworld. Who can say what is impossible? What is very unlikely is that the Hawaiians had a word for honey 200 years ago, since there were no honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century loanword from Greek, thanks to classically-educated missionaries who needed it to translate the Bible. Ross Clark From ratcliff at fs.tufs.ac.jp Fri Nov 6 01:21:11 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Thu, 5 Nov 1998 20:21:11 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > > It is also not impossible for this food which is savored even by bears > > let alone humans to be from a very old word that belongs to > protoworld. > What is often forgotten is that diffusion processes which give rise to > > the GAussian density also have the property that if we divided up the > density into discrete intervals and tested the number at various > levels, the highest is always at 0 which would correspond in the > linguistics case to "no change", in the same way that a drunkard who > takes steps at > random into any direction will most often be found where he started. > There's no law that says that (1) linguistic change is 100% regular > and (2) that > if a sound X changes to Y it cannot change back to X again. I am quite sympathetic with Mark Hubey's ambition to expand the use of quantitative methods in HL, but it won't do to ignore the EMPIRICAL foundations of the discipline. Mathematical models have to be made to serve the evidence. The proposal that given enough time X>Y>X is plausible is based on a hypothesis that the directionality of sound change is random. It isn't. There are clearly preferred (frequently attested) and disprefered (rarely or not attested) directions for sound change. The only systematic work on this I know of is in Ch. 5 of Labov's 1994 "Principles of Language", which only deals with vowels. But every working historical linguist has probably developed his own practical database. Off the top of my head I might suggest that lenition processes (shift of a stop to a homorganic fricative or approximant) are more common than fortition processes, that shifts to a neighboring point of articulation are more common than shifts to a distant one, that among shifts of the latter type shifts from velar to palatal are more common than shifts from velar to uvular, and that shifts from dental to alveolar, or alveolar to palatal are more common than shifts from dental to labial. Without a substantial body of evidence of actual changes, and a statistical determination of likely and unlikely paths of change, we don't even know what types of correspondences to expect in long distance relationships, or what to look for when trying to discover them. The search for identities or 'similarities' may be right, but could just as well be completely wrong. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From mccay at redestb.es Fri Nov 6 12:19:10 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 6 Nov 1998 07:19:10 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: > >>In this respect, what is no doubt special about Castilian (together with >>Basque) vis-`-vis perhaps most other Romance languages in or out of the >>Iberian Peninsula (exceptions are Galician and Romanian) is that these > >Do you mean Roman (or C/S.Italian in general)? Romanian certainly >has /z/ ( < *dj, usually). Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact that besides Castilian, both Galician and Romanian are also exempt from this general Romance development. In Castilian and Galician (but NOT in Romanian, as you point out), there is no /z/ phoneme. I was confusing these two things. From hubeyh at montclair.edu Fri Nov 6 12:21:37 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:21:37 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Sarah G. Thomason wrote: > > ----------------------------Original message---------------------------- > > H.M. Hubey's claim that "there cannot be any proof" that Armenian > /erku/ is not a loanword from a Turkic language is mistaken: there > are at least two, and perhaps more than two, other cognate sets > between Armenian and other IE languages (e.g. Ancient Greek) with > initial correspondences Armenian erk- : other IE dw-. One is a > word for "long", the other a word for "fear/terrible" -- sorry, I > don't have the actual forms in my head or by my desk. There are different ways to reason; one way with certainty is logic. REasoning under uncertainty is probabilistic or fuzzy-logical. Most people who write about science are philosophers and they mostly use logic. Under this particular type of reasoning, proof is possible only in math. No proof of any laws of physics is possible let alone linguistics. Since borrowings are also regularized and since there is a tendency in language to 'regularize' including 'hypercorrection', it does not matter how many sound correspondences are found. The laws of linguistics are based on probability theory (as are laws of physics). At best we can obtain statements to the effect that such and such occurences are unlikely to be due to chance. This, of course, does not rule out 'borrowing'. > Of course, Sumerian "imma" doesn't look very similar to Turkic > "ikki", either; but in the absence of other pairs of words showing > -mm- in Sumerian and -kk- in Turkic, no historical linguist would > accept it as a promising cognate set, especially in the absence > of *systematic* evidence of cognacy elsewhere in the > lexicon (systematic, i.e. with recurring correspondences, as opposed > to scattered similarities of the sort Trask was warning against). Sure, there are 165 of them in Dr. Tuna's book. And the sound change m > k is one of them. And, of course, there are other examples of this sound change. Once again, if regular sound changes are "proof" then Tuna's work has to be treated consistently as are works are treated. > > I must have missed part of this thread: has Hubey given a > definition of "intuitive historical linguistics"? Is it his view > that all historical linguistics that isn't supported by statistics > is "intuitive"? During the last century engineers used the laws of 'hydraulics' which were derived empirically but could not be obtained from the laws of physics. Things changed during the early part of this century. Historical linguistics is circular, especially as it is based mostly on IE. IT says; 1. The set of languages, {x,y,z...} constitutes a family because of 'regular sound correspondences'. 2. 'Regular sound correspondences' indicate a 'language family'. Surely, this is as circular as; 1. I am Napoleon, and my friend here is General Marat, and he can testify that I am Napoleon. 2. I am General Marat and this is Emperor Napoleon. Any two clowns can memorize and repeat these lines. In the case of linguistics, something else is needed. I noted the basics above. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:21:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:21:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > > There is also a mountain of evidence that PIE *o: > u in Armenian, so > the hypothesis that * became by regular sound change > (*dw- > erk-, *o: > u), is vastly superior at interpreting the facts > than the hypothesis that the word was borrowed from Turkic, a > hypothesis that fails to explain the /r/, the /k/ (Turkic /k/ is > aspirated, so should have given Arm. /k`/), and the /u/, and explains > the initial /e/ (phonetically [je] in Armenian) only by invoking a > much less widespread variant of Turkic *iki. Mountain of evidence is sufficient :-) > >Let us not forget that Sumerian for two is "imma" > > The Sumerian for "2" is actually or . As Tanenbaum writes in one of his books; "if you don't like this year's standards wait until next year." :-) The word is 'imma' according to the books and journal articles used by Tuna. If next year, it is changed from 'min' to 'bobooo', then we can discuss that too. The strange thing is that there are others who are making these changes (one of them on the WWW) and as they change things they keep changing things to other cognate forms :-) There is no escape :-) -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:22:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:22:25 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > Who can say what is impossible? What is very unlikely is that the > Hawaiians had a word for honey 200 years ago, since there were no > honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century > loanword from Greek, thanks to classically-educated missionaries who > needed it to translate the Bible. Well, there is nothing like a real test :-) Probability theory can only tell us if something is likely to be due to chance, and not whether it was borrowed or descended from an earlier form of the same language. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:22:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:22:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > I am quite sympathetic with Mark Hubey's ambition to expand the use of > quantitative methods in HL, but it won't do to ignore the EMPIRICAL > foundations of the discipline. Mathematical models have to be made to > serve the evidence. The proposal that given enough time X>Y>X is > plausible is based on a hypothesis that the directionality of sound > change is random. It isn't. There are clearly preferred (frequently This is something that used to bother me a lot and I used to think that there must be some universality. I am still not sure that there is not but I am more sure now that some of what we see is not really universal but rather due to some specific local phenomena. > attested) and disprefered (rarely or not attested) directions for sound > change. The only systematic work on this I know of is in Ch. 5 of > Labov's 1994 "Principles of Language", which only deals with vowels. This has to be from a specific time period and a specific set of languages. There is a similar problem in probability theory, that of stationarity of a signal. It's impossible to prove. No matter how many languages you look into you can only look at a specific time interval. It's hard to say if the sound changes are due to universals or to a particular combination of sounds and phonotactics. However, what I wrote above referred, in general, to any two sounds. It may be that for some specific X and some specific Y, the sound change X > Y for some specific language (i.e. specific set of phonemes and phonotactics) may be irreversible. But in general I do not see any reason to assume that no sound change is reversible given enough time. > But every working historical linguist has probably developed his own > practical database. Off the top of my head I might suggest that lenition > processes (shift of a stop to a homorganic fricative or approximant) are > more common than fortition processes, that shifts to a neighboring point > of articulation are more common than shifts to a distant one, that among > shifts of the latter type shifts from velar to palatal are more common > than shifts from velar to uvular, and that shifts from dental to > alveolar, or alveolar to palatal are more common than shifts from dental > to labial. Without a substantial body of evidence of actual changes, > and a statistical determination of likely and unlikely paths of change, > we don't even know what types of correspondences to expect in long > distance relationships, or what to look for when trying to discover > them. The search for identities or 'similarities' may be right, but > could just as well be completely wrong. Right. However, it is strange on the other hand to see those consonant clusters and lack of vowels in languages like Abaza, Georgian, or Khoisan and its clicks. The foremost question is this: if there is such a universal trend (say toward lack of cases, or toward voicing, or from stops to fricatives, or approximants) how then did the language (any language) get those stops in the first place? Or how did some language get consonant clusters at all? HOw did a language get so many cases? Etc Etc. That is the main reason I am not so sure anymore about those alleged universals, and instead try to explain them as a part of a greater and long range pattern due to interactions of specific types of languages. The most important thing to remember in these occurrences is to think about phenomena at multiple scales. You can see its workings in physics most clearly and precisely. We have lots of equations for describing things at the sub-atomic level. But most of them are useless for chemistry. At the same time, we have descriptions of phenomena at the level in which we treat the material as continuous things (i.e. we call it continuum mechanics) and in which we do not even think about the atomic composition of matter. It works. Most of what you see around you was designed and created using the equations of continuum mechanics at best. At the next level up, we treat things a lumped objects, and even ignore the continuum aspects. That works too. The same applies to linguistics changes. There are many scales at which changes occur, and if we mix up these levels we create opposing ideas. > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe > Senior Lecturer, Arabic and Linguistics, > Dept. of Linguistics and Information Science > Tokyo University of Foreign Studies > Nishigahara 4-51-21, Kita-ku > Tokyo 114 Japan -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From larryt at cogs.susx.ac.uk Fri Nov 6 12:27:44 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Fri, 6 Nov 1998 07:27:44 EST Subject: Hawaiian In-Reply-To: <31E5B5928AE@antnov1.auckland.ac.nz> Message-ID: ----------------------------Original message---------------------------- On Thu, 5 Nov 1998, Ross Clark wrote: > Who can say what is impossible? What is very unlikely is that the > Hawaiians had a word for honey 200 years ago, since there were no > honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century > loanword from Greek, thanks to classically-educated missionaries who > needed it to translate the Bible. Nice one! I didn't know that, even after watching about 700 David Attenborough programs on TV. A pity, though, since it ruins one of my favorite examples of chance resemblance. But at least it becomes a particularly interesting example of borrowing. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 10 23:15:27 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:15:27 EST Subject: rhotacism from Ray Hickey In-Reply-To: <36428F6E.B04C8B45@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > But in general I do not see any reason to assume that no sound > change is reversible given enough time. I assume we are speaking of phonological change, and not of things like spelling pronunciation and the influence of prestige varieties. That being so, certain sound changes are indeed reversible. For example, PIE */t/ generally changed to theta in Germanic, but this theta has changed back to /t/ in the continental Scandinavian languages. This is `Rueckverwandlung', or `reversal'. It is not particularly common. But some sound changes are quite irreversible. Consider loss. In the ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting /h/ was later lost. I predict confidently that the Greeks will never reverse this change by re-introducing those long-gone /s/s, since no ordinary speaker has the faintest idea where the /s/s used to be, or even that they were ever there at all. And the spontaneous change of nothing to /s/ is not something that has been observed very often, if indeed ever at all. Another type of irreversible change is merger. Once accomplished, a merger cannot be reversed by purely linguistic means -- though it *can* sometimes be reversed by external means, such as by the influence of another variety which has not undergone the merger. For example, the vowels of `toe' and `tow', and of `nose' and `knows', were formerly distinct in English, but they have merged in all varieties except for some rural areas of England (and Wales?). Again, I am confident that the merger cannot now be reversed. > Right. However, it is strange on the other hand to see those > consonant clusters and lack of vowels in languages like Abaza, > Georgian, or Khoisan and its clicks. The foremost question is this: > if there is such a universal trend (say toward lack of cases, or > toward voicing, or from stops to fricatives, or approximants) how > then did the language (any language) get those stops in the first > place? Or how did some language get consonant clusters at all? HOw > did a language get so many cases? Etc Etc. Consonant clusters can and do arise from various sources. Perhaps the most frequent is the loss of intervening vowels. For example, in the English of southern England, `police' is pronounced /pli:s/, `collect' is /klekt/, `correct' is /krekt/, `collapse' sounds like `claps', and so on, producing many new instances of word-initial clusters. In many varieties of English, words like `camera', `chocolate' and `family' have lost their medial vowels, and in England the same is true of `medicine', which is /medsin/ -- all now with consonant clusters which were formerly absent. As is well known, similar things happen in French, in which `small' is pronounced /pti/. Pyrenean varieties of Basque have undergone extensive syncope, so that common `you are' (phonetically [sara]) becomes , `they are' becomes , and so on, producing numerous initial and medial clusters which were formerly absent from the language. But there are other mechanisms, such as unpacking. In some varieties of Basque, the historical palatal nasal /n~/ has been unpacked into a cluster /jn/, and in some varieties of French palatal /n~/ has been unpacked into the cluster /nj/. English has acquired some final clusters by excrescence: `vermin' to regional `varmint', `no' to `nope', `amiddes' to `amidst', `betwix' to `betwixt', and so on. And note also cases like `empty' and `thunder', whose /p/ and /d/ were formerly absent but have been inserted by epenthesis, presumably to ease the transitions between unlike sounds. The rise of cases has been much investigated. Perhaps the single most frequent source of new cases is the reduction of postpositional phrases. For example, both Hungarian and Finnish have many more cases than can be reconstructed for their Proto-Uralic ancestor, and the origins of many of these cases are well understood: they derive from the reduction of postpositional phrases. In the same way, the Basque comitative case-ending <-ekin> `with' derives from the reduction of a postpositional phrase *<-en kidean> `in the company of'. New plosives can arise by simple fortition. For example, original /w/ has been strengthened to /b/ in some American languages, and the glide /j/ (= English ) has been strengthened to a variety of fricatives, affricates and plosives in many varieties of Basque, with some of these being devoiced as well. And some western varieties of Basque have changed the sequence /ua/ first to [uwa] and then to [uba], acquiring new instances of /b/ in the process. But there are other sources of new plosives. For example, Dutch /sx/ has been dissimilated to /sk/ in Afrikaans, resulting in new instances of /k/. So, all of the questions that Mark asks are interesting ones, but they have answers which have been largely worked out in the only way possible: by looking at the evidence. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From delancey at darkwing.uoregon.edu Tue Nov 10 23:16:52 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Tue, 10 Nov 1998 18:16:52 EST Subject: rhotacism from Ray Hickey In-Reply-To: <36428B3F.CF3BFD02@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > Historical linguistics is circular, especially as it is based mostly on > IE. > IT says; > > 1. The set of languages, {x,y,z...} constitutes a family because of > 'regular > sound correspondences'. > 2. 'Regular sound correspondences' indicate a 'language family'. Could you give us a couple of examples of recognized language families that have been established on the basis of this kind of reasoning? There are a few, but it's not the standard methodology at all. Not Indo-European, for sure, which was first proposed and established on the basis of extensive correspondences in morphological paradigms. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Tue Nov 10 23:19:18 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:19:18 EST Subject: The good Dr. Tuna In-Reply-To: <36428C1B.656F3324@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > > >Let us not forget that Sumerian for two is "imma" [Miguel Carrasquer Vidal] > > The Sumerian for "2" is actually or . > As Tanenbaum writes in one of his books; "if you don't like this year's > standards wait until next year." :-) > The word is 'imma' according to the books and journal articles used by > Tuna. If next year, it is changed from 'min' to 'bobooo', then we can > discuss that too. The strange thing is that there are others who are > making these changes (one of them on the WWW) and as they change things > they keep changing things to other cognate forms :-) There is no escape > :-) I do not know Tuna's work, and I do not know if he is a specialist in Sumerian. But apparently he is not, if he has indeed had to extract his Sumerian forms from other people's publications. And relying on secondary sources for languages you are not personally acquainted with is a very dangerous practice: you often wind up merely repeating and propagating other people's errors, as well as perhaps introducing a few more of your own, through misunderstanding. Here's a brief example, which is all too typical of so many of the attempts I have seen at comparing languages carried out by investigators not personally familiar with the languages they are comparing. Please note that I am *not* trying to get at any of the people named below. I am merely pointing out the dangers of such work. In 1985, V. Chirikba published a comparison of Basque with various North Caucasian languages, mainly Abkhaz, arguing that his comparisons constituted evidence for a genetic link. Chirikba's work was cited in part by Shevoroshkin and Manaster Ramer in their article in the 1991 Lamb and Mitchell book. S and M-R conclude that there is "some plausibility to many of [C's] comparisons". By this I suppose they mean that the comparisons look good on the page. Fine, but unfortunately the Basque *data* on the page are mostly wrong. C is a specialist in Caucasian, but he doesn't know Basque, and he has relied upon various secondary sources for his Basque data. Bad move: these sources, whatever they were, were clearly anything but reliable. Moreover, C has introduced some further errors of his own, mainly in his erroneous segmentations. The 34 of C's comparisons repeated by S and M-R are numbered 78-111. On the Basque side, some of the forms are correct. However. (When I say that a Basque word is "arbitrarily segmented", I mean that C has arbitrarily extracted a portion of it which does not match and thrown it away without explanation.) [78] is cited wrongly, badly wrongly. [80] does not exist. [81] is hopelessly misglossed. [84] is slightly misglossed. [89] is a bimorphemic form containing a case-suffix but wrongly cited as monomorphemic, even though the case-suffix itself constitutes item [83]. [90] consists of another bimorphemic (case-suffixed) form wrongly cited as monomorphemic, plus a second truly monomorphemic item unrelated to the first (and wrongly glossed). [91] does not exist, but obviously represents a garbled attempt at citing a stem whose meaning is utterly different from what is suggested. [92] does not exist. [93] does not exist. [95] is wrongly glossed and assigned to the wrong part of speech. [96] is wrongly cited. [98] is cited only in a secondary sense. [99] does not exist, and is moreover arbitrarily segmented. [100] is wrongly glossed. [101] is cited only in a secondary regional form, and is arbitrarily segmented. [102] is arbitrarily (and impossibly) segmented. [104] is a transparent compound wrongly presented as monomorphemic. [105] is cited only in a regional secondary variant, and it is both wrongly glossed and assigned to the wrong part of speech. [108] does not exist. [109] is cited only in a regional variant lacking the initial /g/ found elsewhere, and is glossed only in a secondary sense. [111] is cited as monomorphemic, even though the second half of it is precisely the item cited separately as item [80]. So, of the 34 Basque items compared, 21 are erroneous or non-existent. Of the remaining 13, three are single segments, seven are monosyllables in which only a single segment matches anything in Caucasian, and one more contains a sequence of three segments unmatched in the Abkhaz comparandum. The case therefore rests almost entirely upon the erroneous Basque data, while the real Basque data lend no support. It is very, very dangerous to try to do comparisons on languages you do not have a specialist knowledge of. You have to assume that everybody who has intervened between the native speakers and you is virtually infallible. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From l.campbell at ling.canterbury.ac.nz Tue Nov 10 23:20:01 1998 From: l.campbell at ling.canterbury.ac.nz (Lyle Campbell) Date: Tue, 10 Nov 1998 18:20:01 EST Subject: Hawaiian meli Message-ID: ----------------------------Original message---------------------------- Ross Clark's example of Hawaiian -meli- 'honey' as a loanword from Greek via classically-educated missionaries translating the Bible is a very nice one, as Larry Trask pooints out. Larry notes that it ruins one of his favorite examples of chance resemblance, but becomes a particularly interesting example of borrowing. I suppose you could sort of raise the ante on interestaing cases of chance resemblance and borrowing by throwing Maori -mieri- 'honey' into the mix. A comparison of Hawaiian -meli- and Maori -mieri- (bother the -i- vowel difference) might seem to suggest a Polynesian cognate set (throw in Niuean -meli- 'heney' as well, also apparently from Greek), but the Maori word is actually a French loanword (from French -miel- 'honey'), courtesy apparently of early French Catholic missionizing activities in New Zealand, which soon faded in the country. (There are not many French loans in Maori, but a favorite is -wi:wi:- 'French' < French -oui- 'yes'). As Larry Traks points out, this -meli- / -mieri- false cognate is no longer a case of sheer accidental similarity, in that both are from Indo-European languages, but we still have accident to thank for it in a way, in that by sheer happenstance Hawiian ended up with a Greek form and Maori with a French one (which happen to be related languages), not something that would have been expected. Lyle Campbell From hubeyh at Montclair.edu Tue Nov 10 23:21:07 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Tue, 10 Nov 1998 18:21:07 EST Subject: The good Dr. Tuna Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > I do not know Tuna's work, and I do not know if he is a specialist in > Sumerian. But apparently he is not, if he has indeed had to extract his > 1. You mean you would not believe me if I said 2+2=4 because I am not a mathematician? 2. There's a thing called "Argument from Authority". It is one of the classic fallacies of logic. 3. What exactly is there to not being a Sumerian except that you spent N years more on it than someone else? Either the words published in the "scholarly journals" are close to the sound and semantic shape or are not? I have seen people on the WWW do similar things. I noted one cognate and sent email. he changed it. Now there's another one that I noted, which is a cognate, and if I send email to him, he will probably change that too :-) 4. On the ANE list, scholarly discussion is defined as discussion of ideas published in scholarly journals. Scholarly journals are journals in which scholars publish. Scholars are those who publish their ideas in scholarly journals. What is the point of all this? Either what is written uses commonly accepted reasoning or it does not. 5. The simple facts are that there are simple rules in linguistics and it does not take longer than a few minutes to get the hang of the basic idea, but the justification is often missing, and the practice is usually pretty muddled. So far I have found no book on historical linguistics in which there is a clear algorithm for reconstructing protoforms. And i have asked many people about this. They don't know either. Sometimes they can justify it easily sometimes it is a guess and sometimes it is no more than a concensus. You can't get scientific truth by a vote unless it can be clearly explained and backed up by scientific methodology. Some of the worst linguists are those who merely repeat what they have managed to memorize. It's better to become a gardener than to get a PhD in that manner. Let us also recall some of the things done by amateurs. Gauss, one of the greatest mathematicians of all time, had to publish his own book on Arithmetic and Group theory. You can find similar stories in every field. Either you should get your own copy of Tuna's book or get it from a library and xerox it and then make up your mind instead of using the IBM FUD factor (fear, uncertainty, doubt) in the minds of the readers. > Sumerian forms from other people's publications. And relying on > secondary sources for languages you are not personally acquainted with > is a very dangerous practice: you often wind up merely repeating and > propagating other people's errors, as well as perhaps introducing a few > more of your own, through misunderstanding. That is pretty useless. Nobody can learn every language and nobody can be a mathematician, physicist, engineer, computer scientists, chemist, economist, etc. That is not how it works at all. In fact, there is a thing called "specialization of labor" and has been around for a long time. In fact, it is used by linguists all the time, since they defer to specialists all the time, even to the point of committing the "argument from authority" fallacy. Where are the books on Sumerian? Where are all the books on Hittite? I have a batch of them. It took be 6 months to collect some and xerox the others. Does this mean that I must not trust that the OI has done poor work and that I should not rely "on others". What if I start to look at the original text? What exactly is there that is going to take 12 years to learn? Some symbols arranged in some order with some presumed meanings which can be gleaned from multilingual transcriptions. What is there to gain? Sure, maybe I can spot some things, or maybe I can fit another set of words to the original. But something that has gone on for a long time and which creates such long sentences must be, in the main, correct. One does not need to know much more than simple probability theory to know that. And that is exactly what people use to reach their conclusions, even if they use probability theory only intuitively. > > > Here's a brief example, which is all too typical of so many of the > attempts I have seen at comparing languages carried out by investigators > not personally familiar with the languages they are comparing. Please > note that I am *not* trying to get at any of the people named below. > I am merely pointing out the dangers of such work. > Maybe you are a poor linguist and not Manaster-Ramer :-) That is why so many people fight all the time in linguistics. And there is a way to make sure that things can be done so that they are comparable. That way obviously is to present the results in a way in which all scientists do. That is why they use math, and that is why eventually it will be used in linguistics. Finally, once again, you are arguing by pseudo-induction. Let me ask if this argument works: Once, I broke my leg going up an escalator. So don't go up escalators. This is not even induction which itself does not work for science. Here is a case of 'bad induction": The farmer feeds the goose 35 days in a row. So he will feed it on the 36th day. No. He might decide to have roast duck instead. Here is another bad case of induction: Every saturday for the past 10 weeks it rained. So it will rain on the 11th. Wrong. So why do you think that I would take your argument seriously? Surely, there are bad linguists as well as bad engineers. And there is bad linguistics as well as good. So what? How do you propose to fix this problem? Let me guess. You want to have a stamp by the International Linguistics Association and you want to stamp every book that even touches upon linguistics as bad and good. Is that a solution? Well, the arguments often made by linguists, (like a few already made here) are essentially informal versions of this. And they stink as badly as the formal version would stink. > > In 1985, V. Chirikba published a comparison of Basque with various North > Caucasian languages, mainly Abkhaz, arguing that his comparisons > constituted evidence for a genetic link. Chirikba's work was cited in > part by Shevoroshkin and Manaster Ramer in their article in the 1991 > Lamb and Mitchell book. S and M-R conclude that there is "some > plausibility to many of [C's] comparisons". By this I suppose they mean > that the comparisons look good on the page. Fine, but unfortunately the > Basque *data* on the page are mostly wrong. C is a specialist in > Caucasian, but he doesn't know Basque, and he has relied upon various > secondary sources for his Basque data. Bad move: these sources, > whatever they were, were clearly anything but reliable. Moreover, C has > introduced some further errors of his own, mainly in his erroneous > segmentations. > > The 34 of C's comparisons repeated by S and M-R are numbered 78-111. On > the Basque side, some of the forms are correct. However. > > (When I say that a Basque word is "arbitrarily segmented", I mean that C > has arbitrarily extracted a portion of it which does not match and > thrown it away without explanation.) > > [78] is cited wrongly, badly wrongly. > > [80] does not exist. > > [81] is hopelessly misglossed. > > [84] is slightly misglossed. > > [89] is a bimorphemic form containing a case-suffix but wrongly cited as > monomorphemic, even though the case-suffix itself constitutes item [83]. > > [90] consists of another bimorphemic (case-suffixed) form wrongly cited > as monomorphemic, plus a second truly monomorphemic item unrelated to > the first (and wrongly glossed). > > [91] does not exist, but obviously represents a garbled attempt at > citing a stem whose meaning is utterly different from what is suggested. > > [92] does not exist. > > [93] does not exist. > > [95] is wrongly glossed and assigned to the wrong part of speech. > > [96] is wrongly cited. > > [98] is cited only in a secondary sense. > > [99] does not exist, and is moreover arbitrarily segmented. > > [100] is wrongly glossed. > > [101] is cited only in a secondary regional form, and is arbitrarily > segmented. > > [102] is arbitrarily (and impossibly) segmented. > > [104] is a transparent compound wrongly presented as monomorphemic. > > [105] is cited only in a regional secondary variant, and it is both > wrongly glossed and assigned to the wrong part of speech. > > [108] does not exist. > > [109] is cited only in a regional variant lacking the initial /g/ found > elsewhere, and is glossed only in a secondary sense. > > [111] is cited as monomorphemic, even though the second half of it is > precisely the item cited separately as item [80]. > > So, of the 34 Basque items compared, 21 are erroneous or non-existent. > Of the remaining 13, three are single segments, seven are monosyllables > in which only a single segment matches anything in Caucasian, and one > more contains a sequence of three segments unmatched in the Abkhaz > comparandum. The case therefore rests almost entirely upon the > erroneous Basque data, while the real Basque data lend no support. > > It is very, very dangerous to try to do comparisons on languages you do > not have a specialist knowledge of. You have to assume that everybody > who has intervened between the native speakers and you is virtually > infallible. > > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From hubeyh at montclair.edu Tue Nov 10 23:21:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:21:25 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > That being so, certain sound changes are indeed reversible. For > example, PIE */t/ generally changed to theta in Germanic, but this theta > has changed back to /t/ in the continental Scandinavian languages. This > is `Rueckverwandlung', or `reversal'. It is not particularly common. So what does that mean? If there are only 20 consonants and 10 vowels there are (30)(30)-30 = 870 possible sound changes and out of this only 30 can be changing back to the original. > But some sound changes are quite irreversible. Consider loss. Why irreversible? don't languages sometimes add sounds? > Another type of irreversible change is merger. Once accomplished, a > merger cannot be reversed by purely linguistic means -- though it *can* What happened to adding of sounds? > But there are other mechanisms, such as unpacking. In some varieties of > Basque, the historical palatal nasal /n~/ has been unpacked into a > cluster /jn/, and in some varieties of French palatal /n~/ has been > unpacked into the cluster /nj/. So sounds can be added after all. > English has acquired some final clusters by excrescence: `vermin' to > regional `varmint', `no' to `nope', `amiddes' to `amidst', `betwix' to > `betwixt', and so on. And note also cases like `empty' and `thunder', > whose /p/ and /d/ were formerly absent but have been inserted by > epenthesis, presumably to ease the transitions between unlike sounds. Ditto. > So, all of the questions that Mark asks are interesting ones, but they > have answers which have been largely worked out in the only way > possible: by looking at the evidence. CAn you look at evidence that you don't have? This is only what we can see. So then what is wrong with "if it can happen once, it can happen twice" or "if it happened in the past, it can happen in the future". That means that unless there are many many examples collected over many many years of changes that "never" seem to occur, then it is still "all possibly go". -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:21:52 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:21:52 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Scott DeLancey wrote: > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > > > Historical linguistics is circular, especially as it is based mostly on > > IE. > > IT says; > > > > 1. The set of languages, {x,y,z...} constitutes a family because of > > 'regular > > sound correspondences'. > > 2. 'Regular sound correspondences' indicate a 'language family'. > > Could you give us a couple of examples of recognized language families > that have been established on the basis of this kind of reasoning? > There are a few, but it's not the standard methodology at all. > Not Indo-European, for sure, which was first proposed and established > on the basis of extensive correspondences in morphological paradigms. Aren't morphological paradigms also part of 'regular sound change"? > Scott DeLancey > Department of Linguistics > University of Oregon > Eugene, OR 97403, USA > > delancey at darkwing.uoregon.edu > http://www.uoregon.edu/~delancey/prohp.html -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:23:19 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:23:19 EST Subject: [Fwd: Turkic-Sumerian Connection Message-ID: Here are some of the words, some of them with additions by me to the original post by someone else. -------------------------------------------------- Here are some more additions. In a couple more iterations we can have more words for everyone to look at. On Mon, 3 Nov 1997, TIMUR KOCAOGLU wrote: >( Sm. = Sumerian, Tk. = Turkic, % = the loss of the consonant) > >A. Sound equations which CANNOT be recognized easily: > >a. Initial consonants: D, g, m, n, S, Sh, % > >I. Sm. .D ~ Tk. .y, % > >1. Sm. dar : 'spalten, zerschneiden, zerstoeren' (D.71; MSL, III, 100) > Tk. yar-: 'yarmak' [to break in the middle] (DLT, I, 399; KBI, 523) > >2. Sm. dib : 'Band' (D. 83) > Tk. yip : 'ip' [string, rope] (KBI, 546) Note that in some Kipchak dialects these words would be /car/ and /cip/. Furthermore, in d-Bolgharic they would be /dar/ and /dip/. Here are a few more: Sm. dir : zerspalten, zerstoren, umwerfen (D.86) Tk. yir : yirmek (DLT, III, 58) Sm. dirig : 'to be excessive, to be too much, too many (Grd.341) 'ubervoll sein, voll sein (D.87), diri 'ubergrosss'. Tk. irig : 'sert, kaba, hasin, gayretli' (KBI,199) iri 'kaba, sert', 'large, huge, voluminous, big' (Rd. 546) This is the word that showed up in the name Eridanus (i.e. iri) by Herodotus. And again, Doerfer has reconstructed proto-Turkic as *d > *d' > c > y > 0. In some cases the initial y >0 in Turkic languages such as cip/ip, cilan/yilan/ilan, cirt/yirt, cer/yer. These would have originally been dip, dilan, dirt, der, etc according to Doerfer, and Doerfer did this decades before Tuna's book came out, and Doerfer is a big name in Altaic/Turkic studies. Furthermore the old name for the Yayik River (Volga/Don?) shows up in Greek as daichs/daix exactly as it should have been. And it shows up circa 200 BC. Other d-Bolgharic words have been deciphered on Runic writings such as daga, dog, der, in addition to one that was already known, dilom (yilan=snake). Sm dirra : 'Hulfe' (D.87) Tk. yarI " (yardim) (help) Sm. tir : 'country' (MSL,III.87) Tk. yir : 'yer, toprak, yeryuzu' (KBI, 546)(land, earth) Sm. tu(5): 'waschen, baden, libieren' (D.206) Tk. yu(w): yikamak (to wash) > >II. Sm. .g ~ Tk. %, .y > >1. Sm. gamar : 'wuctig sein' (D.41) > Tk. agIr : 'aghIr' [heavy] (DLT, I, 52) > >2. Sm. garim : 'Fluss-Aue' (MSL, III, 109) > Tk. arIq : 'Irmak' [river, canal] Sm. gaz : 'to crush' (Grd.356), gaza (zerbrechen (MSL,III.143) Tk. ez : 'to crush, to pound', ezme 'crushed, pounded' Sm geme: 'Magd' (MSL,III.1250 Tk. eke : 'buyuk kiz kardes' (DLT,I,685) (older sister) Sm. gishig : 'Tur' (D.130), 'door' (EHG,436) Tk. eshik : 'kapi' (KBI,206) (door) >III. Sm. .m ~ Tk. .K > >1. Sm. mal : (Emesal) 'to stay' (for gal), (Grd. 384) > Tk. kal-: 'kalmak' [to stay] (DLT, I, 41; KNI, 215) > >2. Sm. marun : 'Ameise' (D. 160) > Tk. karIncha: 'karInca' [ant] (DLT, I, 501; III, 375) Sm. mir : 'anger' (MSL,IV,35) Tk. kiz : 'to be angry, cross..." SM. mu : 'name, fame' (Grd.388, MSL,V,220) Tk. ku : "un, shan' (DLT,III,212) (fame) >IV. Sm. .n ~ Tk. .y > >1. Sm. nad(9) : 'sich niederlagen, beschlafen; sich lagern' (D. 168; MSL, III,152) > Tk. yad- : 'yaymak, doshemek, sermek' [to spread, to lay down] (DLT, I, 15) The name "yayik", or "daichs" comes from this root. Sm. nigin : 'Summe' (D,171,MSL,III,111) Tk. yigin : 'yigin,kume, yigilmis' This word shows up closer to original form in Balkar as nigish. Sm. nunuz : 'bead' (EHG, 21; Falkenstein, 29) Tk. yinchu: 'inci,cariye (DLT,I,273) Sm. nurum : 'Licht' (D.170) Tk. yaruk : 'isik, aydinklik, parlak' (DLT,I,46)(light) >V. Sm. .S ~ Tk. .y, % > >1. Sm. sar : 'schreiben' (MSL, III, 113), 'to write' (Grd. 403) ~ shar id. > Tk. yaz-: 'shashmak, yanIlmak, chozmek, yazmak' [to write with > other meanings] (DLT, I, 192; II, 20, III, 59) Sm. sig : 'prime, good' (Salonen,22) Tk. yig : 'yeg,iyi, daha iyi' (good) Sm. silig : (II2) 'Hand' (D.182) Tk. elig : 'el' (DLT, I, 72; 4KBI,145) (hand) Sm. sheg : 'rain' (Grd. 412) TK. yag : 'yagmak' (to rain, to precipitate) Sm. shir : 'singen und spielen' (MSL,III,150) Tk. yir : 'kosma, turku, hava,.." This word is still 'jir' in Balkar, Tatar, and 'ir' in kyrgyz. 'Shiir' is poem Turkish and related words mean poem in lots of Turk* languages. Sm. zag : 'Grenze' (MSL,III,85) 'border' (MSL,V,70) 'shoulder, outer edge, boundary, border' (Grd,432) Tk. yaka : 'taraf,yan,civar' (EUSz,280) Tuna missed "chek" which still means 'border' today in KB. Szekely (Chek-eli) in Hungary probably means "borderlands" and Czechs were probably named that way the same way Ukrain, and Krajina got their names. BTW, in Kipchak languages, 'yaka' would be more like 'jaga'. >VI. Sm. .Sh ~ Tk. .ch > >1. Sm. sag : 'small child' (MSL, III, 78) > Tk. chagha: 'yeni dogmush' [new born infant] (YTsz. 48) Sm. sag : 'schlagen' (D.175) Tk. chak : 'chakmak, vurmak' (EUSz,58) The word for nail, "chivi" in Turkish, and "chuy" in KB seem related. The word was probably more like "chuk" for nail (although it now means something else in slang). Sm. zibin : 'Insect' (D.120) Tk. chibin : 'sinek' ... (i.e. fly) >VIII. Sm. .u ~ Tk. .kV/a > >1. Sm. ubur : 'weibliche Brust' (MSL, III, 145; D. 102), 'teats' (Grd. 426; Falkenstein, 26) > Tk. koguz: 'goeghuez' [breast] (KBI, 274; EUSz, 114: DLT, I, 366) Sm. ud: 'day, time (in general' (Grd.425) ud 'day' (MLSb) the consonants after the initial vowel of the word: d, d, m, r, sh > >I. Sm. d/ ~ Tk. d/ > >1. Sm. adakur : 'ein opfergefass fuer Getraenke' (Sm. Lw.)(Akk. Hwb. 9) > Tk. adak : 'ichki kadehi' [wine glass] )Hs. S. 559; Nh. F. 370-8; Muh. 7) Sm. gid : 'l. entfernen' (D.60) Tk. id : 'salmak, gondermek.." Although in most Turk* languages this seems to be something like "jiber" in KB we still find this in the form "iy". Sm. ud : 'Zeit' (D.104) Tk. o"d : 'zaman, vakit..'(DLT,I,44...) (time) >II. Sm. d/,. ~ Tk. n/, . > >1. Sm. dugud : 'schwer' (MSL, III, 141) > Tk. yogun : 'kalin, yoghun' [thick, heavy] (KBI, 549; DLT, III, 29) Sm. kid(2): 'Sonne' (D.149) (il-SHAMASH..) TK. kun : 'gun, gunes,..' (sun, day) SM. mud : 'blood' (Grd. 389) Tk. kan : 'kan..' (blood) Sm. udu : 'sheep' (Grd. 427), udu, 'Schaf' (MSL,III,111) Tk. kong : 'koyun, (sheep) Sm. shid: 'number, voting board' (MSL,V,15); shiti 'Rechnung, 'Zahl' (MSL,III,144) Tk. san : 'sayi, sayma...' (to count) >III. Sm. VmV ~ Tk. VKV > >1. Sm. amash : 'Schafhuerde' (MSL, III, 145), 'Umfriedung, Stall' (D. 13) > Tk. agIl : 'aghIl, koyun yataghi' [sheepfold] (DLT, I, 65, 73) Sm. imma : 'two' (Emesal, (Falkenstein) Tk. ikki : SM. umush: 'discernment' (Grd, 428), 'Verstand' (D.108) Tk. ukush : 'anlayish; (understanding) >IV. Sm. r/ ~ Tk. z/ > >1. Sm. bur : 'to spread abroad, to disperse of (a thing)' (Grd. 336; MSL, III,140, 170) > 'undo' (especially a spell),' 'to make a hole' (MSL, III, 67), buru 'harvest (moun)" (Grd. 336) > Tk. buz-: 'bozmak, yIkmak' [to demolish, to ruin] (DLT, III, 8) Sm. gur(5) : 'zerbrechen, zerschneiden,abtrennen' (D.55) Tk. u"z: 'to break'... Sm. har :'dig,dig quickly' (Prince,176) Tk. kaz : dig Sm. mir: 'anger' (Emesal) (MSL,IV<35) Tk. kiz: (to anger) Sm. sur : 'to squeeze, to press out (oil, juice)' (Grd. 408) Tk. suz : 'suzmek: (to filter, to squeeze out) >V. Sm. sh/ ~ Tk. l/ > >1. Sm. ashsha : 'six' (Falkenstein, 41) > Tk. altI : 'altI' [six] (EUSz. 13) Sm. tush: 'seat, to sit' (MSL,III,58), 'to dwell (in a place) Tk. ol : 'bulunmak, kalmak..." (to be) Sm ush : 'dead, to die' (Grd.431) Tk. o"l : (die) DLT,IT,38) This word shows up in both versions in Karachay i.e. aush (to die). >VI. c. The word endings in ae, g, m, Vr/z > >I. Sm. ae. ~ Tk. An. > >1. Sm. mae : 'I' (Grd. 386) > Tk. men : 'ben' [I] (KBI, 309, DLT, I, 20) SM. zae : 'you' (sing) Tk. sen : >II. Sm. g. ~ Tk. ng. > >1. Sm. asha(g) : 'field' (Grd. 326) > Tk. alang : 'alan, duz ve achIk yer' [field, open space] (DLT, I, 135) Sm. dag : 'daybreak, morning, dawn' (D.43) Tk. tang : (daybreak, morning) Sm. kalag : 'to be strong, vigorous, have power' (Grd.349) Tk. kalIng: 'kalabalik, cok suru, kalin, kesif' >III. Sm. m. ~ Tk. K. > >1. Sm. alim : 'Steppentier, Widder' (D. 13) > Tk. elik : 'geyik' [deer] (ETY, Ir. 97, II, 90; ....) Sm. shurum : 'a cattle stable' (D.201) Tk. surug : 'suru' (herd) >IV. Sm. CVr/z. ~ Tk. Cr/chV. > >1. Sm. dingir : 'Gott' (D. 84), 'god' (Grd. 341) > Tk. tengri : 'TanrI' [god] (DLT, I, 53, 68) > 'goek, sema' [sky] (DLT, III, 377) Sm. dubur : 'Hode' (D.78) Tk. yumru : 'top gibi yuvarlak' (Mn.Gz.78, v5)..(round) >B. Sound equations which CAN be recognized easily: > >1. Sm. agar : 'lead (metal)' (EHG. 34, 58) > Tk. agIr : 'aghIr' [heavy] (DLT, I, 52) > >2. Sm. azgu : 'neck-stock (for use with animals)' (Grd. 331) > Tk. asgu : from as- 'asmak' [to hang] 'asgu' [hanger] (Dsz. 342) > 'asgI', 'asku', 'askI' > >3. Sm. bulug : 'Grenze, Grenzegebiet' (D. 31) > Tk. bulung: 'koshe, bucak, zaviye' [corner] (DLT, II, 371) > >4. Sm. di : 'to speak' (Grd. 342) > Tk. ti-: 'demek' [to say, to speak] (DLT, III, 231) Sm Tk Meaning esh es blow gim kipi/kimi like,as hum kom Lager, stall iduga yidig perfume kad kada knupfen kash kach speed, run, ki kI-L make, do ku ko werfen,lagen,grunden, koymak kur koru guard, watch,protect men men ich, I nammu neme how much, what sum sun geben, give te(ga) teg touch, attain tin tIn Leben, life, breath u u schlaf, uyku (sleep) umush yumush hizmet, vazife,werk ush us verstand,akil zag sag right side >Prof. Tuna compares about 165 Sumerian words with 149 Turkic, 7 >Mongolian, and 1 Akkadian words. > >Prof. Tuna discussed and argued with his thesis and examples with the >following scholars: > >1. 1971 in Harvard University (Boston): Erica Reiner, Guterbock, >Herbert Paper, Omeljan Pritsak, Denis Sinor. > >3. In 1974: with George Cardona & Henry Hoenigswald [bothIndo-Europeanists], >John Faught [American Native Languages], Ake Sjoberg [Sumerologist], >Earl Leichty [Akkadian], Omeljan Pritsak [Turcologist, Slavist]. > >In conclussion, Prof. Tuna argues that there is an historical link >between the Sumerian and the Turkic languages in the past, and whether >they are related or not is not relevant at this point. But, with >these 168 words, he thinks he has established a sound ground for this >"historical link." > >Any comments? > >Timur Now, the serious question of historical linguistics cannot be avoided. Everyone talks about 'regular sound change'. But not too many can answer the question of 'how many'. This data of Dr. Tuna (who is a linguist with a PhD from University of Pennsylvania) cannot be evaluated correctly until and unless we have an idea of how many cognates are required to establish "genetic relationship". Something this important, something that strikes deep into the very core of historical linguistics methodology should have shaken any normal scholar into action. PS. The Sumerian Dictionary from the University of Pennsylvania is by Ake Sjoberg. PPS. I can easily find 30-50% cognates between Turkic and the Sumerian list published by Halloran on the WWW, without even trying too hard. PPPS. If anyone wants more, I have a 20 page paper titled "Story of Life & Death, and Love & War" that treats related topics in historical linguistics via examples of sound changes in Turkic languages. Finally, there have been a few people who have done some mathematical calculations of how many words can be found to be cognates accidentally or how many pairs of cognates are needed to establish geneticity. The numbers are quite small, say 3-7 pairs (by people like Cowan, Bender, Greenberg). I have also made calculations of this type, much more thorough than these people and even tested it via a computer simulation. They can be found in my book, "Mathematical and Computational Linguistics", Mir Domu Tvoemu, Moscow, Russia or in electronic form from http://www.1stbooks.com. The chapter that deals with these computations can be found on my website for free. Enjoy your readings! Regards, Mark Hubey http://www.csam.montclair.edu/~hubey -------------- next part -------------- An embedded message was scrubbed... From: unknown sender Subject: no subject Date: no date Size: 11331 URL: From larryt at cogs.susx.ac.uk Tue Nov 10 23:23:43 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:23:43 EST Subject: The good Dr. Tuna In-Reply-To: <364381BA.3D0EC98D@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H. M. Hubey wrote: [on Tuna's knowledge of Sumerian] > 1. You mean you would not believe me if I said 2+2=4 because I am > not a mathematician? Nope. I mean only that I am unwilling to believe that something is true of Sumerian merely because Tuna has asserted it -- especially if he is not a specialist in Sumerian. > 2. There's a thing called "Argument from Authority". It is one of > the classic fallacies of logic. But I am not arguing from authority. I am merely pointing out that facts are facts, independently of what anybody, authoritative or not, thinks about them. But an authority, by definition, is far more likely to know what the facts are. In Dr. Tuna's case, his cited Sumerian word for `two' is either right or wrong. If it is not in accord with the facts, then it is wrong, no matter *who* he is quoting. > 3. What exactly is there to not being a Sumerian except that you > spent N years more on it than someone else? `Sumerianist', I presume. A Sumerianist is somebody with a detailed knowledge of the Sumerian language. Reading somebody else's book on Sumerian does not make you a Sumerianist, just as reading a book on how to play baseball does not make you a baseball player. > 5. The simple facts are that there are simple rules in linguistics and > it does not take longer than a few minutes to get the hang of the basic > idea, Really? You mean all of us have wasted our time in spending years learning our trade, when all we needed to do was to sign up for Doctor Hubey's Patented Ten-Minute Education in Linguistics? Gosh. OK. Here's a problem from my field. The four major regional variants of the Basque word for `ear' are as follows: So: what's the proto-form? And what "simple rule" should be invoked to discover it? > So far I have found no book on historical linguistics in which there > is a clear algorithm for reconstructing protoforms. That's because there isn't one. Reconstructing proto-forms is not the sort of thing can be reduced to an algorithm. Reconstruction requires a profound knowledge of the languages involved and a good understanding of language change -- at least. The process is not determinate, and it cannot be mechanized. Historical linguists can no more use algorithms to reconstruct languages than historians, archeologists and paleoanthropologists can invoke algorithms to reconstruct the bits of the past they're interested in. We all just have to dig out the available evidence, sweat over it, and then try to figure out what happened. > Some of the worst linguists are those who merely repeat what they > have managed to memorize. It's better to become a gardener than to > get a PhD in that manner. Nobody in the English-speaking world gets a PhD in linguistics by memorizing things. But it *is* a good idea to learn the facts, since external examiners have a tiresome habit of expecting the candidate to know the facts. > Let us also recall some of the things done by amateurs. Gauss, one of the > greatest mathematicians of all time, had to publish his own book on > Arithmetic and Group theory. You can find similar stories in every field. So you can. But you appear to be arguing as follows: *some* amateurs have been successful, therefore *all* amateurs are successful. But this is precisely the deeply flawed inductive reasoning which you have (wrongly) imputed to me. In my experience, most amateur linguists are incompetent, and some of them are downright crazy. If you thing I'm being unreasonable, then answer this question: would you be willing to fly on the maiden flight of an airliner designed by an untrained amateur designer and piloted by an untrained amateur pilot? Even if they've both spent "a few minutes" learning "a few simple rules"? [on the danger of relying on secondary sources] > That is pretty useless. Nobody can learn every language. True, of course, but this sad observation cannot change the fact that working on languages you don't know is dangerous. > and nobody can be a mathematician, physicist, engineer, computer > scientists, chemist, economist, etc. That is not how it works at > all. In fact, there is a thing called "specialization of labor" and > has been around for a long time. In fact, it is used by linguists > all the time, since they defer to specialists all the time, even to > the point of committing the "argument from authority" fallacy. Not guilty. Look. One of the non-existent "Basque" words cited as a comparandum by Chirikba in that article I mentioned is the alleged * `this same'. Why do I declare it to be non-existent? Because some authority has said so? No, not at all. I have never heard this word from any Basque-speaker. I have never read it in any Basque text. I have never encountered it in any specialist work on Basque. I cannot find it in any of my numerous Basque dictionaries, some scholarly, others popular. My Basque-speaking friends do not know it, and my specialist colleagues have never heard of it. I am therefore arguing, not from authority, but from evidence. [on learning Sumerian] > What exactly is there that is going to take 12 years to learn? Some symbols > arranged in some order with some presumed meanings which can be gleaned > from multilingual transcriptions. What is there to gain? Mr. Hubey, are you suggesting that one need not spend years studying Sumerian in order to know Sumerian? Time for a reality check, I think. ;-) > Surely, there are bad linguists as well as bad engineers. And there is > bad linguistics as well as good. So what? How do you propose to fix > this problem? It is beyond my power to fix it. All I can do is to point out errors when I encounter them -- which, in my case, is pretty damn often. > Let me guess. You want to have a stamp by the International > Linguistics Association and you want to stamp every book that even > touches upon linguistics as bad and good. Is that a solution? Well, > the arguments often made by linguists, (like a few already made > here) are essentially informal versions of this. And they stink as > badly as the formal version would stink. Mr. Hubey, I fear that you do not know very much about linguistics. Perhaps that ten minutes wasn't quite long enough after all. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Tue Nov 10 23:24:16 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:24:16 EST Subject: The good Dr. Tuna Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > > Nope. I mean only that I am unwilling to believe that something is true > of Sumerian merely because Tuna has asserted it -- especially if he is > not a specialist in Sumerian. > But I am not arguing from authority. I am merely pointing out that > facts are facts, independently of what anybody, authoritative or not, > thinks about them. But an authority, by definition, is far more likely > to know what the facts are. How are both true? Tuna only points out Sumerian words as found in scholarly journals, which presumably are written by Sumerologists or at least read and reviewed by Sumerologists. Tuna merely lists the Sumerian words with their meanings as told by Sumerologists. But now you are claiming that you will not believe Tuna if he is not a Sumer specialist which means that you will only believe authorities in Sumerian. To me it sounds like you believe that Tuna has falsified evidence, and also believe that only Sumerologists should make use of Sumerian words in any journal or book. > > 5. The simple facts are that there are simple rules in linguistics and > > it does not take longer than a few minutes to get the hang of the basic > > idea, > > Really? You mean all of us have wasted our time in spending years > learning our trade, when all we needed to do was to sign up for Doctor > Hubey's Patented Ten-Minute Education in Linguistics? Gosh. Why don't you post some complex rules of historical linguistics and let us see how long it takes to comprehend. > > OK. Here's a problem from my field. The four major regional variants > of the Basque word for `ear' are as follows: > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > discover it? I think you should try 1) reading what is written instead of what you think is written 2) explaining what I asked you last year on another list On the other list and other lists when I ask "experts" to explain what rules are used to construct protoforms and why they can't be found in textbooks, I notice that there is a lot of hemming and hawing. It sounds like Truman's refrain about economist; "I wish I had some one-handed economists". When I ask for rules on constructing protoforms (i.e. algorithms) there is no answer. When I ask why the field is soft and fuzzy, people like you get insulted and shout that it is a real science. Either there are rules for constructing protoforms or there aren't. My original question is/was why there is no algorithm for producing protoforms. Either what you practice is a science or it is not. If it is magic you don't have to explain it. If it is science it should be possible to see it in writing in some book. As for the "regular sound change rule" as applied to Basque ear, it is plainly possible to see that the algorithm is a "partial Caesar cipher". A Caesar cipher (the easiest cipher to crack) consists of changing every letter (sound) to another one by shifting them by an integer n mod M. In the case of linguistics it is "partial" because only some sounds get changed. In the case above if we make the equivalence l=h=g=* we can write the set as {be*arri,biarri}. Now we equate *=0=# and e=i=@ and obtain b@#rri, where @=ei and #=l=h=g=0. Now we go back to the original problem of language universals, and whether h>g is more common than g > h etc. This is why I asked the question of how protoforms are constructed in the first place. How, if it is not clear which is more common (e.g. h>g or g>h) then do linguists (i) choose one, AND, (ii) at the same time claim that it is a science. > In my experience, most amateur linguists are incompetent, and some of > them are downright crazy. I will refrain from telling you what my experience with (some) linguists is :-) However I could easily tell everyone what I know of your competencies as displayed already on other lists, but I will also let that pass and let you demonstrate it on your own the same way you demonstrated it on other lists (re: context-free grammars vs contex-sensitive grammars, definition of language, etc). All you have to do is answer the question which you have dodged for over a year. I will repeat it: How, if it is not clear which is more common (e.g. h>g or g>h) then do linguists (i) choose one, AND, (ii) at the same time claim that it is a science. Of course, I already know the answer; just pleases me to watch you trip over yourself :-) or try to evade it, (again, and again). > > True, of course, but this sad observation cannot change the fact that > working on languages you don't know is dangerous. It's even more dangerous to take guesses on topics which should be answered and answerable by mathematics. > Mr. Hubey, are you suggesting that one need not spend years studying > Sumerian in order to know Sumerian? In order to produce a list of cognates all you need is a dictionary. There are other things that will help, but since no linguist seems to do this, it does not matter. > It is beyond my power to fix it. All I can do is to point out errors > when I encounter them -- which, in my case, is pretty damn often. Pointing out an error is not the same thing as trying induction from one single example and pretending that you seem to have discovered a law. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mghiselin at casmail.calacademy.org Tue Nov 10 23:24:47 1998 From: mghiselin at casmail.calacademy.org (Michael Ghiselin) Date: Tue, 10 Nov 1998 18:24:47 EST Subject: Cladistic language concepts Message-ID: ----------------------------Original message---------------------------- Dear Dr. Dyen: Yes, we can separate different parts of the genome according to how rapidly they change. The DNA that codes for RNA changes relatively slowly though it depends on the part of the RNA molecule. It is thought that the most conservative parts of the RNA molecule are areas where a change would abolish function: and since the function is making proteins that would be a lethal. Mitochondrial RNA evolves faster than nuclear RNA and the reason seems to be that there is less of that kind of constraint. I am not sure what you mean by rates of branching. If there is rapid speciation the probability that there will be evidence of this in the data is greater if the molecule in question is evolving rapidly. But remember that branching rate is a different thing from the rate of substitution. It would be wonderful if all substitutions took place at a constant rate, for then we would have a "molecular clock" but we don't Sincerely, Michael Ghiselin From hubeyh at montclair.edu Tue Nov 10 23:25:02 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:25:02 EST Subject: [Fwd: [Fwd: Wake up call for the semester :-)]] Message-ID: ----------------------------Original message---------------------------- How do historical linguists evaluate these? > H.M.Hubey wrote: > > > > Other lists are doing it, why not us :-) > > > > How about these: > > > > 1. Meroitic -k, Barea -ge: Fenno-Urgic -k 'to' (e.g. Ingrelian ala-k > > 'down'. > > > > 2. Meroitic -te, Nubian -do locative suffix 'in': Old Turkish -ta, -da > > 'in' Finnish -ta 'in' > > > > 3. Meroitic -k feminine suffix: Mongolian -k-chin feminine of > > adjectives; Meroitic kdi 'woman': Turkish kari 'woman' (correspondence > > d:r looks better than d:ss but to make the matter even more surprising, > > there is one Eastern Turkish language, where the word for woman is > > kissi!) > > > > 4. Meroitic t demonstrative, Nubian ter 'he' etc. Mongolian tere 'he' > > 'that', Finnish 'te' 'this one' (I used te instead of ta-unlaut) > > > > 5. Old Nubian -ka accusative suffix: Old Turkish -g, -ig, Mongolian -g, > > -gi accusative suffix. > > > > 6. Old Nubian -ka dative suffix: Old Turkish -qa, -ke dative suffix > > > > 7. Old Nubian -n(a) genitive suffix: Mongolian -in, -n, Fenno-Ugric -n > > genetive suffix. > > > > 8. Old Nubian -r 'intentive' verbal suffix; Old Turkish -r, Finno-Ugric > > -r factitive verbal suffix. > > > > 9. Meroitic tar 'give' causative verbal affix (according to Dr. Priese) > > Old Nubian tir 'give' causative verbal affix: Old Turkish -tur 'give' > > causative verbal affix > > > > 10. Old Nubian -a participle, conjunctive converb: Old Turkish -a > > conjunctive converb > > > > 11. Old Nubian -ra predicative converb: Mongolian -ra final converb > > > > 12. Old Nubian -sa verbal participle praeteriti: Mongolian -san > > participle praeteriti > > > > 13. Old Nubian -s verbal suffix, praeteritum: Fenno-Ugric -s verbal > > suffix, praeteritum (cf. Old Nubian ki-s-in 'you came' with Wogulian > > min-s-en 'you came') > > > > 14. Old Nubian -men (-m-en) negation of verbs: Old Turkish -ma negation > > of verbs > > > > 15. Old Nubian -in,-en verbal suffix, 'you' 2 sg: Wogulian -en verbal > > suffix 'you' 2. sg. > > > > 16. Old Nubian possessive pronoun=genetive of personal pronoun (ir > > 'you', in-na 'your'): Old Turkish the same ('sen' 'you', san-ing > > 'your'), Mongolian the same (chi 'you', chinu 'your') > > > > 17. Old Nubian -t, -it deverbal nouns: Old Turkish -t, -it,-id deverbal > > nouns > > > > 18. Old Nubian -ki deverbal nouns: Turkish -ki abstract nouns, > > Finno-Ugric -k deverbal nouns > > > > 19. Old Nubian min 'what', Mongolian men 'what': Wogulian men 'what', > > Hungarian mi 'what'; > > > > 20. Old Nubian -guria 'because of': Turkish -gore 'because of' > > > > This connects Eastern-Sudanic (Old Nubian) with Uralo-Altaic. > > > > OK. I spill the beans: these are from Fritz Hintze's article. That is > > the reason for the strange spellings. Anyway, I think family > > relationships based on less than this have already been proposed and > > accepted in some cases in the Americas and in Africa. Is that right? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From tonybreed at juno.com Tue Nov 10 23:25:59 1998 From: tonybreed at juno.com (D. Anthony Tschetter-Breed) Date: Tue, 10 Nov 1998 18:25:59 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- Forgive me for being anectdotal, but when topic comes up of sound changes from X to Y and back to X, I always think of the English: askan > ask > ax (in some dialects, for example Black American English) One could propose that this is not an example of X>Y>X either on the grounds that the "ax" form is not an innovation but rather a retention of an earlier form, or perhaps that it's not systematic. I don't know. Any thoughts? -Tony ___________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866] From larryt at cogs.susx.ac.uk Tue Nov 10 23:26:22 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:26:22 EST Subject: Doing historical linguistics (part 2) In-Reply-To: <3644B37B.BC3CE929@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Sat, 7 Nov 1998, H.M.Hubey wrote: [LT] > > OK. Here's a problem from my field. The four major regional variants > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > discover it? > I think you should try > 1) reading what is written instead of what you think is written > 2) explaining what I asked you last year on another list > On the other list and other lists when I ask "experts" to explain > what rules are used to construct protoforms and why they can't be > found in textbooks, I notice that there is a lot of hemming and > hawing. That's because there are no "rules", in the mechanical or algorithmic sense of the term. Performing good reconstructions requires both knowledge of the languages and experience of the craft. Asking a linguist how to do reconstruction is not like asking a mathematician how to solve a differential equation. It's more like asking a professional cyclist how to ride a bike. (Not a good analogy, but the best I can do off the top of my head.) Your experts are hemming and hawing because they can't find any simple way of explaining the procedure to a novice, not because they don't know how to do it. > It sounds like Truman's refrain about economist; "I wish I had > some one-handed economists". When I ask for rules on constructing > protoforms (i.e. algorithms) there is no answer. When I ask why > the field is soft and fuzzy, people like you get insulted and shout > that it is a real science. No, not insulted, just exasperated. And anyway you've already had your answer: there are *no* algorithms for constructing proto-forms, and there cannot be. > Either there are rules for constructing protoforms or there aren't. Oh, there are certainly rules, but there are no algorithms. For example, given the observed variant forms of the Basque word for `ear', there is only one reconstruction that obeys all the rules. > My original question is/was why there is no algorithm for producing > protoforms. Either what you practice is a science or it is not. If > it is magic you don't have to explain it. If it is science it should > be possible to see it in writing in some book. No; this doesn't follow. In chemistry, for example, there are no rules for discovering new classes of compounds, but chemists frequently discover new classes of compounds nonetheless, and they are pleased to call this activity "science". Scientific activity is not, in general, algorithmic in nature. Algorithms, in fact, are in most cases only useful for obtaining answers to practical problems, and not in obtaining new knowledge. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 10 23:26:41 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:26:41 EST Subject: Doing historical linguistics (part 3) In-Reply-To: <3644B37B.BC3CE929@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Sat, 7 Nov 1998, H.M.Hubey wrote: [LT] > > OK. Here's a problem from my field. The four major regional variants > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > discover it? > As for the "regular sound change rule" as applied to Basque ear, > it is plainly possible to see that the algorithm is a "partial > Caesar cipher". A Caesar cipher (the easiest cipher to crack) consists > of changing every letter (sound) to another one by shifting them > by an integer n mod M. In the case of linguistics it is "partial" > because only some sounds get changed. > In the case above if we make the equivalence l=h=g=* we can write the > set > as {be*arri,biarri}. Now we equate *=0=# and e=i=@ and obtain b@#rri, > where @=ei and #=l=h=g=0. > Now we go back to the original problem of language universals, and > whether h>g is more common than g > h etc. This is why I asked the > question of how protoforms are constructed in the first place. > How, if it is not clear which is more common (e.g. h>g or g>h) then > do linguists (i) choose one, AND, (ii) at the same time claim that > it is a science. Neither h > g nor g > h strikes me as a very plausible change, at least not as a single step. If we allow these to be multi-step changes, then I would expect g > h to be more frequent than the reverse. However, none of this is relevant, because your partial Caesar cipher is of no obvious relevance to my question. Not one of the four attested forms is a plausible ancestor for the other three, and the original form must have been different from all of them. In fact, our preferred reconstruction is *, which accounts perfectly for all four of the recorded variants. But, of course, the probable correctness of this reconstruction is only obvious if you are intimately acquainted with the facts of Basque. Without such knowledge, * looks no more plausible than, say, * or * -- both of which are quite impossible. > I will refrain from telling you what my experience with (some) > linguists is :-) However I could easily tell everyone what I > know of your competencies as displayed already on other lists, but > I will also let that pass and let you demonstrate it on your own > the same way you demonstrated it on other lists (re: context-free > grammars vs contex-sensitive grammars, definition of language, etc). Oooooh, Mr. Hubey. [Note of explanation: Some time ago, on another list, Mark asked me to explain what a context-free grammar was. I did my best, but apparently I was unsuccessful.] > All you have to do is answer the question which you have dodged for > over a year. I will repeat it: > > How, if it is not clear which is more common (e.g. h>g or g>h) then > do linguists (i) choose one, AND, (ii) at the same time claim that > it is a science. Mr. Hubey, I'm afraid this question is unintelligible. It's rather like asking "What is the commonest word for `dog' on the planet?" and expecting the answer to be of some relevance to doing historical linguistics. Anyway, this business of "choosing" between h > g and g > h is something you have invented for yourself. We don't do historical linguistics by choosing from a Chinese-restaurant menu. [LT] > > Mr. Hubey, are you suggesting that one need not spend years studying > > Sumerian in order to know Sumerian? > In order to produce a list of cognates all you need is a dictionary. No, Mr. Hubey, a thousand times no. This is a common misconception among non-linguists, but it is grossly false. To produce a list of cognates you need far more than a dictionary. To start with, you need to know what you're doing. Trying to compile lists of cognates with nothing but dictionaries -- a common but forlorn activity -- is about as sensible as trying to construct a theory of music by counting the occurrences of all the notes in the entire repertoire of operas. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Tue Nov 10 23:27:04 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:27:04 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Sat, 7 Nov 1998, H.M.Hubey wrote: > My Sumerianist colleagues assure me that there exist a number of > (especially older) works on Sumerian which are outdated and which > contain errors of fact and interpretation, and they complain that > various attempts at seeking relatives for Sumerian have relied upon > these defective and unreliable sources. Yes, on the one hand, no science is immune to error, and on the other hand it is unlikely that pages and pages of materials can be "mostly wrong" and still convey more or less the correct message. If there are going to be changes it will probably be with finer shades of meaning. If I now write a book of cognates from Halloran's list, then some of your "Sumerianist" colleagues will again change them. The chances are, reasoning probabilistically, that small changes will create small effects and large scale changes will void past Sumerian works. Since when does a field get a thorough overhaul? Do you think Hittite needs a thorough overhaul? > > So, the question is not whether Tuna has taken his forms from published > sources, but rather how reliable those sources are. And the only people > who are in a position to judge the reliability of Tuna's sources are > those who have a good specialist knowledge of the language. > > Moreover, even when a Sumerian word is cited correctly, it does not > follow that it is a valid comparandum in comparative enterprises. Many > Sumerian words are attested only in late texts, and not in early ones, > and hence cannot be used as comparanda. (This I know is true.) IT does not matter. Because the Sumer-Turk cognates come from a "proto-Euphratic" substrate and these people happened to be around that neighborhood a couple of thousand years BC. That means that "Turk" (whatever it is) peoples were not living in the forests of the Altai-Siberia region. There is too much evidence against it and Turko-Sumer cognates are just a small part of the story. In other cases, a word may have a transparent etymology within Sumerian, > say as a compound or a derivative, and hence it too is unavailable as a > comparandum. (I don't know if this is the case in Sumerian, but it is > very commonly the case with other languages invoked in remote > comparisons.) In still other instances, a word may have changed its > form or meaning substantially during the historical period of the > language, in which case it is only the earliest form and meaning which > is available for comparison, and not the later ones. So what? In NJ among the hispanics "mines" has replaced "mine" because of hypercorrection. That is one mechanism of regularization and things like that happen all the time. What difference does that make except give ammunition to those who want to make Sumerian fall from the sky? > In short, merely finding a Sumerian word cited in the specialist > literature, even if that literature is reliable (which it may not be), > does not automatically permit you to cite that word as a comparandum in > seeking genetic links. Sure it does. It is roots that count, and the root is presumably in many words that have that root and are derived from it. Where did you find these rules? > A few examples from Basque, since I don't know Sumerian. Among the > Basque words frequently cited as comparanda in seeking genetic links are > the following. > > `dry'. This looks very similar to something in Caucasian and > even more similar to something in Berber. But the earliest sense of the > word is `barren, sterile', and it has developed semantically, via > `dried-up' (of a well or a spring) to `dry'. So you can't have the > modern sense. So what? Many words shift meaning over time. Do you think humans 30,000 years ago had 100,000 words in their lexicons? How about 100,000 years ago? Was there ever a time when humans had maybe 1000 words? How can you be sure that none of these roots has survived (albeit in semantically shifted form)? You can't. That is so simple, it makes me wonder about the motives of people who become so irrational. What are you afraid of losing? > `woman'. The second half looks like words for `woman' or > `wife' in several other languages, and eager comparativists have > brandished this delightedly while throwing away the first half as > meaningless junk. But the earlier sense of this word was `girl', and it > is transparently bimorphemic, from , the regular combining form of Wow. YOu mean there is no semantic connection between girl, woman and female? Do the hipsters and hippies who used the word "bitch" for women make a gigantic error, and there is no connection between female dog and female person? > `female' (which itself is borrowed from Occitan), plus <-kume> And where is that borrowed from? em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), emesal (female speech in Sumerian), emcek (breasts, udder), meme (breast), emzirik, etc etc. Do you understand the probabilistic implication of such patterns? > `offspring, child'. So you can't have the word at all, and you > certainly can't assign the `female' element to the second half. Are you trying to convince me that historical linguistics works only with "exact" meanings and "exact" phonological/phonetic/sound shapes of words? Are you kidding? > `burdock', `limpet'. This looks strikingly like something in > Caucasian. But it is a transparent (and fully regular) borrowing of > Latin `burdock' and its Spanish descendant `limpet'. So > you can't have this one either. Accidents happen all the time. That is the whole purpose of learning and understanding probability theory and statistics. AT least then you don't have to repeat yourself over and over trying to prove general statements by giving 2-3 examples. Did you read what I wrote about the farmer who fed his duck for 33 days? How much value do these examples have? None for me, because I got the general idea years ago after 2-3 examples. I already know that generalities cannot be proven or assumed because of 2-3 examples. I also know how probability theory works, and I can recognize those who are ignorant of it and hence constantly make the same mistake over and over and over again. I also have to tell you that there is no such thing as "proof by repetetion". If you want to pile up your examples and then try statistics please be my guest, go right ahead. > All genuine Basque words, all found in any halfway decent dictionary, > but all totally unavailable as comparanda in seeking relatives for > Basque. People who cite such data as comparanda are wasting their time. Hardly. It is more likely that you are wasting your time trying to convince me that I should take your 2-3 examples as proof and substitute "proof by repetetion" and "proof by example" for logic, probability theory and fuzzy set theory, and the rest of math. > So, I'm not accusing Tuna of falsifying data. But, if he doesn't have a > profound knowledge of Sumerian, and is merely extracting words > incomprehendingly from other people's work, then he is running a very > grave risk of obtaining meaningless results -- just as many others have > done in extracting Basque words from dictionaries without having any > knowledge of the language. Fortunately, most people have common sense and that eventually will drive them to accept conclusions derived from probability theory and logic, and the rest of math because they know that the world around them was built by people who use all the tools in this bag of cheap tricks called math. What is funny is why when you extoll the virtues of correlation-regression analysis (Labov?) to high heaven in your book, you seem to have turned against math on this list? Is that because other linguists extolled Labov and you decided you had to do it? OR is it because you can understand CR analysis but don't understand probability theory? Or is it because you think Labov was a linguist and I am not? OR is it because you think Sumerian is too valuable to allow to be related even in the remotest sense to barbarians when truly civilized Aryans are hanging around? For what reason do you find it necessary to stick to this irrational manner of discourse? > > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:27:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:27:25 EST Subject: Doing historical linguistics (part 2) Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Sat, 7 Nov 1998, H.M.Hubey wrote: > > [LT] > > > > OK. Here's a problem from my field. The four major regional variants > > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > > discover it? > > > I think you should try > > > 1) reading what is written instead of what you think is written > > 2) explaining what I asked you last year on another list > > > On the other list and other lists when I ask "experts" to explain > > what rules are used to construct protoforms and why they can't be > > found in textbooks, I notice that there is a lot of hemming and > > hawing. > > That's because there are no "rules", in the mechanical or algorithmic > sense of the term. Performing good reconstructions requires both > knowledge of the languages and experience of the craft. Ehhh. "craft"?????????????? You mean it is not a science? >Asking a > linguist how to do reconstruction is not like asking a mathematician how > to solve a differential equation. It's more like asking a professional No it isn't. A mathematician can tell you how to solve it. > cyclist how to ride a bike. (Not a good analogy, but the best I can do > off the top of my head.) Your experts are hemming and hawing because I know how to ride bikes and I can tell you how to do it. > they can't find any simple way of explaining the procedure to a novice, > not because they don't know how to do it. Then there must at least be some "fuzzy rules" but I forgot you don't like that either. Maybe there are tendencies and propensities and they can be analyzed using Markov process concepts and you don't like that either. Nevertheless you are sure they can't be used despite not knowing what they are and how they can be used. Is this some special form of logic available only to historical linguists or only to you? Despite all of these problems, nevertheless, you are sure it is a real science deserving of the Nobel prize and does not need any improvement whatsoever because it has already reached the pinnacle of what science can be? Is this the logic and conclusion you expect me to accept? > > It sounds like Truman's refrain about economist; "I wish I had > > some one-handed economists". When I ask for rules on constructing > > protoforms (i.e. algorithms) there is no answer. When I ask why > > the field is soft and fuzzy, people like you get insulted and shout > > that it is a real science. > > No, not insulted, just exasperated. And anyway you've already had your > answer: there are *no* algorithms for constructing proto-forms, and > there cannot be. That again, is the wrong conclusion. What you want to say is this: "I do not know how to construct protoforms and I do not know anyone else who does." Certainly you cannot know whether or not there can ever be an algorithm and you cannot know in what manner they might be constructed. > > Either there are rules for constructing protoforms or there aren't. > > Oh, there are certainly rules, but there are no algorithms. For > example, given the observed variant forms of the Basque word for `ear', > there is only one reconstruction that obeys all the rules. OK. PRoduce the rules, and I will create the algorithms. > > My original question is/was why there is no algorithm for producing > > protoforms. Either what you practice is a science or it is not. If > > it is magic you don't have to explain it. If it is science it should > > be possible to see it in writing in some book. > > No; this doesn't follow. In chemistry, for example, there are no rules > for discovering new classes of compounds, but chemists frequently > discover new classes of compounds nonetheless, and they are pleased to > call this activity "science". I don't believe that. Once again, you are confusing yourself for the scientific community. When someone says "I don't know" it could mean "I don't know but others might know." "I don't know but probably nobody else knows either." "I know that nobody knows." .... > Scientific activity is not, in general, algorithmic in nature. A lot of people would disagree. For example, the great philosopher Popper, or Lakatos. > Algorithms, in fact, are in most cases only useful for obtaining answers > to practical problems, and not in obtaining new knowledge. YOu might be partially right here, but only that; "might" and "partial". But don't give up hope yet. There is a thing called "data mining" and "knowledge discovery science". We will lick that problem too :-) Why do you think there are conferences in which people try to create "artificial scientists" and "artificial intelligence". > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From bwald at HUMnet.UCLA.EDU Tue Nov 10 23:28:09 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Tue, 10 Nov 1998 18:28:09 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- In passing, I noticed Alan King acknowledge a correction by Miguel Carrasquer. Alan wrote: >Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact >that besides Castilian, both Galician and Romanian are also exempt from >this general Romance development. In Castilian and Galician (but NOT in >Romanian, as you point out), there is no /z/ phoneme. I was confusing >these two things. I am curious about this situation because I have long had the impression that what is historically distinctive of Castillian and the Spanish-area around it, but not extending to Portuguese, Catalan or other adjacent Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among them. That is, I never thought that the ancestor of Castiliian, Spanish, whatever, was exempt from the EARLIER Romance process of intervocalic voicing of Latin -s- (among other sounds), but that by LATER developments it devoiced the resulting -z- (in most environments -- in general,in effect). Similarly, /zh/, as in ancestral "*g*ente" 'people' (current "*h*ente"), was devoiced, something like /zh/ > /sh=c,/ (ichlaut) (> h), cf. ojala (o*h=x*ala:) < Arabic in*SH*alla:h 'God willing'. As far as I know, such fricative devoicing is a distinctive feature of the divergence of Spanish from other Romance languages, and Spanish should be celebrated as a language which exemplifies historical devoicing of (certain) fricatives (the ones made with the tongue against the palate). P.S. Interesting in Spanish is the tenacity of the devoicing, so that rather than revoice intervocalically, /s/ has a tendency (found elsewhere) to reduce to "pure" devoicing, i.e., /h/ (and ultimately loss of the segment). Though /s/ > /h/ also occurs before consonants, the tenacity is reminiscent of the previous movement of /sh/ to the back, to /x/ or /h/, rather than toward "revoicing" in any environment. (Unlike /s/, /sh/ only occurred before vowels. /s/ could easily be neutral or even associated with the phoneme /sh/ before consonants, cf. Portuguese, and, indeed, the numerous studies of s > h variable dialects of Spanish find that s > h is more favored to occur before a consonant than a vowel, cf. the history of French /s/+consonant.) In a separate message, Robert Ratcliffe wrote: >But in general I do not see any reason to assume >that no sound change is reversible given enough time. It should be made clear that the issue is directionality of sound change. The issue of "reversibility" usually implies that the set of words affected by the first sound change remain distinct at the time of the second sound change. However, the "given enough time" qualification suggests that preservation of the identity of the class of words is not at issue (since other changes will most often obscure the original word class affected). The issue is only that a second sound change HAPPENS to be the reverse of the first one -- but historically has nothing to do with the first change. According to my understanding, Latin s > z and then Spanish z > s illustrates this point. From R.J.Penny at qmw.ac.uk Tue Nov 10 23:28:26 1998 From: R.J.Penny at qmw.ac.uk (Ralph Penny) Date: Tue, 10 Nov 1998 18:28:26 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- Rhotacization of /s/ to [r] before /d/ commonly occurs in Cantabria, at least in the rural varieties I investigated in _El habla pasiega_ and _Estudio estructural del habla de Tudanca_. Ralph Penny > ----------------------------Original message---------------------------- > > > > > > The change of s to r before another consonanant is widespread in southern > > Spain and also in Madrid, so that los dedos may be articulated as lor > > dedoh and buenos dias as buenor diah. > > it is nice to read it, because I never heard a change like "buenor > diah" in Madrid, nor in Toledo, Galicia nor in Southern Spain. You say > "widespread", but I never, *really never*, met it. > > You can hear following (sub)standard variants: > > in Madrid > 1) buenos dias > 2) bueno dia > in the South (from Toledo toward andalucia) > 3) buenoeh diah > 4) buenoh diah > > a Galician should use the standard variation, also "buenos dias", but > "bos dias" (in Galician). > > > Have you ever heard "buenor diah" ? Where do you get such informations? > I would like to read about it. > > j.m. > Ralph Penny School of Modern Languages Department of Hispanic Studies Queen Mary and Westfield College Mile End Road London E1 4NS. Tel: +44 171 775 3139 Fax: +44 181 980 5400 From fcosw5 at mail.scu.edu.tw Tue Nov 10 23:33:13 1998 From: fcosw5 at mail.scu.edu.tw (Steven Schaufele) Date: Tue, 10 Nov 1998 18:33:13 EST Subject: Ket-Na-Dene affiliation? Message-ID: ----------------------------Original message---------------------------- Dear Colleagues, My wife has just forwarded to me a Reuters story (posted on www.cnn.com) about an article by Merritt Ruhlen appearing this week in the Proceedings of the National Academy of Sciences, claiming an affiliation between Ket, the sole surviving Yeniseian language, and the Na-Dene family of North America. The Reuters story begins with the (to a historical/comparative linguist) ominous statement, `A few words in one of the planet's most obscure languages support the theory that Native Americans left Asia in several separate migrations', and goes on to note that Ruhlen's paper `gives examples of 36 words that are similar in the two language families, including the words for birch bark, children and rabbit'. (However, the only actual lexemes mentioned in the Reuters story are the Ket word for `birch bark' -- it's simply asserted that `in several existing Na-Dene languages it is pronounced similarly' -- and the Ket and Koyukon words for `breast'.) The Reuters story goes on to reassure us that the whole hypothesis isn't really based only on 36 lexemes: `Ruhlen found enough other similarities to convince him of the link. "I just picked ouit 36 for this article that looked like the best and most obvious and strongest," he said.' I'm wanting to know, does anybody subscribing to this EBB know anything about this? Is there anything to this proposed affiliation? And if not, is anybody doing anything about clarifying the issue for the general public? I note with great trepidation that the Reuters article trots out mouth-watering statements like `Related words are often easy to spot -- for instance the German word "mutter" is similar to its English counterpart "mother", while the Russian word "brat" looks very much like "brother" and is similar to the Latin root for words like "fraternal".', but completely ignores any mention of such nasty little things as false friends, etc. Considering the trouble i go to in my introductory courses explaining how unreliable this kind of argument is, i'm worried about a story like this in the general press. Best, Steven -- Steven Schaufele, Ph.D., Asst. Prof. of Linguistics, English Department Soochow University, Waishuanghsi Campus, Taipei 11102, Taiwan, ROC (886)(02)2881-9471 ext. 6504 fcosw5 at mail.scu.edu.tw http://www.prairienet.org/~fcosws/homepage.html ***O syntagmata linguarum liberemini humanarum!*** ***Nihil vestris privari nisi obicibus potestis!*** From DISTERH at UNIVSCVM.SC.EDU Tue Nov 10 23:40:23 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Tue, 10 Nov 1998 18:40:23 EST Subject: HISTLING is back Message-ID: Dear Colleagues, As you have doubtless noticed, HISTLING is back from her weekend away from the computer. I apologize for having inconvenienced your mailboxes with the 20 postings I've just sent out. Dorothy Disterheft From delancey at darkwing.uoregon.edu Wed Nov 11 19:37:15 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Wed, 11 Nov 1998 14:37:15 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- > My wife has just forwarded to me a Reuters story (posted on www.cnn.com) > about an article by Merritt Ruhlen appearing this week in the > Proceedings of the National Academy of Sciences, claiming an affiliation > between Ket, the sole surviving Yeniseian language, and the Na-Dene > family of North America. The Reuters story begins with the (to a This idea has been floating around for a while, among the usual suspects. The proposal of a Yeniseian-Sino-Tibetan link was first made quite a few years ago, I believe, though I don't know remember (if I ever knew) who was the original proponent. (I think this is something which Starostin has endorsed, but I'm sure the original proposal is much older than that). And Na-Dene-Sino-Tibetan goes back to Sapir, so it must be true. So, by transitivity, Yeniseian-Na-Dene must be true, too. Simple logic. Actually, though seeing Ruhlen's name attached to any proposal for genetic grouping always makes me suspicious, I'd be very interested to see these data. I spent a couple of days several years ago straining my almost nonexistent Russian and groaning through some Ket verb paradigms, and one of the striking things about the language is the existence of semantically obscure elements in the prefix string, apparently lexically empty preverbs of some sort, which indeed reminded me a lot of Athabaskan "thematic" prefixes. Not the sort of resemblance I'd be ready to hang a major proposal for trans-Pacific relationship on, but it's a rather unusual typological feature. > I'm wanting to know, does anybody subscribing to this EBB know anything > about this? Is there anything to this proposed affiliation? And if > not, is anybody doing anything about clarifying the issue for the > general public? Good question, but--does the general public really want to know better? Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Wed Nov 11 18:20:34 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Wed, 11 Nov 1998 13:20:34 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- On Tue, 10 Nov 1998, Steven Schaufele wrote: > My wife has just forwarded to me a Reuters story (posted on www.cnn.com) > about an article by Merritt Ruhlen appearing this week in the > Proceedings of the National Academy of Sciences, claiming an affiliation > between Ket, the sole surviving Yeniseian language, and the Na-Dene > family of North America. The Reuters story begins with the (to a > historical/comparative linguist) ominous statement, `A few words in one > of the planet's most obscure languages support the theory that Native > Americans left Asia in several separate migrations', and goes on to note > that Ruhlen's paper `gives examples of 36 words that are similar in the > two language families, including the words for birch bark, children and > rabbit'. (However, the only actual lexemes mentioned in the Reuters > story are the Ket word for `birch bark' -- it's simply asserted that `in > several existing Na-Dene languages it is pronounced similarly' -- and > the Ket and Koyukon words for `breast'.) The Reuters story goes on to > reassure us that the whole hypothesis isn't really based only on 36 > lexemes: `Ruhlen found enough other similarities to convince him of the > link. "I just picked ouit 36 for this article that looked like the best > and most obvious and strongest," he said.' > I'm wanting to know, does anybody subscribing to this EBB know anything > about this? Is there anything to this proposed affiliation? I haven't seen this particular article, but Ruhlen has in fact been pushing a link between Yeniseian (the family to which Ket belongs) and Na-Dene for some time. Take a look at chapter 4 of the following book: Merritt Ruhlen (1994), On the Origin of Languages, Stanford: Stanford University Press. (Do not confuse this with Ruhlen's 1994 book of nearly identical title published by Wiley.) This chapter, co-written with Sergei Starostin, proposes 300+ Proto-Yeniseian reconstructions, plus external comparisons with Na-Dene, Basque, Abkhaz-Adyghe, Nakh-Dagestan, PIE, Basque, Burushaski, and Nahali. All the items mentioned by Steven are there, though the details are perhaps different: for example, no Na-Dene comparandum is proposed for the Proto-Yeniseian word for `birch bark'. My own view is that the comparisons on offer are devoid of value. They consist of nothing more than the usual miscellaneous lookalikes. The Basque comparanda, I can testify, are characterized by the most awful collection of errors. Here are the particular items mentioned by Steven, with only Proto-Yeniseian (PY) and Na-Dene (ND) comparanda; diacritics are omitted: `birch back': PY *. No ND comparandum. `breast': PY *; Haida `heart', Tlingit `heart', Kutchin `breast', Tahltan `breast', Hare `breast', Mattole `breast'. `children': PY *; Haida `child', `son', Tlingit `child', `son', Eyak `child', `son', Navajo <[gamma]e?> `son'. `rabbit': PY * ~ *: Eyak `rabbit', Slave `rabbit', Tsetsaut `rabbit', Navajo `rabbit'. > And if not, is anybody doing anything about clarifying the issue for > the general public? This is not easy. Last year the Times of London published a solemn article about a particularly imbecilic book "proving" that Etruscan was Basque, and even added an editorial praising this "achievement". I wrote a letter to the Times drawing attention to the absurdity which they had perpetrated, but I never got so much as a reply, let alone anything in print. It appears that even seemingly serious publications prefer eye-catching drivel to sober assessment. > i'm worried about a story like this in the general press. Me too, but what can we do? Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Wed Nov 11 18:19:46 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 11 Nov 1998 13:19:46 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > > Aren't morphological paradigms also part of 'regular sound change"? > > No, they're not. > > But more to the point, contrary to what Mr Hubey seems to be > suggesting, the little couplet above is not an instance of circular > reasoning. It's merely the same statement phrased two different ways. > Or rather, 2. is a statement of a general principle, of which 1. is a specific application. > > What Mr Hubey may be trying to articulate is the superficially > circular-looking: > > 1. A,B,C are a language family ===> A,B,C have regular sound > correspondences. > > 2. A,B,C have regular sound correspondences ===> A,B,C are a language > family. > Perhaps only some people use it circularly, or many linguists use it circularly, but it is in use. > But the attribution of circularity rests on a misreading of the > relations between the propositions in 1. and 2. as the same. 1. is a > causal relation -- regular sound correspondences result from the > definition of a language family, plus the fact that sound change is > regular. But borrowings also create regular sound correspondences. 2. is a progression from evidence to inference -- we observe > regular sound correspondences, from which we conclude these languages > are a family. (Whether this is an accurate account of what we > actually do is not the question here.) It's no more circular than: > > 1. Patient has measles ===> patient has spots on face. > 2. Patient has spots on face ===> patient has measles. This rests on something different. 1. Patient has measles (definition comes from some other place, but may include spots on the face). These days the defn would come from being able to culture the bacteria. Then the spots on the face and measles correlate. The time depth is short and one can see a non-measles person get it, get sick, etc. Because of the correlation of measles and spots, 2 then becomes an implication. But that does not work so in historical linguistics because we never had a record of any language family (knowing its relatives, etc) but all of it rests on a larger theory of which regular sound correspondence must be a part. So it is a whole mess of correlations which lead towards that conclusion. More to the point it is based on this reasoning. 1. These languages have too many things in common. IOW, there are many words in these languages which can be made to look like each other with similar meanings and which could not be due to chance. 2. If that is not due to chance then either they got these words from each other or the words are all descended from a common language. 3. We have plenty of evidence (what?) that these languages did not get these words from each other. 4. Therefore these words in these languages must all come from an earlier common source. This is how it is supposed to work, but you can see rather easily how and where problems crop up, and where arguments occur. Isn't this basically right? > This may be rough and ready diagnostic practice, but it isn't a > logical fallacy. > > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From martinez at eucmos.sim.ucm.es Wed Nov 11 16:41:09 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 11 Nov 1998 11:41:09 EST Subject: Gr. #s- > #h- Message-ID: ----------------------------Original message---------------------------- > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting and of course a problem like Gr. hu^s / s^us "pig" From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:40:56 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:40:56 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- At the risk of getting caught in the crossfire, I'd like to interject myself in the debate between Larry Trask and Mark Hubey. I'm in agreement with about 90% of what LT has to say, and have to say (with no malice intended) that MH's postings reveal a great ignorance of the field of linguistics. Nonetheless I do think that MH is asking some legitimate questions and deserves a better response from the profession than the old refrain that the esoteric knowledge is only revealed to those who have joined the secret brotherhood. re the following sequence in particular: H. M. Hubey wrote: > So far I have found no book on historical linguistics > in which there is a clear algorithm for reconstructing protoforms. And > > i have asked many people about this. They don't know either. Sometimes > > they can justify it easily sometimes it is a guess and sometimes it is > no more than a concensus. MH again: > > On the other list and other lists when I ask "experts" to explain > > what rules are used to construct protoforms and why they can't be > > found in textbooks, I notice that there is a lot of hemming and > > hawing. LT's response: > That's because there are no "rules", in the mechanical or algorithmic > sense of the term. Performing good reconstructions requires both > knowledge of the languages and experience of the craft. Asking a > linguist how to do reconstruction is not like asking a mathematician > how > to solve a differential equation. It's more like asking a > professional > cyclist how to ride a bike. (Not a good analogy, but the best I can > do > off the top of my head.) Your experts are hemming and hawing because > they can't find any simple way of explaining the procedure to a > novice, > not because they don't know how to do it. My view: That "reconstruction" is (or even could be) a matter of algorithm, or rule is a widely held misunderstanding outside the field, no doubt supported by a misinterpretation of the technical term "reconstruction" in its ordinary sense. We cannot literally "reconstruct" anything. A "reconstruction" is a hypothesis, neither more nor less. Like any hypothesis in any field of science we arrive at it by guesswork, by intuition, by imagination, by accident. There is no path of deductive reasoning, no "discovery procedure", no algorithm which leads from the data to the hypothesis. The rigor in historical linguistics as in the natural sciences is not in the way in which hypotheses are reached but in the way in which the hypotheses are submitted to the test of the data. A good reconstruction should be testable in principle-- it should be a specific hypothetical prediction about the way a particular language was spoken by a particular group of people at a particular point in time. A reconstruction is virtually never (directly) testable in practice-- it is so only on those extremely rare occasions when new texts of ancient languages are unearthed. How do we test it then?-- By implication. Each reconstruction (of a proto-phoneme for example) has implications for the whole system of the proto language (the whole phonological system, eg), for the development path leading from the proto-language to the attested languages (the sequence of sound changes, eg), and for the forms of the reflexes in the descendant languages. Only the last is directly observable, of course, and only this real data can be used to rule out a proposed reconstruction absolutely. The proposed development path implied by the reconstruction cannot be tested as right or wrong but only as plausible or implausible based on the statistical frequency of observed patterns of change. Simlarly the implications which a reconstruction has for the whole system can only be tested under the assumption of the "uniformitarian principle" - the assumption that prehistoric languages were not fundamentally different in kind from attested languages and hence should not show structural anomalies of a type not found in attested languages (languages shouldn't be reconstructed with no vowels, or all verbs, for example). Probability and statistics is involved in all of this, and much could be gained from making it more explicit. Much could also gained if historical linguists made an effort to base assumptions about plausible directions of changes on an explicit, accesible, statistically analyzable body of evidence of changes, rather than on implicit personal knowledge. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From mcv at wxs.nl Wed Nov 11 16:39:30 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:39:30 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36463306.60B6E2F@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: >Larry Trask wrote: >> `female' (which itself is borrowed from Occitan), plus <-kume> > >And where is that borrowed from? > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), >emesal (female speech in Sumerian), emcek (breasts, udder), meme >(breast), emzirik, etc etc. This illustrates the fact that a little knowledge of the languages involved is never a bad thing. Larry should have been slightly more precise by indicating that Basque was not borrowed from vanilla Occitan, but from Gascon/Bearnais. Then it immediately becomes clear that the word is simply Latin FEMINA, which regularly becomes hemna (hemm@ ~ hemno) in Gascon (f > h). The ultimate root is IE *dheH1- "to suck" (*dheH1-mHn-oH2 > femina). The coincidence with PTurkic *eme "woman" is entirely coincidental. And so is, with even more reason, Sumerian where the element doesn't mean "woman" at all, but "language, tongue". The word means "thin, refined" (it used to be thought that was "woman", and one can still find that in older books, but in fact the Sumerian word was ). >Do you understand the probabilistic implication of such patterns? Yes. In short: "Coincidences happen" and "Garbage in, garbage out". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From Georg at home.ivm.de Wed Nov 11 16:39:11 1998 From: Georg at home.ivm.de (Ralf-Stefan Georg) Date: Wed, 11 Nov 1998 11:39:11 EST Subject: The good Dr. Tuna In-Reply-To: <364381BA.3D0EC98D@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- >Where are the books on Sumerian? > >Where are all the books on Hittite? I have a batch of them. It took >be 6 months to collect some and xerox the others. Does this mean that >I must not trust that the OI has done poor work and that I should not >rely "on others". What if I start to look at the original text? > >What exactly is there that is going to take 12 years to learn? A textbook, or well a handbook of, say, Hittite is but a snapshot of common knowledge about such a language at a very precise given point in time. Just as an example: I consider myself well-trained in Indo-European and did study Hittite (including cuneiform) for a while. Yet I'd not even think about citing a Hittite form which may be crucial for some argument of mine without consulting my Hittitologist friends before, who work night and day with the actual texts. Hittite is language which is amply (re: the number of texts), but still quite fragmentarily (re: its vocabulary and morphology) attested. *Every day* during the summer campaigns new texts are unearthed, transcribed, translated and eventually published. Every year at least, some of them lead to sometimes considerable changes in our knowledge of the language. Words are re-assigned to different inflectional classes, meanings are fixed, even the *readings* of cuneiform signs are fixed sometimes even today, previously well-known words, which already made it into the dictionaries are unveiled as errors of previous investigators and so on, and so on. It is quite unlikely that things like that happen too often for languages like Danish, Mandarin Chinese or Guugu-Yimidhirr (one modern language, for which a constant reanalysis of synchronic data is however quite common is Ket), but for a language whose records are three millennia old this is quite common and should surprise nobody. The same holds for the other cuneiform languages, like Sumerian, Hurrian, Urartaean aso. In the case of Hattic, one monograph published last year manages to render almost the entirety of previous publications on this language obsolete (yet Hattic had been assigned to various language families before, too bad). This has nothing to do with the languages being so utterly difficult and unmanagable, the writing system is, the cultural setting of the texts is, and the fragmentary nature of the stone tablets after so many centuries is (there is actually *one* (1) entirely undamaged text of any interesting length in Hittite, the famous bronze-tablet contract, found in the eighties). Learning Hittite does *not* take 12 years, it takes your whole life. Making use of the present state-of-the-art (which may be different after two years) for linguistic purposes takes a phone-call to the experts (but having spent 12 years learning Hittite before that surely helps to understand what they are talking about ;-). St.G. Stefan Georg Heerstrasse 7 D-53111 Bonn FRG +49-228-69-13-32 From Georg at home.ivm.de Wed Nov 11 16:38:47 1998 From: Georg at home.ivm.de (Ralf-Stefan Georg) Date: Wed, 11 Nov 1998 11:38:47 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- >I'm wanting to know, does anybody subscribing to this EBB know anything >about this? The Yenisseyan-Na-Dene affiliation is subscribed to mainly by Sergei Starostin and his Moscow-based school of Nostraticists. It forms part of a much larger grouping, called Dene-Caucasian, comprising no less than the following languages and families: Basque, Iberian (yes, Iberian, though nobody knows anything about Iberian !), Burushaski, Yenisseyan, "North-Caucasian" (comprising NW- and NE-Caucasian which are not generally grouped together by Caucasianists), Hurro-Urartean (lumped with N-Cauc.), Kusunda (an extinct language of Nepal, on which little is known; Bengtson, another supporter of this grouping, recently informed me that Kusunda has been taken out of the grouping; I'm inclined to call this a step in the right direction ;-) and Na-Dene (I'm currently unsure whether further North American lgs. have made it into the family yet), and, sorry I forgot, the whole of Sino-Tibetan, and of course Sumerian. It is clear that one of the goals of this grouping is to hoover up most languages of the Old World, which are currently thought to be isolates. Larry Trask has shown on numerous occasions that (the supporters of) this theory treat(s) Basque data in a less than competent manner; the same can be said about Yenisseyan and much of the Tibeto-Burman (part of Sino-Tibetan) data I've seen in connection with this theory. Parts of this giganto-macro-grouping have some history, though: a Yenisseyan-Sino-Tibetan connection has been en vogue in the earlier days of Yenisseyology with investigators like Donner, Bouda and the outsider Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as I remember this was based on typological resemblances only), Sumerian has been compared to pretty much everything, so has Basque, but I think it is safe to say that a Basque - (unspecified) Caucasian connection has probably lured more early researchers than anything else (probably originally instigated by ergativity, which once made up for a quite exotic look-and-feel of a language - which is hardly the case today). Readers may already have inferred from my slightly ironic tone that I'm personally disinclined to buy much of this (I have working experience with Tibeto-Burman, Yenisseyan, NW-Caucasian, NE-Caucasian and a bit of Burushaski). St.G. Stefan Georg Heerstrasse 7 D-53111 Bonn FRG +49-228-69-13-32 From mcv at wxs.nl Wed Nov 11 16:38:31 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:38:31 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- Steven Schaufele wrote: >My wife has just forwarded to me a Reuters story (posted on www.cnn.com) >about an article by Merritt Ruhlen appearing this week in the >Proceedings of the National Academy of Sciences, claiming an affiliation >between Ket, the sole surviving Yeniseian language, and the Na-Dene >family of North America. Old news. You can find the Yeniseian ~ Na-Dene comparisons (as well as Yen ~ NCauc./SinTib./Basque/Burush./Nahali) in Ruhlen's 1994 book "On the Origin of Languages" (make sure it's the Stanford UP one, not Ruhlen's "The Origin of Language", Wiley & Sons, 1994). There's also a chapter on Na-Dene itself. For an updated list of Proto-Yeniseian reconstructions (with some comparisons to North Caucasian and Sino-Tibetan only -- the Na-Dene, Basque, Burushaski, Nahali ones were added by Ruhlen [and very poorly so for Basque, I must say]), you can consult or download Sergej Starostin's online database at http://starling.rinet.ru/ (also contains online etymological materials on North Caucasian [Nikolaev & Starostin's NCED], Dravidian, Altaic, Sino-Tibetan, Chuckchi-Kamchatkan, as well as Ozhegov, Zalizniak and Mueller's Russian(-English) dictionaries). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Wed Nov 11 16:38:19 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:38:19 EST Subject: s > r (Iberian) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- bwald wrote: >I am curious about this situation because I have long had the impression >that what is historically distinctive of Castillian and the Spanish-area >around it, but not extending to Portuguese, Catalan or other adjacent >Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among >them. The phenomenon does extend to "apitxat" Valencian Catalan, i.e. the central Valencian dialects of Valencia city and province (not N and S Valencian of Castello' and Alacant provinces). The rules are: -v- > -b- venia = /be'nia/ (also in std. Cat. /b@'ni@/) -z- > -s- casa = /'kasa/ (std. Cat. /'kaz@/) -Z- > -S- pujar = /pu'Sar/ (std. Cat. /pu'Za/) -dZ- > -tS- metge = /'metSe/ (std. Cat. /'medZ@/) I think "apitxat" also shares with Lleidata` Z > dZ initially and after /n/, so that the tongue twister ("16 judges of a courthouse eat liver from a hanged man") becomes: std. lg. /sEdz@ ZudZ at s dun ZudZat menZ at n fedZ@ dun p at nZat/ Lleidata` /sedze dZudZes dun dZudZat mendZen fedZe dun pendZat/ Valencia` /setse tSutSes dun tSutSat mentSen fetSe dun pentSat/ The big difference with Castilian is that there is no Z/S > x development (nor dz/ts > /T/). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From martinez at eucmos.sim.ucm.es Wed Nov 11 16:38:07 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 11 Nov 1998 11:38:07 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > But some sound changes are quite irreversible. Consider loss. In the > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting > /h/ was later lost. I predict confidently that the Greeks will never > reverse this change by re-introducing those long-gone /s/s, yes. as a product of analogy, see the -s- futures etc. From mccay at redestb.es Wed Nov 11 16:37:36 1998 From: mccay at redestb.es (Alan R. King) Date: Wed, 11 Nov 1998 11:37:36 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Benji Wald wrote: >In passing, I noticed Alan King acknowledge a correction by Miguel >Carrasquer. Alan wrote: > >>Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact >>that besides Castilian, both Galician and Romanian are also exempt from >>this general Romance development. In Castilian and Galician (but NOT in >>Romanian, as you point out), there is no /z/ phoneme. I was confusing >>these two things. > >I am curious about this situation because I have long had the impression >that what is historically distinctive of Castillian and the Spanish-area >around it, but not extending to Portuguese, Catalan or other adjacent >Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among >them. That is, I never thought that the ancestor of Castiliian, Spanish, >whatever, was exempt from the EARLIER Romance process of intervocalic >voicing of Latin -s- (among other sounds), but that by LATER developments >it devoiced the resulting -z- (in most environments -- in general,in >effect). Oh dear, I put my foot in it again. Apologies. I didn't really mean to say that, I can't even imagine how it happened. Must have been in a hurry. Of course I am wrong again, and you're right to pull me up on it. What I should have said (see if I get it right this time) is that the intervocalic s > z development, general in Romance (all? most?), was later reversed in Castilian and Galician, as opposed to most other Romance languages. In modern Castilian and Galician there is no /z/ phoneme. In most other Romance languages there IS such a phoneme, AND it comes from intervocalic /s/ in Latin. I know of one other modern Romance language which, synchronically, does not have /z/ deriving, historically, from Latin intervocalic /s/. As pointed out, Romanian (unlike Castilian and Galician) DOES have a /z/ phoneme, but from different historical sources than the /z/'s in the other Romance languages we're talking about. Whew! I hope that's better. So thanks for the correction. But what I was (we were) talking about, with regard to which this was a mere tangential point, was this (this is a reiteration of what I've already said, or tried to): (1) I believe that all languages in the Iberian peninsula, in their present (synchronic) forms, voice /s/ before voiced consonants in those cases (varieties) where the sibilant is not subject to some more radical transformation (rhotacism, aspiration, loss...). Even peninsular Basque, for what it's worth. (2) As regards the (synchronic) structural consequences of this voicing, we must differentiate between: a. languages which possess a /z/ PHONEME (and also a /s/ phoneme, of course); here we can talk about neutralization in the context in question. b. languages which LACK (synchronically, of course) a /z/ PHONEME; here we can only talk about conditioned allophones of /s/. Within the Iberian Peninsula, the languages in the a. group are Catalan and Portuguese. Those in the b. group are Castilian, Galician and peninsular Basque. Some varieties of Castilian (and also of Galician) are exempt from this classification because they do other things with their sibilants, such as aspirate them or drop them. And some dialects of Galician, and of Castilian (I thought and still think), rhotacize in the pre-voiced-consonant position. (I can provide documentation for this for Galician, and actually already have done so; for Castilian, which I have not investigated systematically, it is merely an impression, and I stand to be corrected if need be!) I say nothing about other Iberian languages or dialects, such as Asturian, owing merely to ignorance, but the corresponding information is most welcome. I wonder what I mangled this time?! In humility, Alan From drc at antnov1.auckland.ac.nz Wed Nov 11 16:37:17 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Wed, 11 Nov 1998 11:37:17 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Tue, 10 Nov 1998 18:21:52 EST > Reply-to: hubeyh at montclair.edu > From: "H.M.Hubey" > Organization: Montclair State University > Subject: Re: rhotacism from Ray Hickey > To: HISTLING at VM.SC.EDU > ----------------------------Original message---------------------------- > Scott DeLancey wrote: > > > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > > > > > Historical linguistics is circular, especially as it is based mostly on > > > IE. > > > IT says; > > > > > > 1. The set of languages, {x,y,z...} constitutes a family because of > > > 'regular > > > sound correspondences'. > > > 2. 'Regular sound correspondences' indicate a 'language family'. > > > > Could you give us a couple of examples of recognized language families > > that have been established on the basis of this kind of reasoning? > > There are a few, but it's not the standard methodology at all. > > Not Indo-European, for sure, which was first proposed and established > > on the basis of extensive correspondences in morphological paradigms. > > Aren't morphological paradigms also part of 'regular sound change"? No, they're not. But more to the point, contrary to what Mr Hubey seems to be suggesting, the little couplet above is not an instance of circular reasoning. It's merely the same statement phrased two different ways. Or rather, 2. is a statement of a general principle, of which 1. is a specific application. What Mr Hubey may be trying to articulate is the superficially circular-looking: 1. A,B,C are a language family ===> A,B,C have regular sound correspondences. 2. A,B,C have regular sound correspondences ===> A,B,C are a language family. But the attribution of circularity rests on a misreading of the relations between the propositions in 1. and 2. as the same. 1. is a causal relation -- regular sound correspondences result from the definition of a language family, plus the fact that sound change is regular. 2. is a progression from evidence to inference -- we observe regular sound correspondences, from which we conclude these languages are a family. (Whether this is an accurate account of what we actually do is not the question here.) It's no more circular than: 1. Patient has measles ===> patient has spots on face. 2. Patient has spots on face ===> patient has measles. This may be rough and ready diagnostic practice, but it isn't a logical fallacy. Ross Clark From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:36:15 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:36:15 EST Subject: probability & sound change (nee rhotacism from R.H.) Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > Robert R. Ratcliffe wrote: > > > >The proposal that given enough time X>Y>X is > > plausible is based on a hypothesis that the directionality of sound > > change is random. It isn't. There are clearly preferred (frequently > > This is something that used to bother me a lot and I used to think > that > there must be some universality. I am still not sure that there is not > > but I am more sure now that some of what we see is not really > universal but rather due to some specific local phenomena. This is not a matter of opinion, or something that can be decided simply by thinking about it. It is a matter of evidence, and that is what I meant by 'empirical'. Labov's conclusion, based on case studies of some twenty languages, is: ".. there are no directions of vowel shifting that are forbidden to speakers of human languages, but ... some directions are taken far more than others" (Principles of Linguistic Change 1994, p.116). Specifically he proposes that in chain shifts, long vowels rise, short vowels fall, and back vowels move to the front. > > attested) and disprefered (rarely or not attested) directions for > sound > > change. The only systematic work on this I know of is in Ch. 5 of > > Labov's 1994 "Principles of Language", which only deals with vowels. > This has to be from a specific time period and a specific set of > languages. Labov's sample (mistake in title, should be Prinicples of Linguistic Change, sorry) is small by the standards of typology, and most of the languages are European. Further work in the diachronic typology of phonological systems is certainly desirable. Nonetheless, as empirical scientists we have to prefer a theory consistent with some evidence (directional sound change) over a theory (random sound change) which has not been shown to apply to any evidence. It is certainly possible that directionality effects are part of a short term cycle which evens out to zero in the long run, but I haven't seen any evidence which would support this view. > There is a similar problem in probability theory, that of stationarity > of a signal. It's impossible to prove. No matter how many languages > you > look into you can only look at a specific time interval. It's hard to > say if the sound changes are due to universals or to a particular > combination of sounds and phonotactics. This passage is a bit cryptic. I haven't talked about universals, but only about emprically observed statistical tendencies, and haven't discussed explanatory factors at all. It would certainly be meaningless to say that something is "due to" a statistical tendency. The length of the time interval is only limited by the historical record, but even so, it isn't necessarily a problem. If directional effects are only part of a long term cycle X>Y>X (a cycle longer than the historical record), this should still be demonstrable, based on the fact that not all languages should be at the same point in the cycle. In some langues X>Y should be the trend in the attested period, in others Y>X. Breadth of the sample can substitute for depth. > However, what I wrote above referred, in general, to any two sounds. > It > may be that for some specific X and some specific Y, the sound change > X > Y for some specific language (i.e. specific set of phonemes and > phonotactics) may be irreversible. But in general I do not see any > reason to assume that no sound change is reversible given enough time. You're right, of course. Directional trends relate only to what is probable in sound change, not to what is possible or impossible. On the other hand it occurs to me that there is one type of sound change which IS irreversible, that is X>0, as found in mergers, phoneme loss, or assimilation in clusters. > Right. However, it is strange on the other hand to see those consonant > > clusters and lack of vowels in languages like Abaza, Georgian, or > Khoisan and its clicks. The foremost question is this: if there is > such a > universal trend (say toward lack of cases, or toward voicing, or from > stops to fricatives, or approximants) how then did the language (any > language) get those stops in the first place? Or how did some language > get consonant clusters at all? Directionality does not mean teleology. To say that sound change moves in a non-random direction is not to say that it is going anywhere in particular. One might propose (although I didn't) that over time phoneme inventories get smaller, or that 'marked' phonemes tend to be lost in favor of unmarked. I have yet to see a teleological proposal of this type which can survive the test of the evidence. The directionality trends discovered by Labov appear ultimately due to physiological constraints both articulatory and auditory. In other words the direction of sound change is constrained by the present not by the future. I suppose this is what you mean by talking about local versus universal (i.e. short term vs. long term). There are no long term directional trends (as far as the evidence now goes). But we cannot conclude from this that the sum of short term directional trends adds up to zero or no directionality or random directionality. The prequisite to a realistic probalistic model of language change is a systematic research program in diachronic typology. We have to establish the probable direction of change on the basis of statistical data from observed changes-- not on the basis of a priori reasoning. I suspect the ultimate model will look something like those typhoon maps we see here (don't know if you have them in other countries): There is a circle showing where the storm is and from the outer edges of the circle there is a wedge shaped projection showing the area into which the storm is likely to move. At the big end of the wedge there is another bigger circle showing where the storm is likely to be at the end of a certain period (24 hrs, e.g.). From this big circle another wedge is projected, with a bigger mouth showing where the storm should go in the next period, and so on. > The same applies to linguistics changes.There are many scales at which > changes occur, and if we mix up these levels we create opposing ideas. A priori we don't know what models or scales apply to linguistic changes. We can't know until we have tried to apply them, that is test them against the data. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From artabanos at mail.utexas.edu Wed Nov 11 16:34:52 1998 From: artabanos at mail.utexas.edu (Tom Wier) Date: Wed, 11 Nov 1998 11:34:52 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- "D. Anthony Tschetter-Breed" wrote: > Forgive me for being anectdotal, but when topic comes up of sound changes > from X to Y and back to X, I always think of the English: > > askan > ask > ax (in some dialects, for example Black American English) > > One could propose that this is not an example of X>Y>X either on the > grounds that the "ax" form is not an innovation but rather a retention of > an earlier form, or perhaps that it's not systematic. I don't know. Any > thoughts? The form in question I believe is actually connected to other dialects in the South which themselves go back to nonstandard dialects in England. OE had both "bscian" and "bcsian", and dialects in America were not homogenously from Southern England around London (on which dialect the standard was based) or anything, so I find it far more likely that it's merely a retention of the earlier form. ======================================================= Tom Wier ICQ#: 4315704 AIM: Deuterotom Website: "Cogito ergo sum, sed credo ergo ero." "The heart is deceitful above all things, and desperately corrupt; who can know it?" Jeremiah 17:9 ======================================================== From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:34:32 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:34:32 EST Subject: s > r (Iberian) correction Message-ID: ----------------------------Original message---------------------------- bwald wrote: > In a separate message, Robert Ratcliffe wrote: > > >But in general I do not see any reason to assume > >that no sound change is reversible given enough time. > This wasn't me. But Mark Hubey responding to something I wrote. Have to be careful with these interlinear discussions. My point was that some directions of change are more or less probable not that any are necessarily impossible. But some changes ARE irreversible, as Larry Trask pointed out earlier in this thread-- namely phoneme loss, mergers, and I would add assimilation in clusters (okto >> otto type changes). -- +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From hubeyh at montclair.edu Wed Nov 11 16:32:26 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 11 Nov 1998 11:32:26 EST Subject: Hawaiian meli Message-ID: ----------------------------Original message---------------------------- Lyle Campbell wrote: > > ----------------------------Original message---------------------------- > Ross Clark's example of Hawaiian -meli- 'honey' as a loanword from Greek > via classically-educated missionaries translating the Bible is a very nice > one, as Larry Trask pooints out. Larry notes that it ruins one of his > favorite examples of chance resemblance, but becomes a particularly > interesting example of borrowing. I suppose you could sort of raise the > ante on interestaing cases of chance resemblance and borrowing by throwing > Maori -mieri- 'honey' into the mix. A comparison of Hawaiian -meli- and > Maori -mieri- (bother the -i- vowel difference) might seem to suggest a > Polynesian cognate set (throw in Niuean -meli- 'heney' as well, also > apparently from Greek), but the Maori word is actually a French loanword > (from French -miel- 'honey'), courtesy apparently of early French Catholic > missionizing activities in New Zealand, which soon faded in the country. > (There are not many French loans in Maori, but a favorite is -wi:wi:- > 'French' < French -oui- 'yes'). As Larry Traks points out, this -meli- / > -mieri- false cognate is no longer a case of sheer accidental similarity, > in that both are from Indo-European languages, but we still have accident > to thank for it in a way, in that by sheer happenstance Hawiian ended up > with a Greek form and Maori with a French one (which happen to be related > languages), not something that would have been expected. These are nice examples. But that is like the fisherman's game. The one that got away was the biggest of them all. How are we to take into account all of those words that are said to be cognate in IE when it is quite possible that many (most) might be left over a substratum that was living in that neighborhood for many thousands of years and had stabilized so that changes were taking place very slowly? Of course, this is supposed to point out that false matches can occur. True. IT is also possible that we might miss out on real cognates. Maybe they cancel each other. That is why statistics is for. YOu take N measurements and average them, and the random deviations cancel. It also tells you how much confidence you can have in your results. That is a side benefit. > Lyle Campbell -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From johanna at uclink.berkeley.edu Thu Nov 12 01:27:35 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Wed, 11 Nov 1998 20:27:35 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36463306.60B6E2F@montclair.edu> Message-ID: ----------------------------Original message---------------------------- Mark Hubey writes: >----------------------------Original message---------------------------- > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), >emesal (female speech in Sumerian), emcek (breasts, udder), meme >(breast), >emzirik, etc etc. > >Do you understand the probabilistic implication of such patterns? > I'll take the liberty of commenting, as I understand the probabilistic implication of such patterns. What are the chances of finding, in any language, a word consisting of a vowel (any vowel? or just a non-round vowel?) followed by [m], with or without any following material of any shape, having one or the other of the set of 7 glosses cited above plus presumably any other meaning that might have to do with females? That boils down to: what are the chances of finding a sequence Vm(-) in any of a wide (or open-ended) set of meanings? Success is virtually guaranteed, since you get to keep looking until you find something that fits. I've computed the chance of finding resemblances in which two similar consonants occur in the same order in words with similar senses. If you're allowed: (1) a couple of distinctive features' leeway in defining generic consonants (so e.g. p, b, and f are all taken to match) (2) and much phonotactic leeway (so e.g. epte, fad, upatha, bdezolg, pet, puot, etc., etc. are all taken to match because they all have a p, f, or b followed later in the word by a t, d, or th, and no intervening consonant) (3) and up to five senses' leeway (e.g. 'black', 'night', 'dark', 'soot', 'shadow' or any other set of five meanings you consider related; or e.g. 'fingernail', 'finger', 'hand', 'arm', 'claw') then the event probability of such a match is 0.04, and 25 such matches out of a pre-specified 100-word list are required to reach the 95% confidence level on a binary test of two languages. (That's the conventional minimum level for deciding that the number and degree of resemblances are not random. I got the number of 25 from a binomial probability table for an event probability of 0.04 and 100 trials.) For one-consonant sets like Mark's, over half of the 100-word list would have to consist of matches in order to reach 95% confidence. All this is if you prespecify the 100 glosses, prespecify the range of 5 for each, and prespecify the generic consonants. And choose in advance the two languages you want to compare. If you get to look through all words (i.e. entire dictionaries) of any languages, then the required numbers of matches go up. So the probabilistic implication of such patterns is nil, unless you have over 50 of them out of some standard 100-word list. Such are the hazards of doing open-ended searches. (This brings us to Yeniseian and Na-Dene, of which more anon.) The computation of probability in limited searches is described (briefly) in my paper 'The comparative method as heuristic' in M. Durie and M. Ross, eds., The Comparative Method Reviewed (Oxford UP, 1996). I'm working on a fuller explanation. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From johanna at uclink.berkeley.edu Thu Nov 12 01:28:08 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Wed, 11 Nov 1998 20:28:08 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Steven Schaufele writes: ... a Reuters story (posted on www.cnn.com) about an article by Merritt Ruhlen appearing this week in the Proceedings of the National Academy of Sciences, claiming an affiliation between Ket, the sole surviving Yeniseian language, and the Na-Dene family of North America. ... I'm wanting to know, does anybody subscribing to this EBB know anything about this? Is there anything to this proposed affiliation? And if not, is anybody doing anything about clarifying the issue for the general public? - - - - - - - - - - - - - - - - - - - - - - - - - I have seen the article and discussed it with a couple of journalists who contacted me about it. Here is the gist of what I said: The burden of proof is on Ruhlen to demonstrate that his resemblances exceed what is expected by chance. There are well-known tests, standards, criteria, etc. for doing this in linguistics and in statistics. He has not invoked or applied any such criteria, so his findings must be assumed to be chance resemblances until and unless he shows otherwise. Ruhlen lists 36 word sets that he considers resemblant and indicative of genetic relatedness between Yeniseian and Na-Dene. In fact these 36 sets are not numerous enough and not closely enough resemblant to exceed the range of chance. I've worked out the chances of finding words with two similar consonants in the same order, with similar but not necessarily identical meanings, in two languages. Out of a fixed list of 100 meanings chosen in advance (this is analogous to asking "what is the probability that similar forms will mean 'water' in both languages?", and so on for another 99 glosses), this is how many resemblant sets it takes to exceed the range of chance and show that relatedness is likely: 2-consonant words with the very same meaning: 7 2-consonant words with similar meanings (modeling this as a search that allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or whatever; these must also be specified in advance): 25 1-consonant words (or 2-consonant words with one resemblant consonant and one non-resemblant one) with the same meaning: 27 1-consonant words with 5 senses' leeway each: over 50 Ruhlen has: 15 words with two resemblant consonants in the same order 2 words with two resemblant consonants but in different orders or with other consonants intervening ('children' and 'foot') 18 words with one resemblant consonant 1 word with perhaps no resemblant consonant ('name', where the initial glottal stop in the Proto-Yeniseian form is found in no daughter language and must there be phonetic detail of the pronunciation of initial vowels rather than a true structural consonant; and where similarly there doesn't seem to be any evidence for reconstructing a glottal stop as initial, or even as non-initial but the first consonant in the root, for Na-Dene). For simplicity and to give Ruhlen the benefit of the doubt these can be described as 17 two-consonant resemblances and 19 one-consonant resemblances. 11 of Ruhlen's sets have considerable semantic leeway; most have some leeway; for only a few is the sense really the same. So this can be modeled as a search of up to perhaps five senses. In practice this means the researcher finds a word in one of the languages and gets to cast about in the other language looking for a word of similar form in a meaning plausibly connected to that of the first word. Occasionally s/he is lucky and finds a resemblance in the very same sense; usually s/he has to search through a few close senses; and sometimes s/he has to search longer. So even if Ruhlen had started out with a strict list of 100 glosses and a list of 5 senses' worth of leeway for each, he would need about 50 sets of the kind he has. (I'm not calculating this with any precision. I figure he has about half of the 25 2-consonant words he needs, so he also needs to have about half of the 50+ 1-consonant words he needs, and the sum of about half of 27 and about half of over 50 is somewhere around 50.) But he didn't set up a closed list of glosses. His wordlist is not 100 but some open-ended number. At least the total number of words found in the source on Yeniseian he draws on, or maybe the total of those plus any additional meanings found in his various Na-Dene sources. Or maybe it's the average total number of root words per language on earth. I don't know quite how to model this, but he clearly has a wordlist of well over 100. So he needs correspondingly more than the about 50 words he would need if he used a pre-specified closed list of 100. My calculation is for comparisons of actual daughter languages, while Ruhlen compares protolanguages. A protoform has a greater chance of occuring in the protolanguage than in any daughter language. (This is because a protoform occurs in the protolanguage by definition, while for any cognate set it's likely that one or another daughter language will happen to lack a cognate.) Therefore, for comparisons of protolanguages larger numbers of resemblant words will be needed than for daughter languages. For instance, Ruhlen's set 'hunger' has Haida as its only Na-Dene representative, and 'river' has Haida as the only form with a two-consonant resemblance. Some of the sets have no Haida cognate, e.g. 'foot', 'birch bark', 'boat'. From the point of view of methodology, this means he allowed himself to adduce evidence from either Haida or from Eyak-Athabaskan-Tlingit, and the chances of finding a resemblant in one or another family are greater than finding a resemblant in one particular one. Again, this can be modeled as a search in which if the researcher doesn't find a resemblant in Eyak-Athabaskan-Tlingit s/he gets to look for one in Haida. In some sets an ejective consonant in Na-Dene corresponds to a sequence of consonant plus vowel plus glottal stop in Yeniseian (e.g. 'birch bark'); in others (e.g. 'clay') an ejective just corresponds to a Yeniseian consonant with no glottal stop elsewhere in the word; in still others there is a Yeniseian glottal stop but no ejective anywhere in Na-Dene (e.g. 'lake', 'word'). Thus Ruhlen seems to have allowed himself to scan consonants in whatever way maximized resemblances. For all of these reasons, Ruhlen's resemblances are no more numerous and no closer than would be expected to come up by chance. They are exactly the sort of thing one finds when looking through dictionaries and casting about to find resemblances with few constraints on how the search is to be conducted. This is all assuming that the linguistic analysis and the protoforms are sound, matters I cannot judge. Any errors in data, analysis, or reconstruction increase the number of resemblances that must be adduced if they are to be regarded as non-accidental. The paper is not up to date on human genetics and archeology. It mentions only the Greenberg-Turner-Zegura (1986) theory of three American settlements, not mentioning all the work on mitochondrial DNA which has been done since 1987. The claim that "the first migration of the Amerinds [occurred] about 11,000 years ago" is no longer current; the well-dated, well-accepted Monte Verde site in southern Chile at ca. 12,500 years ago shows that the first Americans entered well over 11,000 years ago. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From bwald at HUMnet.UCLA.EDU Thu Nov 12 01:32:44 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Wed, 11 Nov 1998 20:32:44 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- D. Anthony Tschetter-Breed writes: >Forgive me for being anectdotal, but when topic comes up of sound changes >from X to Y and back to X, I always think of the English: >askan > ask > ax (in some dialects, for example Black American English) This example is problematic because OE asc(i)an and acs(i)an have both continued UNINTERRUPTED into current dialects of English, e.g., there are still British dialects that have "aks" (it's common in some working class areas of urban Mersey and Lancashire, for example). It cannot be assumed that Black English "reinvented" ask > aks, rather than received it from relevant British dialects, and has preserved it while most American dialects have not. This is NOT a clear case of reversal of a *sound change* that you want. Meanwhile, it turns out that there indeed have been spontaneous reversals ask > aks. For example, many New York City working and lower class speakers use "aks". It is clear that the metathesis was spontaneous in this case, for the following reasons. In New York City, "short a" (as in "*a*sk") is RAISED to [E] (a mid to high front tense vowel with some inglide for its final off-glide). However, this raising is highly conditioned. It occurs in closed syllables before voiceless fricatives, but NOT before voiceless stops. Thus, "short a" as in "back, bat, tap, tax", etc. remains LOW. The NYC "aks" speakers, however, RAISE "short a" in "ask", i.e., they pronounce it "Eks", and it contrasts with the low vowel is "axe". From this we realise that FIRST, a > E /_s(k) etc (as in "past", "task", "bath", "raft", even "bash" etc) and THEN sk > ks / #E_# (ONLY IN THIS WORD, not, for example, in "task" vs. "tax", etc.) I gave an example of a real reversed sound change, for what it's worth, in my last message. Latin > Romance s > z /V_V and LATER, Central Spanish z > s (but without intervocalic conditioning). That was NOT a matter of -s- surviving from Latin in some dialects of Spanish-to-be and then rediffusing. On the contrary, it was a later sound change that happened to reverse the previous and much older sound change. No mysterious "invisible hand" guided Spanish to reverse s > z BACK to s (and indeed the conditioning is different). It just happened -- because both directions are possible (under certain conditions -- certainly NOT ****z > s /V_V). Essentially, such "reverse" changes have no more intrinsic interest than to serve as a warning that in more time-compressed and less documented cases that *s = s etc does not necessarily mean that there NEVER were any intervening stages between a reconstruction and "unchanged" documented reflex. I don't remember why this issue arose. Maybe in the context of somebody wanting to pose borrowing from some language into another language when the reflex cannot be accounted for in terms of any attested stages of either donor or recipient languages. P.S. One comment on one of Hubey's dyspeptic replies to Trask: Larry wrote: > Mr. Hubey, are you suggesting that one need not spend years studying > Sumerian in order to know Sumerian? Mark replied: >In order to produce a list of cognates all you need is a dictionary. Unh Unh. Wrong! As Larry implied in his message about wrong morpheme cuts, you also need a GRAMMAR (which dictionaries usually don't supply, at least not in sufficient detail). Dictionaries do not generally analyse WORDS into MORPHEMES, esp if the morphemes are derivational affixes of varying degrees of obscurity. Amateurs (and even some professionals) produce a lot of crap by using dictionaries (of languages whose grammars they are ignorant of), and arbitrarily inserting morpheme boundaries to fit preconceived cognates into procrustean molds by getting rid of the "bad" stuff. Such practices conceitedly ape the more judicious use of the technique by more RESPONSIBLE linguists, who nevertheless sometimes make mistakes with that reconstructive technique. At least when responsible linguists indulge in the technique (responsible amateurs included) they continue to take seriously the issue of the historical significance of the stuff they snipped off. From hubeyh at Montclair.edu Thu Nov 12 01:33:22 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Wed, 11 Nov 1998 20:33:22 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), emzirik, etc etc. > > This illustrates the fact that a little knowledge of the languages > involved is never a bad thing. Larry should have been slightly more > precise by indicating that Basque was not borrowed from vanilla > Occitan, but from Gascon/Bearnais. Then it immediately becomes clear > that the word is simply Latin FEMINA, which regularly becomes hemna > (hemm@ ~ hemno) in Gascon (f > h). The ultimate root is IE *dheH1- > "to suck" (*dheH1-mHn-oH2 > femina). The coincidence with PTurkic > *eme "woman" is entirely coincidental. And so is, with even more I don't know that PT 'eme' is 'women'. I only pointed out that 'em' is an old word in Turkic, and might well be one of those that belongs in PW (protoworld) for those who believe in it. > > reason, Sumerian where the element doesn't mean > "woman" at all, but "language, tongue". The word means "thin, > refined" (it used to be thought that was "woman", and one can > still find that in older books, but in fact the Sumerian word was > ). Related words again. This is one of the reasons why the creation of a standard "semantic" space is so important for historical linguistics. All the powerful tools of math are unavailable to linguists because of that, and probably because of this, articles like those Times, etc will keep appearing despite the fact that some hate it. If there is no standard and no impartial referee, that is the way it will always be. If sociologists and psychologists are now fighting over correlation coefficients it is time for linguists to bite the bullet. > > >Do you understand the probabilistic implication of such patterns? > > Yes. In short: "Coincidences happen" and "Garbage in, garbage out". That is another one of those things linguists say. I will refrain from further comments. There are not really too many things to do in the face of uncertainty. Everyone from rocket scientists, engineers, economists, biologists, to computer scientists, workers in speech recognition, synthesis, phonetics and psychologists does the same thing. Except of course linguists. But then again, nobody writes articles in Times that the moon is full of green cheese, or that the world is supported on the back of turtles. > > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From hubeyh at Montclair.edu Thu Nov 12 01:34:26 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Wed, 11 Nov 1998 20:34:26 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > ----------------------------Original message---------------------------- > At the risk of getting caught in the crossfire, I'd like to interject > myself in the debate between Larry Trask and Mark Hubey. I'm in > agreement with about 90% of what LT has to say, and have to > say (with no malice intended) that MH's postings reveal a great > ignorance of the field of linguistics. Nonetheless I do think that MH > I think what is necessary is to get going discussing what is being done and how, and something good might come out of it. > is asking some legitimate questions and deserves a better response from > the profession than the old refrain that the esoteric knowledge is only > revealed to those who have joined the secret brotherhood. > That is the real reason I post what I do. There are computer programs that paint and compose music. Is it really that hard to believe that linguistic reconstruction is no less structured? I bring this up, because for a long time the anti-AI crowd used arguments similar to those offered often on linguistics lists for why AI would be impossible. I will try to reply to the rest of the comments later from home. > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From delancey at darkwing.uoregon.edu Thu Nov 12 01:34:44 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Wed, 11 Nov 1998 20:34:44 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Ralf-Stefan Georg wrote: > Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially > Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as > I remember this was based on typological resemblances only), Sumerian has Pretty much, yes--their both having tone weighed heavily in the hypothesis (since we now know that in both families tone is a secondary development in the languages and branches where it occurs this loses any force it might ever have had), and Sapir saw a fundamentally monosyllabic structure to both. (That may not be as completely far-fetched a way of thinking about Athabaskan as it sounds). But he did dig up a handful of actual resemblants. Some can be found in Victor Golla's _Sapir-Kroeber Correspondence_, especially letter 332: Tlingit k'a 'surface' : Navaho k'a 'surface' : Tibetan k'a 'surface' Chinese t'an 'charcoal' : Haida s-t'An 'charcoal' OChin ti 'this' : Ath. di 'this' OChin ti 'pheasant' : Ath di 'partridge' Nadene k'u 'hole' : Indo-Chinese [sic] k'u 'hole', with several supporting forms from different N-D and S-T languages. Pretty standard "long-ranger" stuff. In an addendum to that letter he has a very nice word-family comparison involving words having to do with 'tie', 'twist', 'rope', 'snare', 'trap', which is the only evidence I've ever seen presented for Sino-Na-Dene that made me feel even for a minute like I'd like to see more. I had heard that in his notebooks he had more extensive comparisons, so one day while I was at the APS looking for Chinookan stuff I spent an hour looking for ST-ND stuff, but didn't find any. > been compared to pretty much everything, so has Basque, but I think it is > safe to say that a Basque - (unspecified) Caucasian connection has probably > lured more early researchers than anything else (probably originally > instigated by ergativity, which once made up for a quite exotic > look-and-feel of a language - which is hardly the case today). I wish I could remember, or had saved, the citation. I can't even remember the journal--but I did once see an article, by some English individual, from the 1940's or so I think, claiming a Tibetan-Basque relationship, pretty much entirely on the grounds that they were both ergative. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From drc at antnov1.auckland.ac.nz Thu Nov 12 12:46:10 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 12 Nov 1998 07:46:10 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Wed, 11 Nov 1998 12:22:12 -0500 > From: "H.M.Hubey" > Reply-to: hubeyh at montclair.edu > Organization: Montclair State University > To: Ross Clark > Cc: HISTLING at VM.SC.EDU > Subject: Re: rhotacism from Ray Hickey > Ross Clark wrote: > > > > > Aren't morphological paradigms also part of 'regular sound change"? > > > > No, they're not. > > > > But more to the point, contrary to what Mr Hubey seems to be > > suggesting, the little couplet above is not an instance of circular > > reasoning. It's merely the same statement phrased two different ways. > > Or rather, 2. is a statement of a general principle, of which 1. is a specific application. > > > > What Mr Hubey may be trying to articulate is the superficially > > circular-looking: > > > > 1. A,B,C are a language family ===> A,B,C have regular sound > > correspondences. > > > > 2. A,B,C have regular sound correspondences ===> A,B,C are a language > > family. > > > > Perhaps only some people use it circularly, or many linguists use it > circularly, but it is in use. Could you cite an example or two? I don't mean of 1. or 2. in use separately, but of the two used together as a fallacious syllogism. > > > But the attribution of circularity rests on a misreading of the > > relations between the propositions in 1. and 2. as the same. 1. is a > > causal relation -- regular sound correspondences result from the > > definition of a language family, plus the fact that sound change is > > regular. > > But borrowings also create regular sound correspondences. Yes, just as things other than measles can produce spots on the face. We need to take such things into consideration if we want to raise our competence in historical linguistics (or medical diagnosis) from this very rudimentary level. > > 2. is a progression from evidence to inference -- we observe > > regular sound correspondences, from which we conclude these languages > > are a family. (Whether this is an accurate account of what we > > actually do is not the question here.) It's no more circular than: > > > > 1. Patient has measles ===> patient has spots on face. > > 2. Patient has spots on face ===> patient has measles. > > This rests on something different. > > 1. Patient has measles (definition comes from some other place, but may > include spots on the face). These days the defn would come from being > able > to culture the bacteria. Then the spots on the face and measles > correlate. > The time depth is short and one can see a non-measles person get it, get > sick, etc. > > Because of the correlation of measles and spots, 2 then becomes an > implication. > > > But that does not work so in historical linguistics because we never had > a record of any language family (knowing its relatives, etc) but all of > it > rests on a larger theory of which regular sound correspondence must be a > part. Well, we do in fact have records of various language families. What are you trying to say here? > So it is a whole mess of correlations which lead towards that > conclusion. > > More to the point it is based on this reasoning. > > 1. These languages have too many things in common. IOW, there are many > words in > these languages which can be made to look like each other with similar > meanings > and which could not be due to chance. > > 2. If that is not due to chance then either they got these words from > each other > or the words are all descended from a common language. > > 3. We have plenty of evidence (what?) that these languages did not get > these > words from each other. > > 4. Therefore these words in these languages must all come from an > earlier common > source. > > This is how it is supposed to work, but you can see rather easily how > and where > problems crop up, and where arguments occur. > > Isn't this basically right? Yes, problems crop up and arguments occur, to be sure. I recognize 1-4 as a rough outline of the reasoning by which one arrives at a hypothesis of genetic relatedness among languages. Rather than argue about details, I'd like to know where you're going with it. Are we finished with the idea that it's logically circular? Ross Clark From hubeyh at montclair.edu Thu Nov 12 12:47:06 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:06 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- bwald wrote: > Mark replied: > >In order to produce a list of cognates all you need is a dictionary. > > Unh Unh. Wrong! As Larry implied in his message about wrong morpheme > cuts, you also need a GRAMMAR (which dictionaries usually don't supply, at > least not in sufficient detail). Dictionaries do not generally analyse > WORDS into MORPHEMES, esp if the morphemes are derivational affixes of > varying degrees of obscurity. Amateurs (and even some professionals) > produce a lot of crap by using dictionaries (of languages whose grammars > they are ignorant of), and arbitrarily inserting morpheme boundaries to fit > preconceived cognates into procrustean molds by getting rid of the "bad" > stuff. Such practices conceitedly ape the more judicious use of the > technique by more RESPONSIBLE linguists, who nevertheless sometimes make > mistakes with that reconstructive technique. At least when responsible > linguists indulge in the technique (responsible amateurs included) they > continue to take seriously the issue of the historical significance of the > stuff they snipped off. Let me rephrase it. It should be obvious to everyone by now that Starostin is way ahead of the game than the 99% of the subscribers. It should be clear by now that if he has already built up a database of lexicons of various languages with their meanings, writeable in ASCII, Unicode etc, he has singlehandedly done what should have been done by the linguistics community. But that is not all. If there are people who are writing programs to paint (yes, produce art), and compose music, it takes no genius to see that even if Starostin only wrote (or got a student to write) a brute-force, dumb program on a commodity grade PC, he can uncover relationships that many humans cannot do, even if they collaborate. The reason for this will take too long to explain. But given a set of words (and their meanings) even a brute-force program can keep cranking 24 hours a day to produce cognates via regular sound changes, clusters, and things that a typical linguist does not even know exists. It's too bad that the attitude of most linguists is, in fact, the most damaging to themselves and their own professions. But, that is the way evolution is. Short term goals and intuition only go so far. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:47:23 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:23 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > I've worked out the chances of finding words with two similar consonants in > the same order, with similar but not necessarily identical meanings, in two > languages. Out of a fixed list of 100 meanings chosen in advance (this is > analogous to asking "what is the probability that similar forms will mean > 'water' in both languages?", and so on for another 99 glosses), this is how > many resemblant sets it takes to exceed the range of chance and show that > relatedness is likely: > > 2-consonant words with the very same meaning: 7 > 2-consonant words with similar meanings (modeling this as a search that > allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or > whatever; these must also be specified in advance): 25 > 1-consonant words (or 2-consonant words with one resemblant consonant and > one non-resemblant one) with the same meaning: 27 > 1-consonant words with 5 senses' leeway each: over 50 I erased the rest not because it is not important but because I want to ask about this. Is it not true that the most important consideration in probability theory is knowing the sample space? In other words, when "matches" due to chance are being calculated, should not the fact that the two languages have (or seem to have) the same set of phonemes enter into the calculation? In other words, the sample space should consist of the phonemes that the languages could have had (along with the phonemes that they do have) but do not? The calculations should involve conditional probabilities. No? Secondly, I also made some calculations. But mine is not for phonemes and does not take into account phonemes for the reason that they cause more complications, and do not take into account that the same speech space available for humanity is divided up differently and into different number of chunks (phonemes) in different languages. The fact that out of possible M phonemes if languages seem to have a particular set of N phonemes that in itself has to be accounted for. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:47:59 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:59 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), > >emzirik, etc etc. > > > >Do you understand the probabilistic implication of such patterns? > > > > I'll take the liberty of commenting, as I understand the probabilistic > implication of such patterns. What are the chances of finding, in any > language, a word consisting of a vowel (any vowel? or just a non-round > vowel?) followed by [m], with or without any following material of any > shape, having one or the other of the set of 7 glosses cited above plus > presumably any other meaning that might have to do with females? No. I was thinking of something more. If words (phonological shapes and their meanings) did not cluster there would be no such thing as etymology. If some word X or its reflexes shows up in a family in random scatter but is strongly represented in another family, what does that imply? Secondly, if words of two different languages were generated independently of each other, the if there are chance occurences it is not only the number of them that matters but also the patterns. But counting up the numbers shows no indication of patterns. Word generation is a Markov process so the tests should be run on some model that purports to be a model of language evolution/development. > That boils down to: what are the chances of finding a sequence Vm(-) in > any of a wide (or open-ended) set of meanings? > > Success is virtually guaranteed, since you get to keep looking until you > find something that fits. Sure, it is too easy. > > I've computed the chance of finding resemblances in which two similar > consonants occur in the same order in words with similar senses. If you're > allowed: > > (1) a couple of distinctive features' leeway in defining generic > consonants (so e.g. p, b, and f are all taken to match) > > (2) and much phonotactic leeway (so e.g. epte, fad, upatha, > bdezolg, pet, puot, etc., etc. are all taken to match because they all have > a p, f, or b followed later in the word by a t, d, or th, and no > intervening consonant) > > (3) and up to five senses' leeway (e.g. 'black', 'night', 'dark', > 'soot', 'shadow' or any other set of five meanings you consider related; or > e.g. 'fingernail', 'finger', 'hand', 'arm', 'claw') > > then the event probability of such a match is 0.04, and 25 such matches out > of a pre-specified 100-word list are required to reach the 95% confidence > level on a binary test of two languages. (That's the conventional minimum > level for deciding that the number and degree of resemblances are not > random. I got the number of 25 from a binomial probability table for an > event probability of 0.04 and 100 trials.) For one-consonant sets like > Mark's, over half of the 100-word list would have to consist of matches in > order to reach 95% confidence. All this is if you prespecify the 100 > glosses, prespecify the range of 5 for each, and prespecify the generic > consonants. And choose in advance the two languages you want to compare. > If you get to look through all words (i.e. entire dictionaries) of any > languages, then the required numbers of matches go up. Some of these computations were also done on Language. And that list is mainly for quantitative approaches. Yes, but this only takes into account number and not pattern. There are lots of ways of testing things. > So the probabilistic implication of such patterns is nil, unless you have > over 50 of them out of some standard 100-word list. Such are the hazards > of doing open-ended searches. (This brings us to Yeniseian and Na-Dene, of > which more anon.) I was talking about the implication of patterns of words in a given language. One of the biggest problems that I encounter all the time is that neutral evidence works for the advantage of the dominant theory. For example, most of the IE words could be due to the substratum which could have been a family. One can always insist that the reason why IE words resemble each other is because they are all left over from a previous language which was spread out over the same region. > The computation of probability in limited searches is described (briefly) > in my paper 'The comparative method as heuristic' in M. Durie and M. Ross, > eds., The Comparative Method Reviewed (Oxford UP, 1996). I'm working on a > fuller explanation. Yes, I read it about a year ago. I think you had an article on comparison of Hittite with others. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:48:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:48:25 EST Subject: probability & sound change (nee rhotacism from R.H.) Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > but I am more sure now that some of what we see is not really > > universal but rather due to some specific local phenomena. > > This is not a matter of opinion, or something that can be decided simply > by thinking about it. It is a matter of evidence, and that is what I > meant by 'empirical'. Labov's conclusion, based on case studies of some > twenty languages, is: ".. there are no directions of vowel shifting that > are forbidden to speakers of human languages, but ... some directions > are taken far more than others" (Principles of Linguistic Change 1994, > p.116). Specifically he proposes that in chain shifts, long vowels > rise, short vowels fall, and back vowels move to the front. I have nothing about empiricism. But these directions must be local because if it were not, we would not have had those which are supposed to disappear. > Labov's sample (mistake in title, should be Prinicples of Linguistic > Change, sorry) is small by the standards of typology, and most of the > languages are European. Further work in the diachronic typology of Even if all the languages of the world were in the sample, it would still be local (temporally) we do not know what it was like before we recorded these languages. > phonological systems is certainly desirable. Nonetheless, as empirical > scientists we have to prefer a theory consistent with some evidence > (directional sound change) over a theory (random sound change) which has > not been shown to apply to any evidence. It is certainly possible that > directionality effects are part of a short term cycle which evens out to > zero in the long run, but I haven't seen any evidence which would > support this view. Yes, I understand all that. But even if someone produced an equation of evolution, it still cannot be universal because it is still local in time. A time evolution cannot be created from a short sample in time. > > There is a similar problem in probability theory, that of stationarity > > of a signal. It's impossible to prove. No matter how many languages > > you > > look into you can only look at a specific time interval. It's hard to > > say if the sound changes are due to universals or to a particular > > combination of sounds and phonotactics. > > This passage is a bit cryptic. I haven't talked about universals, but > only about emprically observed statistical tendencies, and haven't > discussed explanatory factors at all. It would certainly be meaningless > to say that something is "due to" a statistical tendency. The length of > the time interval is only limited by the historical record, but even so, > it isn't necessarily a problem. If directional effects are only part of > a long term cycle X>Y>X (a cycle longer than the historical record), > this should still be demonstrable, based on the fact that not all > languages should be at the same point in the cycle. In some langues X>Y > should be the trend in the attested period, in others Y>X. Breadth of > the sample can substitute for depth. YEs, given a long enough time and a large enough sample we can say that such and such changes have a tendency to occur. But is it a tendency that is unconditioned or is it a tendency that arose from some reasons which are lost in time? To attack the problem a model is needed, a model of change over time. > You're right, of course. Directional trends relate only to what is > probable in sound change, not to what is possible or impossible. On the > other hand it occurs to me that there is one type of sound change which > IS irreversible, that is X>0, as found in mergers, phoneme loss, or > assimilation in clusters. Don't languages also add vowels and consonants? Japanese suffixes vowels to many borrowed words. Ditto for Finnish. Turkish prefixes vowels to initial consonant clusters. The general idea is that the phonotactics decides. That means some phonotactical rule of some language (even the same language at a different time) can create the conditions which can result in the addition of a vowel. But then again, a model is needed of language evolution. Lass talks about a lot of this in his book. There are things we can think about just from the way things behave in nature. Gell-Mann discusses this in the Santa Fe book. > Directionality does not mean teleology. To say that sound change moves > in a non-random direction is not to say that it is going anywhere in > particular. One might propose (although I didn't) that over time Maybe there is. Dynamical systems have behaviors which can be used as models. For example, the concept of stability. Languages might change very slowly after they have essentially reached stability. That could explain why some languages seem to change fast and others slowly. > phoneme inventories get smaller, or that 'marked' phonemes tend to be > lost in favor of unmarked. I have yet to see a teleological proposal of > this type which can survive the test of the evidence. The directionality > trends discovered by Labov appear ultimately due to physiological > constraints both articulatory and auditory. If that can be shown to follow from physical and auditory causes so much the better. In other words the direction > of sound change is constrained by the present not by the future. I > suppose this is what you mean by talking about local versus universal > (i.e. short term vs. long term). There are no long term directional > trends (as far as the evidence now goes). But we cannot conclude from > this that the sum of short term directional trends adds up to zero or no > directionality or random directionality. No, local means restricted to a small region of space or time. If humans have been speaking for 500,000 years, a 5,000 year span is very small. > The prequisite to a realistic probalistic model of language change > is a systematic research program in diachronic typology. We have to > establish the probable direction of change on the basis of statistical > data from observed changes-- not on the basis of a priori reasoning. Of, course. But even statistics is built on models. > suspect the ultimate model will look something like those typhoon maps > we see here (don't know if you have them in other countries): There is a > circle showing where the storm is and from the outer edges of the circle > there is a wedge shaped projection showing the area into which the storm > is likely to move. At the big end of the wedge there is another bigger > circle showing where the storm is likely to be at the end of a certain > period (24 hrs, e.g.). From this big circle another wedge is projected, > with a bigger mouth showing where the storm should go in the next > period, and so on. Well, that is dynamics already. And there are models (equations) that will do those things. The zeroth order approximation is the asexual family tree model. Something beyond this seems desirable. What kind of a model should it be? > > > The same applies to linguistics changes.There are many scales at which > > changes occur, and if we mix up these levels we create opposing ideas. > > A priori we don't know what models or scales apply to linguistic > changes. We can't know until we have tried to apply them, that is test > them against the data. Yes, and if the present model says that the subtratum has no effect then there's no way to test it against data. The reason is that the neutral data is interpreted to always confirm the status quo. It is circular. Substratum has no effect. The change was internal. Proof: no words from the substratum. If we don't know the substratum how can we tell if there was no effect. Similarly, how do we reach the conclusion that the subtratum had no effect, if we do not consider any change to be due to the substratum but always internal? > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe > Senior Lecturer, Arabic and Linguistics, > Dept. of Linguistics and Information Science > Tokyo University of Foreign Studies > Nishigahara 4-51-21, Kita-ku > Tokyo 114 Japan -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:48:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:48:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > Could you cite an example or two? I don't mean of 1. or 2. in use > separately, but of the two used together as a fallacious syllogism. There are too many cases. I don't feel like embarrasing people and making it worse. > > But borrowings also create regular sound correspondences. > > Yes, just as things other than measles can produce spots on the face. > We need to take such things into consideration if we want to raise > our competence in historical linguistics (or medical diagnosis) from > this very rudimentary level. But that is not all. Human family members resemble each other. That does not mean that unrelated people cannot resemble each other. And despite the fact that we know both we still consider two people who resemble each other to be related unless there's proof to the contrary. We arrive at this through experience. We see families (which we can confirm) resemble each other and therefore create a general inductive rule. For measles, doctors know very well what healty people look like. How many language families has any human experienced? I do not mean the purported/alleged language families. If a human could be created who could live 100,000 years or more and if we can send him to the past to learn dozens of languages, then he would have 'experienced' language families like human families and the way doctors (and others) have the experience of knowing what measles does. No such thing can be done in linguistics so all of it is based on analogy to models from the rest of the world, such as Linnaean trees, etc. That is coupled with some intuitive calculation of whether the resemblence is due to chance. > Well, we do in fact have records of various language families. What > are you trying to say here? There we go again. Do we know these families like we know human families or is this based on some calculation that the occurences cannot be due to chance? > > > 1. These languages have too many things in common. IOW, there are many > > words in > > these languages which can be made to look like each other with similar > > meanings > > and which could not be due to chance. > > > > 2. If that is not due to chance then either they got these words from > > each other > > or the words are all descended from a common language. > > > > 3. We have plenty of evidence (what?) that these languages did not get > > these > > words from each other. > > > > 4. Therefore these words in these languages must all come from an > > earlier common > > source. > Yes, problems crop up and arguments occur, to be sure. > > I recognize 1-4 as a rough outline of the reasoning by which one > arrives at a hypothesis of genetic relatedness among languages. > Rather than argue about details, I'd like to know where you're going > with it. Are we finished with the idea that it's logically circular? But it is not finished. The key here is that we have to know what is due to chance and what is not. Otherwise we can be creating an argument like this: Well, this mathematical method says that X and Y are related but I know that they are not, so the mathematical method is wrong. There are people (yes, real people, and linguists too) who do this. In fact one of the superstarts of sci.lang and linguistics actually argued exactly like this in email to me. > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:49:23 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:49:23 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > My view: > > That "reconstruction" is (or even could be) a matter of algorithm, or > rule is a widely held misunderstanding outside the field, no doubt > supported by a misinterpretation of the technical term "reconstruction" > in its ordinary sense. We cannot literally "reconstruct" anything. A > "reconstruction" is a hypothesis, neither more nor less. Like any > Of course, there is no proof even in physics. I state this regularly. hypothesis in any field of science we arrive at it by guesswork, by > intuition, by imagination, by accident. There is no path of deductive > reasoning, no "discovery procedure", no algorithm which leads from the > data to the hypothesis. The rigor in historical linguistics as in the > natural sciences is not in the way in which hypotheses are reached but > in the way in which the hypotheses are submitted to the test of the > data. A good reconstruction should be testable in principle-- it should > be a specific hypothetical prediction about the way a particular > language was spoken by a particular group of people at a particular > point in time. A reconstruction is virtually never (directly) testable > in practice-- it is so only on those extremely rare occasions when new > texts of ancient languages are unearthed. That is lots of evidence for correctness but not proof. If some new language that was unearthed was remarkably like someone's prediction, that is cause for celebration. How do we test it then?-- By > implication. Each reconstruction (of a proto-phoneme for example) has > implications for the whole system of the proto language (the whole > phonological system, eg), for the development path leading from the > proto-language to the attested languages (the sequence of sound changes, > eg), and for the forms of the reflexes in the descendant languages. > Only the last is directly observable, of course, and only this real data > can be used to rule out a proposed reconstruction absolutely. But here is where the iteration comes in. The first attempt at reconstruction of a protolanguage *X will be based on N languages. If we add the (N+1)st language then *X might have to be changed. We might find another language y to add to the family. How many correspondences do we need? Even worse, if the similarity of the language to other languages is not considered, it will be added to the most established, largest family, and it will continue to snowball. If you want to see if A is more like B or like C you have to have both the compare. This is the "forced binary discrimination" test which is often used (in phonology and phonetics in lingistics). But then those languages that got writing first keep piling up everything because of these factors. > proposed development path implied by the reconstruction cannot be tested > as right or wrong but only as plausible or implausible based on the > statistical frequency of observed patterns of change. To me that says that we need numbers so we can compare languages to others. That means standardization. > implications which a reconstruction has for the whole system can only be > tested under the assumption of the "uniformitarian principle" - the > assumption that prehistoric languages were not fundamentally different > in kind from attested languages and hence should not show structural > anomalies of a type not found in attested languages (languages shouldn't > be reconstructed with no vowels, or all verbs, for example). > Probability and statistics is involved in all of this, and much could be > gained from making it more explicit. Much could also gained if > historical linguists made an effort to base assumptions about plausible > directions of changes on an explicit, accesible, statistically > analyzable body of evidence of changes, rather than on implicit personal > knowledge. Amen! > -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mccay at redestb.es Thu Nov 12 12:50:05 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 12 Nov 1998 07:50:05 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: >The phenomenon does extend to "apitxat" Valencian Catalan, i.e. the >central Valencian dialects of Valencia city and province (not N and S >Valencian of Castello' and Alacant provinces). I'm glad you have introduced some information about "apitxat". As it happens, my companion is a speaker of apitxat, but a non-native one (she is Basque but grew up in the city of Valencia), and she never gives me clearcut data because she is insufficiently confident in her competence in it, and I don't know any other apitxat speakers!!! Also, I get the idea there is some sort of stigma attached to apitxat speech within Catalan, which could be producing unwanted interference. So anyway, we add one dialect of Catalan to the list of varieties without /z/ in the peninsula (the others being the whole of Castilian and (virtually) all of Galician). (And Asturian?) Now: are there any interesting sibilant-transforming phenomena in apitxat, I wonder? (1) Does /s/ voice before voiced consonants? (In Castilian and Galician this is the only context in which we find [z]. My money is on the same thing happening in apitxat.) (2) Does /s/ rhotacize? Does it aspirate? Does it drop? (I'm guessing no to all three...) In fact, the only phenomenon of this type that I have heard of anywhere in Catalan is rhotacism in varieties of Mallorcan (as pointed out by Carrasquer earlier). Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From wbehr at rullet.leidenuniv.nl Thu Nov 12 12:50:35 1998 From: wbehr at rullet.leidenuniv.nl (WB (in Frankfurt today)) Date: Thu, 12 Nov 1998 07:50:35 EST Subject: Ket-Na-Dene affiliation? Message-ID: ----------------------------Original message---------------------------- Just a small note of caution, re: >The Yenisseyan-Na-Dene affiliation is subscribed to mainly by Sergei >Starostin and his Moscow-based school of Nostraticists. (Stefan Georg) My understanding from personal conversations with Starostin last year is that he himself remains non-committal as far as the proposed Ketic-Na-Dene affiliation is concerned, and that he was not particularly amused when Ruhlen arbitrarily added Na-Dene comparanda to his translation of Starostin's original 1982 _Ketskij Sbornik_ article in _On the Origin of Languages_. Ditto for Sumerian, Kusunda and Burushaski, which are all ad- ditions by Bengtson, Blazhek, Ruhlen, Boisson (?) et al., not by Starostin, as far as I can see. It is true though, that Starostin's collaborator on NC, S.L. Nikolaev subscribes to the existence of "Sino-Caucasian Languages in America" (see his articles in Shevoroshkin ed. 1989, 1991) but that doesn't necessarily mean, of course, that _all_ "long-rangers" at Humanities U. in Moscow would agree upon that. For detailed background information on Sapir's view of the Na-Dene -- ST relationship, cf. Alan S. Kaye, "Distant genetic relation- ship and Edward Sapir", _Semiotica_ 91 (3-4): 273-300 (see also J.D. Bengtson, "Edward Sapir and the 'Sino-Dene' Hypothesis",_Anthro- pological Science_ 102 (1994): 207-230). Judging from Kaye's careful account, it would seem that Sapir's initial enthusiasm for the Sino-Athapaskan connection in the early 20ies gradually bleached, possibly under the influence of his student Li Fang-kuei, who was certainly more qualified than any other scholar around to judge _both_ sides of the comparison. (Indeed, one letter to Sapir suggests that the main motivation for accepting FKL as his student was precisely Sapir's interest in verifying the ND-ST hypothesis!). A decade later, in a letter to Nick Bodman (July 21, 1933), Sapir had already become slightly more cautious, saying: "It is true that some years ago I announced that the Nadene group of American Indian languages might be remotely related to Sinitic (Indo-Chinese) group of languages. I still believe this is true but have so far not prepared my notes for publication". Notice however, that Sapir's original motivation for the assumption of a genetic relationship was _not_ exclusively lexical/typlogical but _morphological_. In a letter to Berthold Laufer of 1921 he writes "I have at last found what I had been looking for some time now, namely that in Nadene as in Indo-Chinese there is an alternation of unaspirated and aspirated consonants, the latter of which have causative value." It would thus be interesting to hear more about the diachronic sources of this particular causative formation from a specialist in "Na-Dene" (as well as the current state of opinions about a "Na-Dene" +/- Tlingit itself etc.), especially since the Sino-Tibetan distinction is now com- monly accepted to go back to *s-prefixation of the root. In any case, I really wish someone could edit Sapir's manuscript "Atha- baskan Dictionary, part 2, Sino-Nadene" (preserved in the American Philosophical Sociery library), since his letters to Berthold Laufer, most of which have been published by Hartmut Walravens in the meantime, do not go beyond some rather superficial remarks. Finally I wonder if someone has ever read Laurence Farget's M.A. thesis _Na-Dene and Sino- Tibetan: Historical linguistics and new data towards establishing genetic relationship_ (Lyon 1986), or would know, how to get a copy of it. Best wishes, Wolfgang Behr Kaye quotes It forms part of a >much larger grouping, called Dene-Caucasian, comprising no less than the >following languages and families: Basque, Iberian (yes, Iberian, though >nobody knows anything about Iberian !), Burushaski, Yenisseyan, >"North-Caucasian" (comprising NW- and NE-Caucasian which are not generally >grouped together by Caucasianists), Hurro-Urartean (lumped with N-Cauc.), >Kusunda (an extinct language of Nepal, on which little is known; Bengtson, >another supporter of this grouping, recently informed me that Kusunda has >been taken out of the grouping; I'm inclined to call this a step in the >right direction ;-) and Na-Dene (I'm currently unsure whether further North >American lgs. have made it into the family yet), and, sorry I forgot, the >whole of Sino-Tibetan, and of course Sumerian. >It is clear that one of the goals of this grouping is to hoover up most >languages of the Old World, which are currently thought to be isolates. >Larry Trask has shown on numerous occasions that (the supporters of) this >theory treat(s) Basque data in a less than competent manner; the same can >be said about Yenisseyan and much of the Tibeto-Burman (part of >Sino-Tibetan) data I've seen in connection with this theory. >Parts of this giganto-macro-grouping have some history, though: a >Yenisseyan-Sino-Tibetan connection has been en vogue in the earlier days of >Yenisseyology with investigators like Donner, Bouda and the outsider >Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially >Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as >I remember this was based on typological resemblances only), Sumerian has >been compared to pretty much everything, so has Basque, but I think it is >safe to say that a Basque - (unspecified) Caucasian connection has probably >lured more early researchers than anything else (probably originally >instigated by ergativity, which once made up for a quite exotic >look-and-feel of a language - which is hardly the case today). >Readers may already have inferred from my slightly ironic tone that I'm >personally disinclined to buy much of this (I have working experience with >Tibeto-Burman, Yenisseyan, NW-Caucasian, NE-Caucasian and a bit of >Burushaski). > >St.G. > >Stefan Georg >Heerstrasse 7 >D-53111 Bonn >FRG >+49-228-69-13-32 > From mccay at redestb.es Thu Nov 12 12:50:55 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 12 Nov 1998 07:50:55 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- B. Wald says: >It just happened -- because both directions >are possible (under certain conditions -- certainly NOT ****z > s /V_V). That provoked me to try and think of counterexamples to Benji's latter assertion here. I couldn't come up with an example of z > s /V_V, but I did find one of +voice > -voice /V_V. Furthermore, if I am right in thinking that the "devoicing" is an innovation (and thus did happen) rather than a survival (i.e. unless there was really no change; see the end), then it provides as good an example of X>Y>X as what Benji called "Central Spanish" s > z > s. (Okay, I'm going to be a pain and say it: just what is "Central Spanish" supposed to mean? The phenomenon in question includes the WHOLE of Castilian (i.e. Spanish) in its territorial extension. Are you implicitly calling Portuguese, Catalan and perhaps Basque "peripheral Spanish"?? Sorry, I couldn't hold that one back. May I suggest that perhaps Benji meant something like "Central Ibero-Romance"?) What I have in mind is a dialectal phenomenon within modern Welsh. First the relevant background information: In modern Welsh word stress is normally on the penultimate syllable, and the phenomenon I will describe occurs in the consonant following the stress when another syllable follows (as it usually does except in stressed monosyllables, of course), i.e. normally at the boundary before the last two syllables of a word, for example where the /d/ is in: /'kadair/, /'kader/ or /'kadar/, orthographically , 'chair' < Latin "cat(h)edra" etc. In this post-stress intervocalic position, in standard Welsh, the consonant may be phonetically simple, as in standardly pronounced ['ka:der] (or in slow speech, ['ka:dajr]) or geminate, as in: /'etto/ 'yet, again' The distribution of simple and geminate consonants in this environment is largely predictable (e.g. "voiced" stops are simple, "voiceless" stops are geminate), as is the distribution of short and long vowels in the stressed syllable (long preceding simple consonants, short preceding geminate ones), for which reason the modern orthography reflects neither consonant length nor vowel quantity in these cases. But there are some minimal contrasts too, e.g. 'sing' versus 'whiten'. (Caveat: I have put quotes around "voiced" and "voiceless" because phonetically all stops in Welsh tend to be voiceless (or at least voicing is not critical) and the contrast is realized principally in terms of tenseness and aspiration. To simplify the exposition I shall henceforth largely ignore that fact in the transcriptions and terminology used.) That is in STANDARD Welsh. Now in Northern spoken Welsh, this system has been altered, in that nearly all consonants in the intervocalic post-stress position are pronounced geminate (and correspondingly, all stressed vowels followed by a consonant in non-final syllables are short - but I'm going to focus on the consonants here). So corresponding to standard ['ka:dajr] we will find ['kaddar] in northern Welsh, while ['etto] shows no change since the consonant is already geminate. Long vowels in northern Welsh are found mainly in monosyllabic words (/ka:n/ 'song') and in penultimate syllables when in hiatus with the following vowel (/'diod/ [di:od] 'drink'); otherwise all vowels are short. In northern Welsh, then, and are both ['kannI]. Of these two systems, the standard and the northern, the standard one must represent a diachronically prior stage, for a number of reasons that I won't go into, but some of which are already fairly obvious. Now: apart from this gemination and shortening, but possibly related to it, we find some varieties of Welsh in which, corresponding to consonants like the /d/ in , if the consonants in question are "voiced" stops in standard Welsh, we find "voiceless" stops instead: ['katar] etc. Assuming once again that the standard form is diachronically prior, we then have d > t (or possibly d > dd > t) in post-stress intervocalic position, and similarly b > p and g > k. Irritatingly, this phenomenon has been noted in the south, in an area not contiguous to the northern "geminating" varieties. By the way, we might assume that the same varieties would have had z > s intervocalically, if only Welsh possessed a /z/ phoneme, which it doesn't; such is life! Then again, maybe z > s wouldn't have happened, since the voiced fricatives /v/ and /D/ do NOT undergo devoicing in this context. Tant pis! (Actually, these consonants also seem to resist gemination in the "geminating" northern dialects, even though the preceding stressed vowel is STILL short. That's the reason why I said "nearly all" three paragraphs back.) As may have been noticed, however, the standard Welsh /d/ which in some dialects has become /t/ (by the account I just gave) itself derives ultimately from an earlier /t/. In the evolution from proto-Celtic (or Latin, in the case of loanwords) to modern Welsh, intervocalic voiceless stops were regularly voiced, much as in western Romance languages. The modern intervocalic voiceless stops come from original geminate stops - which explains their gemination in the modern language, as well as the lack of it in the voiced counterparts in standard Welsh. So in an earlier stage of the development of Welsh we have approximately: VttV > VttV VtV > VdV and similarly for /p/ and /k/. In northern Welsh ['kaddar] we then have a further step: 'VCV > 'VCCV (hence 'VdV > 'VddV) while in the varieties that now pronounce ['katar] we presumably have a later reversal of t > d: 'VdV > 'VtV Unless, of course, 'VtV in these varieties is actually a SURVIVAL of Celtic/Latin VtV. Would anyone more expert than myself like to take up the discussion from here? NOTE: For anyone who wants to know the detailed diachronic background of Welsh phonology (in general), I recommend the classic historical grammar by Morris-Jones. From jhewson at morgan.ucs.mun.ca Thu Nov 12 21:20:10 1998 From: jhewson at morgan.ucs.mun.ca (John Hewson) Date: Thu, 12 Nov 1998 16:20:10 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Tue, 10 Nov 1998, Larry Trask wrote: > Oh, there are certainly rules, but there are no algorithms. For > example, given the observed variant forms of the Basque word for `ear', > there is only one reconstruction that obeys all the rules. > > > My original question is/was why there is no algorithm for producing > > protoforms. Either what you practice is a science or it is not. If > > it is magic you don't have to explain it. If it is science it should > > be possible to see it in writing in some book. Actually there is an algorithm that a Systems Analyst and I developed to produce my _Computer-generated Dictionary of Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be stated in a single sentence, as on p.iv of the above: From the data of the daughter languages generate all possible protoforms, then sort alphabetically, and examine all sets of identical protoforms collocated by the sort. The methodology has now been streamlined by John Lowe and Martine Mazaudon, who reported on it in an article in the Bulletin de la Societe Linguistique de Paris. Their "Reconstruction Engine" even reconstructs tones, and can be adapted to the use of any language family. The existence of these algorithms is a massive demonstration of the regularity of sound change. But the final product still has to be worked on by the linguist, given the nature of human languages. There may be more than one possible reconstruction thrown up by the algorithm, where perhaps one or more definitive item of evidence is lacking. If this further evidence can not be obtained by subsequent research, the reconstruction must remain tentative (the same is true for all reconstructions, of course, machine-generated or not). John Hewson, FRSC tel: (709)737-8131 Henrietta Harvey Professor of Linguistics fax: (709)737-4000 Memorial University of Newfoundland St. John's NF, CANADA A1B 3X9 From delancey at darkwing.uoregon.edu Thu Nov 12 21:19:33 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Thu, 12 Nov 1998 16:19:33 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, WB (in Frankfurt today) wrote: > do not go beyond some rather superficial remarks. Finally I wonder if > someone has ever read Laurence Farget's M.A. thesis _Na-Dene and Sino- > Tibetan: Historical linguistics and new data towards establishing genetic > relationship_ (Lyon 1986), or would know, how to get a copy of it. I have a 13-page "abrege" of the thesis ("Na-Dene et sino-tibetain: synthese et nouvelles donnees pour une eventualle parente"), but it is just a brief literature review, with no new evidence presented. I too would like to see the complete thesis. For some reason she seems to attach some importance to Benedict's reconstruction of a two-tone system for Proto-Sino-Tibetan, which is not a very popular idea within the field--I think most Sino-Tibetanists currently reconstruct PST without tone. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Thu Nov 12 21:19:10 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:19:10 EST Subject: Hawaiian meli In-Reply-To: <3648FDB6.2C7D5309@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, H.M.Hubey wrote: > How are we to take into account all of those words that are said to > be cognate in IE when it is quite possible that many (most) might be > left over a substratum that was living in that neighborhood for many > thousands of years and had stabilized so that changes were taking > place very slowly? First of all, languages do not "stabilize". Every living language is always changing, and the only language that doesn't change is a dead one. Now, if some substrate words had managed to enter the ancestor of PIE, that wouldn't matter. A PIE word is a PIE word, regardless of its ultimate origin, and all we are usually trying to do is to reconstruct PIE, and not to go further back. Only if we *do* want to go further back do we need to worry about loan words into Pre-PIE. As for loan words into an already existing PIE, well, that's a problem familiar to all historical linguists. We always have to worry about the possibility that some words in any given language might be loan words. But PIE happens to be a more than averagely convenient language in this respect. PIE words tend rather strongly to adhere to certain patterns of formation, perhaps most typically Root-Formative-Suffix, with all three elements recurring in other words. There is no reason to expect a loan into PIE to conform to such patterns, and hence we perhaps have a better than average chance of spotting loans into PIE. As for the suggestion that PIE consisted mostly of elements from another language, well, that's rather reminiscent of the famous claim that the Iliad and the Odyssey were not written by Homer, but by another poet of the same name. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From sally at isp.pitt.edu Thu Nov 12 21:18:48 1998 From: sally at isp.pitt.edu (Sarah G. Thomason) Date: Thu, 12 Nov 1998 16:18:48 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Your message of "Wed, 11 Nov 1998 13:20:34 EST." Message-ID: ----------------------------Original message---------------------------- One additional problem with the Na-Dene comparison, of course, is that the inclusion of Haida in Na-Dene is (to put it mildly) not universally accepted by specialists. See, for instance, Robert Levine's article in IJAL on the subject some years back -- an article that was attacked by Greenberg in a whole chapter, in his 1987 book on classification of Native American languages. But Levine's article is good nevertheless, and Greenberg's arguments against Levine aren't impressive. -- Sally From larryt at cogs.susx.ac.uk Thu Nov 12 21:18:21 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:18:21 EST Subject: rhotacism from Ray Hickey In-Reply-To: <364969B6.511AF413@eucmos.sim.ucm.es> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Javier Martinez wrote: [LT] > > But some sound changes are quite irreversible. Consider loss. In the > > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting > > /h/ was later lost. I predict confidently that the Greeks will never > > reverse this change by re-introducing those long-gone /s/s, > > yes. as a product of analogy, see the -s- futures etc. Yes, agreed, except that specialists do not seem to be sure whether /s/ was first lost from futures like (from `loosen') and then restored by analogy, or whether it was never lost in the first place because of paradigmatic pressure to retain it. As usual, we cannot tell without textual evidence whether we are looking at an instance of analogical restoration or at an instance of analogical preservation. Analogical preservation is quite well attested. For example, some of the Finnic languages have lost word-final /n/ in all cases *except* where that /n/ was the sole marker of a grammatical category, in which case it has been retained. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Thu Nov 12 21:18:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:18:09 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36557252.64845784@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Miguel Carrasquer Vidal wrote: OK. I'm going to try to catch up a little here. > "H.M.Hubey" wrote: > >Larry Trask wrote: > >> `female' (which itself is borrowed from Occitan), plus <-kume> > >And where is that borrowed from? It is not borrowed at all, but native. Basque means `child, offspring'. This is a native word, reconstructed as *, and apparently attested in the Aquitanian ancestor of Basque as both OMBE- and VMME. This word often occurs as the second element in compounds, in which position it occurs as <-kume>, with an initial /k/. Nobody knows how this /k/ arises, but it is not unique to this morpheme. Very commonly, a vowel-initial Basque word acquires an initial /k/ or /t/ when it serves as the second element in a compound. The origin of this plosive is only partly understood, but in any case it is normal in Basque, and the /k/ in requires no additional explanation. This <-kume> is frequent: `cat', `kitten' `sheep', `lamb' `pigeon', `baby pigeon' And so on, for many examples. The other phonological developments seen here are all regular. The word `woman' is seemingly of rather recent origin in Basque: it has only a single attestation before the 18th century, since when it has become the usual word for `woman', except in the Salazarese dialect, in which it means `girl', which we believe to have been its original meaning, since the etymology is * `female offspring'. The earlier word for `woman' was , which today has been specialized to `wife', except in Salazrese, where it still means `woman'. It will surprise no one to learn that both and are compounds built upon the word `female', whose regular combining form is (look at two of the examples above). We suspect an etymology * `young female'. And, as Miguel C V has pointed out, this is borrowed from Gascon , itself from Latin . Basque contains many loans from Gascon. The /mn/ cluster has been intolerable in Basque at all periods, and, in loan words containing it, it is always reduced to either /m/ or /n/. > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), emzirik, etc etc. The Sumerian word is not relevant, and it has been effectively disposed of by Miguel. Turkish does not mean `mother'; it means `but'. The Turkish word for `mother' is in Anatolian Turkish but in standard Istanbul Turkish, this last apparently being an expressive variant of . Turkish `nipple' is not available for comparison. One of the best-known facts about Turkish is that native Turkish lexical items do not begin with /m/ unless they are imitative words or nursery words. In all likelihood, we are looking at a nursery word here, and nursery words cannot be cited as comparanda, because they are so often created independently. The stem `suck' is the source of the derivatives `nipple', `nipple', and `suckle'; these can only be counted as one word. Finally, is merely a diminutive of `vulva', containing the usual Turkish diminutive suffix <-cik>. Hence all we have here for Turkish is a verb-stem `suck' and a noun `vulva'. And nothing whatever can be concluded from this. It's every bit as impressive as English `ear' and `hear'. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 21:14:45 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 16:14:45 EST Subject: intervocalic DEvoicing can also happen / X > Y > X In-Reply-To: <3.0.3.32.19981112103615.006eaf48@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >As may have been noticed, however, the standard Welsh /d/ which in some >dialects has become /t/ (by the account I just gave) itself derives >ultimately from an earlier /t/. In the evolution from proto-Celtic (or >Latin, in the case of loanwords) to modern Welsh, intervocalic voiceless >stops were regularly voiced, much as in western Romance languages. The >modern intervocalic voiceless stops come from original geminate stops - >which explains their gemination in the modern language, as well as the lack >of it in the voiced counterparts in standard Welsh. So in an earlier stage >of the development of Welsh we have approximately: > >VttV > VttV >VtV > VdV > >and similarly for /p/ and /k/. In northern Welsh ['kaddar] we then have a >further step: > >'VCV > 'VCCV (hence 'VdV > 'VddV) > >while in the varieties that now pronounce ['katar] we presumably have a >later reversal of t > d: > >'VdV > 'VtV > >Unless, of course, 'VtV in these varieties is actually a SURVIVAL of >Celtic/Latin VtV. I'm far from an expert on Welsh, and I'm not even sure I have the basic facts right. What I gather from the above and from Paul Russel's "An introduction to the Celtic languages" is that: -d- > -D- Adam > Addaf -t- > -d- carita:tem > cardod -tt- > -T- cattus > cath Apparently, new geminates -tt- arose from syncope, giving -d- (cadair) and -tt- (eto /etto/) as the only allowed intervocalic stops. Now the lenitions of -d- to /D/ and -t- to /d/ seem to be the oldest. It is also reasonable to suppose that after they had taken place, geminate /tt/ in /kattu(s)/ was simplified to /katu(s)/ (or directly spirantized to /kaTu(s)/), and did in any case not merge with the new /tt/. If -t- in the /katar/ varieties is a survival of Celtic/Latin -t-, then that must mean that *tt did not pass through a stage *t, but spirantized directly, at least in those varieties: d > D t > t tt > T -- > tt As opposed to the rest of Welsh: SW NW d > D D D t > d d dd tt > t > T T T -- > tt tt tt But that would put the /katar/ varieties of Welsh above Cornish and Breton in the branching tree, which doesn't seem very likely. So it must be: SW NW tW d > D D D D t > d d dd t tt > t > T T T T -- > tt tt tt tt What's missing in the above is a variety of Welsh that has eliminated geminates: D, d, T, t. One would expect one. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From johanna at uclink.berkeley.edu Thu Nov 12 21:14:14 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Thu, 12 Nov 1998 16:14:14 EST Subject: Yeniseian and Na-Dene In-Reply-To: <364A801F.95E8163A@montclair.edu> Message-ID: ----------------------------Original message---------------------------- Mark Hubey writes: >Johanna Nichols wrote: >> >> I've worked out the chances of finding words with two similar consonants in >> the same order, with similar but not necessarily identical meanings, in two >> languages. Out of a fixed list of 100 meanings chosen in advance (this is >> analogous to asking "what is the probability that similar forms will mean >> 'water' in both languages?", and so on for another 99 glosses), this is how >> many resemblant sets it takes to exceed the range of chance and show that >> relatedness is likely: >> >> 2-consonant words with the very same meaning: 7 >> 2-consonant words with similar meanings (modeling this as a search that >> allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or >> whatever; these must also be specified in advance): 25 >> 1-consonant words (or 2-consonant words with one resemblant consonant and >> one non-resemblant one) with the same meaning: 27 >> 1-consonant words with 5 senses' leeway each: over 50 > > >I erased the rest not because it is not important but because I want to >ask about this. > >Is it not true that the most important consideration in probability >theory is knowing the sample space? > >In other words, when "matches" due to chance are being calculated, >should not the fact that the two languages have (or seem to have) >the same set of phonemes enter into the calculation? In other words, >the sample space should consist of the phonemes that the languages >could have had (along with the phonemes that they do have) but do not? > >The calculations should involve conditional probabilities. No? > >Secondly, I also made some calculations. But mine is not for phonemes >and does not take into account phonemes for the reason that they cause >more complications, and do not take into account that the same speech >space available for humanity is divided up differently and into >different >number of chunks (phonemes) in different languages. The fact that out of >possible M phonemes if languages seem to have a particular set of N >phonemes that in itself has to be accounted for. > I have two ways of computing the probability of a generic consonant. (1) Languages (as sampled in my database) have an average of about 20 consonant phonemes. One of them has, on average, 0.05 chance of occurring in a randomly chosen position in a randomly chosen form. This is the probability of a specific consonant. Allowing three trials (allowing a search through two to three distinctive features' space, or about three phones) yields a probability of 0.143. (2) Whatever the consonant inventories of the languages under comparison, divide each of them into 7 phonetically coherent spaces. (Variant: divide them into 6 coherent spaces, and count lack of any consonant -- e.g. initial V rather than C -- as a seventh possibility. This makes it possible to accommodate Rotokas, with its 6-consonant system.) This way too we get a probability of 0.143 (1/7 = 0.143). Neither of these procedures guarantees fair coverage of vastly different frequencies of different consonants, language-specific or family-specific preferences of different consonants or consonant classes for different phonotactic positions, and the like. I hope that some of these differences get ironed out by putting consonants together in phonetic groupings. Still, the metric is only approximate. It enables us to point out that 36 resemblant sets, half of them with only one resemblant consonant, isn't enough to indicate genetic relatedness unless a very small wordlist was specified in advance. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From larryt at cogs.susx.ac.uk Thu Nov 12 16:37:20 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 11:37:20 EST Subject: X>Y>X In-Reply-To: <364A7E48.2595E67@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, H.M.Hubey wrote: > If there are people who are writing programs to paint (yes, produce > art), and compose music, it takes no genius to see that even if > Starostin only wrote (or got a student to write) a brute-force, dumb > program on a commodity grade PC, he can uncover relationships that > many humans cannot do, even if they collaborate. The reason for > this will take too long to explain. But given a set of words (and > their meanings) even a brute-force program can keep cranking 24 > hours a day to produce cognates via regular sound changes, clusters, > and things that a typical linguist does not even know exists. No; this is not remotely so. The fundamental problem here, I think, is this. Such a dumb brute-force approach is obliged to treat all data on an equal footing. But the first thing you learn when you take up historical linguistics is that you *cannot* treat all data on an equal footing. Anyway, I might point out that just such brute-force programs already exist, that they have already been developed to a certain level of sophistication beyond the maximally dumb, and that they have already been applied to a number of individual cases. The ones I know most about are those developed at Cambridge, and these are interesting. However, these programs, interesting as they are, have certain inherent weaknesses. First, they cannot prove a linguistic relationship. At best, they can conclude that a genetic relationship is likely at the confidence level of 95%, or 99%, or whatever. And even these impressive-looking levels generally only arise in cases in which linguists have already established that a genetic link exists. Second, and more seriously, they cannot distinguish relatedness from non-relatedness. If you feed in data from, say, English, Dutch, French and Chinese, what you get is a tree in which English and Dutch are the two closest languages, French is somewhat more distantly connected, and Chinese is more distantly connected still. That is, the programs cannot distinguish an unrelated language from a distantly related language. One more thing. One of these programs has the curious habit of reporting a strong link between French and Hungarian at the 95% confidence level or above. Not IE and Hungarian, you understand: just French and Hungarian. Brute force or not, a mere dumb program is capable of reaching conclusions which any knowledgeable linguist knows are just plain wrong. Properly designed programs, in the hands of skilful linguists, are potentially capable of becoming a useful tool -- but certainly not a replacement for ordinary work in historical linguists. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 16:36:38 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 11:36:38 EST Subject: s > r (Iberian) In-Reply-To: <3.0.3.32.19981112015451.006d35bc@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Now: are there any interesting sibilant-transforming phenomena in apitxat, >I wonder? > >(1) Does /s/ voice before voiced consonants? (In Castilian and Galician >this is the only context in which we find [z]. My money is on the same >thing happening in apitxat.) > >(2) Does /s/ rhotacize? Does it aspirate? Does it drop? (I'm guessing no >to all three...) My direct experience with apitxat is minimal, but I guess your guesses are correct. I'm still wondering where my non-voicing of /s/ in Castilian before /n/, /m/, /l/, /r/ comes from. Aragonese? That's where my father's from. I've just remembered a case of Catalan zetacism: the word "cupboard", vulgo [az'mari]. Could be hypercorrection, could be dissimilation, as opposed to assimilation in *ciresa > cirera "cherry". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From larryt at cogs.susx.ac.uk Thu Nov 12 16:36:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 11:36:09 EST Subject: the Trask-Hubey debate In-Reply-To: <364A34D2.5287B5D3@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, H. M. Hubey wrote: > That is the real reason I post what I do. There are computer programs that > paint and compose music. Is it really that hard to believe that linguistic > reconstruction is no less structured? There is no comparison. *Any* piece of paintwork we might produce counts as a painting. And *any* piece of music we might produce counts as music. But it is *not* true that any "reconstruction" we might produce counts as a reconstruction. A linguistic reconstruction is an attempt at recovering a real but unrecorded piece of prehistory, and it is not an attempt at producing an original work of art. I have little doubt that it would be possible to write a computer program that would chomp its way through any pile of linguistic data we chose to dump into it and spit out some kind of result, according to its instructions. But I see no reason to suppose that such a result would be anything but meaningless. > I bring this up, because for a long time the anti-AI crowd used > arguments similar to those offered often on linguistics lists for > why AI would be impossible. I do not believe this is true. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 16:14:33 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 11:14:33 EST Subject: s > r (Iberian) In-Reply-To: <3.0.3.32.19981111105818.006e93e8@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >the intervocalic >s > z development, general in Romance (all? most?), was later reversed in >Castilian and Galician, as opposed to most other Romance languages. It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, -g-, which must mean that the change was general in most Western Romance, but did not occur in Eastern Romance (S. Italian and Romanian). The exceptions in Western Romance are Mozarabic (partially, and as far as this can be determined) and Aragonese, or at least part of it. In High Aragon we see cases of maintained intervocalic -p-, -t- and -k-. Intervocalic -s- is /s/ now of course, but I wonder whether we have to assume a phase with /z/ for those Aragonese dialects that conserve -p-, -t- and -k-. The Ribargorc,a dialect of Catalan, adjacent to the Aragonese area, has intervocalic /s/, but it may be safely assumed that this is a recent phenomenon. Ribagorc,a` does not maintain intervocalic -p-, -t- and -k-, and there's even some evidence for rhotacism (-z- > -r-) in toponyms of the Pyrinean area (Glosianes > Glorianes). The standard language also has CERESIA > cirera "cherry", an admittedly isolated case of pure intervocalic rhotacism that hadn't been mentioned yet. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From drc at antnov1.auckland.ac.nz Fri Nov 13 01:17:16 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 12 Nov 1998 20:17:16 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Thu, 12 Nov 1998 07:48:55 EST > Reply-to: hubeyh at montclair.edu > From: "H.M.Hubey" > Organization: Montclair State University > Subject: Re: rhotacism from Ray Hickey > To: HISTLING at VM.SC.EDU > ----------------------------Original message---------------------------- > Ross Clark wrote: > > > > Could you cite an example or two? I don't mean of 1. or 2. in use > > separately, but of the two used together as a fallacious syllogism. > > There are too many cases. I don't feel like embarrasing people > and making it worse. > I must insist. I deny that anybody uses 1. and 2. in this way. Prove me wrong. Embarrass somebody. > > > > > But borrowings also create regular sound correspondences. > > > > Yes, just as things other than measles can produce spots on the face. > > We need to take such things into consideration if we want to raise > > our competence in historical linguistics (or medical diagnosis) from > > this very rudimentary level. > > But that is not all. > > Human family members resemble each other. That does not mean that > unrelated > people cannot resemble each other. And despite the fact that we know > both > we still consider two people who resemble each other to be related > unless > there's proof to the contrary. We do? We arrive at this through experience. We > see > families (which we can confirm) resemble each other and therefore create > a general inductive rule. > > For measles, doctors know very well what healty people look like. > > How many language families has any human experienced? I do not mean the > purported/alleged language families. Since I don't know which entities qualify in your mind as "language families" as opposed to "purported/alleged language families", I can't answer this. However, on my own understanding of what a language family is, any competent historical linguist has experienced (has some knowledge of) a variety of language families and of various languages not known to be related. If a human could be created who > could > live 100,000 years or more and if we can send him to the past to learn > dozens of languages, then he would have 'experienced' language families > like human families and the way doctors (and others) have the experience > of > knowing what measles does. > > No such thing can be done in linguistics so all of it is based on > analogy > to models from the rest of the world, such as Linnaean trees, etc. That > is > coupled with some intuitive calculation of whether the resemblence is > due > to chance. You seem to be suggesting that the empirical base of historical linguistics is too small. Unless you have some realistic suggestion as to how it could be signficantly enlarged, we have to live with it. If that disqualifies it as "science" in your opinion, too bad. > > > > Well, we do in fact have records of various language families. What > > are you trying to say here? > > There we go again. Do we know these families like we know human > families or is this based on some calculation that the occurences > cannot be due to chance? Calculation of whether resemblances could be due to chance or not becomes relevant in distant relationships or borderline cases, about which so much argument goes on now. I'm talking about families even Lyle Campbell believes in, where there is no argument. In some cases (eg Latin and Romance) we have the proto-language through direct documentation. We also have recorded histories of many individual languages which tell us a lot about how languages change. > > > > > > 1. These languages have too many things in common. IOW, there are many > > > words in > > > these languages which can be made to look like each other with similar > > > meanings > > > and which could not be due to chance. > > > > > > 2. If that is not due to chance then either they got these words from > > > each other > > > or the words are all descended from a common language. > > > > > > 3. We have plenty of evidence (what?) that these languages did not get > > > these > > > words from each other. > > > > > > 4. Therefore these words in these languages must all come from an > > > earlier common > > > source. > > > Yes, problems crop up and arguments occur, to be sure. > > > > I recognize 1-4 as a rough outline of the reasoning by which one > > arrives at a hypothesis of genetic relatedness among languages. > > Rather than argue about details, I'd like to know where you're going > > with it. Are we finished with the idea that it's logically circular? > > But it is not finished. The key here is that we have to know what is > due to chance and what is not. Otherwise we can be creating an argument > like this: Well, this mathematical method says that X and Y are related > but I know that they are not, so the mathematical method is wrong. There > are people (yes, real people, and linguists too) who do this. In fact > one > of the superstarts of sci.lang and linguistics actually argued exactly > like this in email to me. This is not circularity. Mathematical methods can be wrong. A number of different proposals have been made for calculating the probability of accidental linguistic resemblances. They give different results. Therefore they can't all be right. They may be free of mathematical error, but they do not necessarily yield historically correct conclusions. On the other hand, one would want to ask the linguist in question just what was the basis of his/her certainty that X and Y are _not_ related. Ross Clark From hubeyh at montclair.edu Fri Nov 13 13:21:42 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:21:42 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Wed, 11 Nov 1998, H. M. Hubey wrote: > > > That is the real reason I post what I do. There are computer programs that > > paint and compose music. Is it really that hard to believe that linguistic > > reconstruction is no less structured? > > There is no comparison. > > *Any* piece of paintwork we might produce counts as a painting. And > *any* piece of music we might produce counts as music. But it is *not* > true that any "reconstruction" we might produce counts as a > reconstruction. Any piece of paintwork does not count as painting. The painters have their own black magic guild which decides what is art and what is not. Does that ring a bell? Besides, art and music were almost the last refuge for anti-AI crowds who thought they finally found something computers could not do but which "intellectuals" could do. Of course, they also deemed that doing math (and physics, and engineering, and compsci and ....) was mechanistic, rote, deterministic, rule-following, algorithmic stuff that computers could do, and "real intellectual activity" was things like playing chess, painting, writing poetry, composing music etc. Now they have no place to hide because what we see is that they are also about rules. And now during the last hurrah of anti-AI frenzy you are picking up the mantle and expect to convince me with decades old arguments which have already been blasted to smithereens. > A linguistic reconstruction is an attempt at recovering a real but > unrecorded piece of prehistory, and it is not an attempt at producing an > original work of art. > > I have little doubt that it would be possible to write a computer > program that would chomp its way through any pile of linguistic data we > chose to dump into it and spit out some kind of result, according to its > instructions. But I see no reason to suppose that such a result would > be anything but meaningless. I have no doubt that I can write a program to do exactly what you claim can't be done, just as I had no doubt that music creation or art creation could be automated. I also have no doubt that people like Starostin already have made much progress with his electronic and machine readable database of languages and etymologies. I also have no doubt that soon lots of others will be doing it. I also have very small doubts that I can write a program that can take n words from language A, and m from language B and write a program that can change one set into the other (at least enough of them to dumbfound the skeptics and force them to having N new looks into the "comparative method") using only regular sound changes. The only problem is that I don't have the time or energy to put into such a useless demonstration. It will probably be done by someone (like Starostin) or someone else who is upset at the way linguists spurn statistics and math. After all, it the only thing that will make people sit up and take notice there is no better way than to demolish their toy :-) > > I bring this up, because for a long time the anti-AI crowd used > > arguments similar to those offered often on linguistics lists for > > why AI would be impossible. > > I do not believe this is true. See, Hubey, H.M. (1996) "Topology of Thought", CC-AI: Journal for the Integrated STudy of Artificial Intelligence, Cognitive Science, and Applied Epistomology", vol 13, No.2-3, pp.225-292. > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mcv at wxs.nl Fri Nov 13 13:22:15 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 08:22:15 EST Subject: Doing historical linguistics In-Reply-To: <364A7E48.2595E67@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: >It should be obvious to everyone by now that Starostin is way ahead of >the game than the 99% of the subscribers. It should be clear by now >that if he has already built up a database of lexicons of various >languages with their meanings, writeable in ASCII, Unicode etc, he has >singlehandedly done what should have been done by the linguistics community. I certainly agree that Starostin's databases are a great resource. But that is not all. >But that is not all. > >If there are people who are writing programs to paint (yes, produce >art), and compose music, it takes no genius to see that even if Starostin >only wrote (or got a student to write) a brute-force, dumb program on a >commodity grade PC, he can uncover relationships that many humans cannot do, >even if they collaborate. The reason for this will take too long to explain. Well, why don't you write (or get a student to write) a brute-force program to explain it? We *should* use information technology to assist research in historical linguistics. If only to nip in the bud attempts to relate Bq. "dry" with something in Caucasian ("Warning 23: oldest attested semantics: "barren, sterile"") or Bq. with words for "woman" sounding like /kwVn-/ ~ /kwVm-/ or Sum. with words for "woman" sounding like /em(e)/ ("Error 09: operation not commutative"). Certainly there are people using programs to paint and make music. But the value of the result depends entirely on the human input parameters, and on human selection/rejection/editing of the output. Even computer art is a craft. It takes a specialist (or a gifted person) to get interesting results. A program (brute-force or otherwise) to assist in reconstruction of proto-languages can be made, but surely to be useful it should incorporate existing knowledge about linguistics in its programming or configuration parameters (likely phonological developments and semantic shifts) and it should be fed reliable and complete data, including morphological information, otherwise it's GIGO. It is an illusion to think that we can throw away two centuries of comparative linguistic practice, and let "logic, probability theory and fuzzy set theory" do all the work. Mathematicians aren't going to be replaced by theorem proving automata any time soon, and neither are historical linguists scheduled to be replaced with "proto-language construction programs". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From hubeyh at montclair.edu Fri Nov 13 13:23:02 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:23:02 EST Subject: domina/womina, haber, bad Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > dialect, in which it means `girl', which we believe to have been its > original meaning, since the etymology is * `female offspring'. > > The earlier word for `woman' was , which today has been > specialized to `wife', except in Salazrese, where it still means > `woman'. > > It will surprise no one to learn that both and are > compounds built upon the word `female', whose regular combining > form is (look at two of the examples above). We suspect an I don't recall writing that it was borrowed into Basque. What I did write, however was that Lahovary (Dravidian and the West) writes about the pre-IE and pre-AA languages of the "Mediterranean" and tries to connect them with Dravidian and Caucasian. That 'em' word shows up in them, with related meanings. von Soden, points out that pre-IE and pre-AA languages of the region confused the liquids /r/ and /l/ (especially in the beginnings of words), etc. Obviously, the languages which do not have words beginning with liquids and which confuse them are Dravidian and Altaic. Obviously all of this matters to some people. > etymology * `young female'. And, as Miguel C V has pointed > out, this is borrowed from Gascon , itself from Latin > . Basque contains many loans from Gascon. The /mn/ cluster kap>capere, ulu/uru > ululare, karnash > carnal, tes > seks, oghlan > clan, and now fem > hem > em, etc. > The Sumerian word is not relevant, and it has been effectively disposed > of by Miguel. Turkish does not mean `mother'; it means `but'. Turkish yes, Turkic no. Karachay-Balkar uses ata, ana, atta, anna, appa, akka, amma,... and most of these words can be found in Turkic languages. > The Turkish word for `mother' is in Anatolian Turkish but > in standard Istanbul Turkish, this last apparently being an expressive > variant of . It is also a part of similar changes. I think Chuvash is also anne or ane. > Turkish `nipple' is not available for comparison. One of the > best-known facts about Turkish is that native Turkish lexical items do > not begin with /m/ unless they are imitative words or nursery words. In > all likelihood, we are looking at a nursery word here, and nursery words > cannot be cited as comparanda, because they are so often created > independently. This statement seems to fly in the face of what is known. Babies can create and do babble every sound there is. That means that there is some reason why parents pick up certain types of words. The most likely explanation is that those words already exist in the language and the parents are happy to hear those words. This means that your argument is backwards (ahistorical accretion) and one (or the only) purpose for its existence is the fact that it makes the IE theory look good and the proto-world look bad. Just recently you mentioned that there are 6,000+ languages. In how many of them are "nursery" words present? The list posted to sci.lang (or some other usenet list) by Miguel listed less than 100 languages. That is 100/6,000 which is 1/60. Since when does 1/60 become the standard? By whose arithmetic? > The stem `suck' is the source of the derivatives `nipple', > `nipple', and `suckle'; these can only be counted as > one word. So there is some value to doing clustering analysis and semantic distances, and Swadesh-like lists after all. > Finally, is merely a diminutive of `vulva', containing the > usual Turkish diminutive suffix <-cik>. > > Hence all we have here for Turkish is a verb-stem `suck' and a > noun `vulva'. And nothing whatever can be concluded from this. > It's every bit as impressive as English `ear' and `hear'. That once again shows the classic method. Take each possible cognate one at a time, and ignore them one at a time. When someone spends from 1947 to 1990 patiently collecting words (Dr. Tuna) and presents it all at once, then what? Change the Sumerian words? Once again this gets us in the thick of things. How many such words are needed? If it happens all the time, why don't linguists list them someplace instead of listing the perennial 'bad' (English,Farsi), haber (Latin, German), domina/womina (Italian,Japenese) etc. Maybe this is the real big reason I want things quantified. IT reminds me of the joke about the accountant. At a job interview he is asked "How much is 2+2?". He says "How much do you want it to be?" Is this how science is supposed to be done, or is this how linguistics "science" supposed to be done (according to you)? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mccay at redestb.es Fri Nov 13 13:23:43 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 13 Nov 1998 08:23:43 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: >Apparently, new geminates -tt- arose from syncope, giving -d- >(cadair) and -tt- (eto /etto/) as the only allowed intervocalic >stops. I accept the correction. I was writing from memory, without the relevant manuals to hand. But I don't think this affects the gist of my argument. >But that would put the /katar/ varieties of Welsh above Cornish and >Breton in the branching tree, which doesn't seem very likely. I think it's pretty obvious that /kadar/ > /katar/ is an innovation, not a survival. I was merely saying that it has to be one or the other (and is, in fact, the one, not the other). >What's missing in the above is a variety of Welsh that has eliminated >geminates: D, d, T, t. One would expect one. I'll say what I know (or think I know) about this. As I said before, except in certain cases (e.g. "canu" versus "cannu"), it is predictable in modern Welsh whether a given intervocalic consonant preceded by a stressed vowel will be geminate or simple, and in the case of stops, the standard rule is voiced -> simple, and voiceless -> geminate. So I think it can be argued that gemination may not be phonemic, except for those "certain cases". Thus [t] and [tt] would be allophones of /t/, etc. Given that situation, the issue seems to belong to the domain of surface phonetics. The difference between southern and northern dialects in this respect would then be that in the south only voiceless stop phonemes have the two allophones (simple and geminate) while in the north this pattern has been extended to include voiced stops. SOUTH (more conservative in this respect): /t/ -> [t], [tt] /d/ -> [d] NORTH: /t/ -> [t], [tt] /d/ -> [d], [dd] hence e.g. southern [etto], [ka:der] vs. northern [etto], [kaddar] for /eto/, /kadVr/. Not being a phonetician nor a native Welsh speaker (but I do speak Welsh), my IMPRESSION from contact with speakers of different varieties of Welsh, and reading on the subject, is that: in the NORTHERN dialects, the acoustic effect of both the gemination of the stops and the shortening of the preceding stressed vowels is very striking and "vigorous". At least in many speakers, in a word like /kadar/, the intervocalic consonant seems to "last" considerably longer than the preceding stressed vowel (sic!), which sounds "clipped" to my foreign ear. For the stop to "last" that long, I believe the flow of breath must be interrupted completely in the middle, and that is what it sounds like: [kad] (pause) [dar], probably more accurately [kat] (pause) [tar], with no aspiration. (In my experience, speakers from the Iberian peninsula who learn Welsh tend to identify this as [katar].) in the SOUTHERN dialects, where gemination affects a smaller range of consonants, my general impression is that the distinction is maintained but in a phonetically more low-key manner, and since the distinction is phonetic rather than phonemic, my guess is that there is probably considerable variation in how these consonants are realized. On the other hand, for the southern dialects, the distinction, while not phonemic, retains *phonological* significance in that it correlates with a length contrast in the preceding vowel ([ka:der] versus [et(t)o]), and the vowel length distinction *is* observed (very noticeably if you're more used to hearing northern Welsh). Thus while the simple/geminate contrast is not by itself distinctive in most cases (it not only correlates with length of the adjacent vowel but is furthermore itself usually predictable from other features, such as voice), the fact that it does correlate with (clearly perceptible) vowel quantity probably helps to reinforce the speaker's *competence* regarding consonant length, even if structurally superfluous. Then there are the cases where consonant gemination remains phonemic in southern Welsh, e.g. /ka:ni/ 'sing' versus /kanni/ 'whiten' (orthographically, "canu" and "cannu", and both /kannI/ in northern Welsh). It would be interesting to know whether some southern dialects actually confuse such pairs, but this seems unlikely because the vowel "helps", even if the gemination contrast is weakened. In any case, I have not read or heard of southern dialects in which the gemination contrast is lost completely, both phonologically and phonetically. Alan From mccay at redestb.es Fri Nov 13 13:24:12 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 13 Nov 1998 08:24:12 EST Subject: s > r (Romance) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer writes: >It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, >-g-, which must mean that the change was general in most Western >Romance, but did not occur in Eastern Romance (S. Italian and >Romanian). In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > -z- has taken place. I would have guessed (from my position of overall ignorance about the Italian dialects) that this probably also reflected roughly the situation in dialectal Italian south of the Spezia-Rimini line (north of the line, -s- is also voiced, of course, but so are the stops). Doesn't this establish a precedent clearly opposed to assuming (in the absence of further data, I mean) that other Romance dialects which didn't lenite the stops couldn't have lenited -s-? I would have thought so. It is easy enough to see why the (apparent or real) parallel between voicing of intervocalic stops, on the one hand, and of -s-, on the other, suggest the conclusion that a single sound-change event is involved, rather than two separate phenomena, but that is surely not a foregone conclusion a priori? Among non-Romance languages in which regular voicing of (some) intervocalic consonants has occurred, I believe there are cases of voicing of stops but not of -s-, and of course there are languages with the opposite, voicing of -s- but not of stops (e.g. Old English). An example of the former combination may be Welsh in its development from proto-Celtic via proto-Brittonic. Intervocalic stops were voiced, as I remarked in my previous post and is anyway well known. Intervocalic s was in general aspirated, never voiced. My doubt is as to which happened first. If s aspirated first, it could then be argued that the reason why s wasn't voiced together with the stops is that at the time the stops were voiced, there were no intervocalic sibiliants around (only h's). Even if the Welsh example is dubious (as I said, that would seem to depend on the relative chronology of the voicing of stops and aspiration of -s-), perhaps there are examples in other languages? Coming back to Romance, though, and getting somewhat aprioristic myself, I would have thought that a weightier factor capable, potentially, of influencing the way -s- developed into -z-, would be, not what happened with the intervocalic stops, but what happened with *geminate* consonants in general, or at any rate with -ss- in particular. I'm thinking now in terms of structural economy. In dialects where geminates were simplified, non-geminate -s- either had to evolve to -z- (or at least to something other than -s-) or else be confused with -ss- > -s-. Where, on the other hand, geminates are maintained, there can have been no such pressure. The same argument applies, of course, to the lenition of intervocalic voiceless stops. Thus, if such principles of structural economy operate, we would indeed expect -s- and voiceless stops to voice in the same dialects and in the same contexts, unless other factors intervened to differentiate, because both developments would have been responses to similar pressures. But that *is* aprioristic. Returning to facts, isn't Italian a counter-example? Alan From hubeyh at montclair.edu Fri Nov 13 13:24:48 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:24:48 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > I must insist. I deny that anybody uses 1. and 2. in this way. Prove > me wrong. Embarrass somebody. Lots of people do, and some of them even have PhDs. > > Human family members resemble each other. That does not mean that > > unrelated > > people cannot resemble each other. And despite the fact that we know > > both > > we still consider two people who resemble each other to be related > > unless > > there's proof to the contrary. > > We do? Of course we do. WE make lots of decisions about the world based on a kind of a "theory" of everything that we have created about us based on our life experiences from the moment of birth. Unless we know the people it is similarity that propels us to thinking such things. > > How many language families has any human experienced? I do not mean the > > purported/alleged language families. > > Since I don't know which entities qualify in your mind as "language > families" as opposed to "purported/alleged language families", I > can't answer this. However, on my own understanding of what a > language family is, any competent historical linguist has experienced > (has some knowledge of) a variety of language families and of various > languages not known to be related. But that is not what I asked nor commented on. There lies the circularity again. How do we know that they constitute a family? > You seem to be suggesting that the empirical base of historical > linguistics is too small. Unless you have some realistic suggestion > as to how it could be signficantly enlarged, we have to live with it. > If that disqualifies it as "science" in your opinion, too bad. No, I am not suggesting that. I am merely stating the reasons for belief that language families exist, and that we can come to know them. > Calculation of whether resemblances could be due to chance or not > becomes relevant in distant relationships or borderline cases, about > which so much argument goes on now. I'm talking about families even > Lyle Campbell believes in, where there is no argument. In some cases > It is always relevant. The fact that some people's beliefs can be based on analogy and resemblance to physics or biological families or imitation does not change the facts. (eg Latin and Romance) we have the proto-language through direct > documentation. We also have recorded histories of many individual > languages which tell us a lot about how languages change. We have Latin at some stage and its descendants. The others are then created using analogies, which is what at least some linguists deny. But they also deny that the reasoning (about all those resemblances not being due to chance) is implicitly probabilistic. So then what is left? It's not analogy, it's not induction, it's not probability! > This is not circularity. Mathematical methods can be wrong. A number This is worse than pseudoinduction. Which mathematical method is wrong? No axiomatic system can ever be wrong by definition. > of different proposals have been made for calculating the probability > of accidental linguistic resemblances. They give different results. > Therefore they can't all be right. They may be free of mathematical > error, but they do not necessarily yield historically correct > conclusions. It is not the method, it is the person who uses it. If you bash your finger using a hammer is it the hammer's fault? There are incompetent people in many places. > On the other hand, one would want to ask the linguist in question > just what was the basis of his/her certainty that X and Y are _not_ > related. Why ask a person when there are dozens of books which purport to explain it? > > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From larryt at cogs.susx.ac.uk Fri Nov 13 13:27:20 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Fri, 13 Nov 1998 08:27:20 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, John Hewson wrote: > Actually there is an algorithm that a Systems Analyst and I > developed to produce my _Computer-generated Dictionary of > Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be > stated in a single sentence, as on p.iv of the above: From the data of the > daughter languages generate all possible protoforms, then sort > alphabetically, and examine all sets of identical protoforms collocated by > the sort. Very interesting, but I'm amazed. What particularly catches my attention is the expression "all possible protoforms". How on earth can this label be fleshed out? How do you know that your program generates all possible protoforms? Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Fri Nov 13 13:27:46 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:27:46 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > Neither of these procedures guarantees fair coverage of vastly different > frequencies of different consonants, language-specific or family-specific > preferences of different consonants or consonant classes for different > phonotactic positions, and the like. I hope that some of these differences > get ironed out by putting consonants together in phonetic groupings. How about simulation results? Have you done any ? > Still, the metric is only approximate. It enables us to point out that 36 > resemblant sets, half of them with only one resemblant consonant, isn't > enough to indicate genetic relatedness unless a very small wordlist was > specified in advance. What is your conclusion on how many such pairs are needed? Is this number for the whole language or just using a Swadesh like list? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From Roger.Wright at liverpool.ac.uk Fri Nov 13 13:33:46 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 13 Nov 1998 08:33:46 EST Subject: s > r (Iberian) In-Reply-To: <364adb3b.1343704@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- In fact, the supposed non-voicing of intervocalic plosives in Mozarabic (Ibero-Romance spoken in Moslem Spain) was probably a chimera caused by the Arabic alphabet's application to Ibero-Romance; I think most specialists are rather dubious about it by now. And the Aragonese referred to is high up in the Pyrenees, and not characteristic of the whole of Aragon - RW On Thu, 12 Nov 1998, Miguel Carrasquer Vidal wrote: >It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, >-g-, which must mean that the change was general in most Western >Romance, but did not occur in Eastern Romance (S. Italian and >Romanian). > >The exceptions in Western Romance are Mozarabic (partially, and as >far as this can be determined) and Aragonese, or at least part of it. From Roger.Wright at liverpool.ac.uk Fri Nov 13 13:35:14 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 13 Nov 1998 08:35:14 EST Subject: Stabilized languages In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Larry said: >First of all, languages do not "stabilize". Every living language is >always changing, and the only language that doesn't change is a dead >one. Change is at some times slower than at other times, though. And it just isn't true that "dead" languages don't change. Medieval Latin varied astonishingly widely in time and space, Renaissance Latin was recognizably different again, and even the modern Latin used by the Vatican has been continually acquiring new vocabulary. RW From mcv at wxs.nl Fri Nov 13 21:27:04 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 16:27:04 EST Subject: s > r (Iberian) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Roger Wright wrote: >And the Aragonese >referred to is high up in the Pyrenees, and not characteristic of the >whole of Aragon. Not anymore, although a word like still covers the whole Aragonese area. Place names and medieval documents prove that retention of -p-, -t- and -k- was also wide-spread in Lower Aragon until the 15th century or so. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Fri Nov 13 21:26:49 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 16:26:49 EST Subject: s > r (Romance) In-Reply-To: <3.0.3.32.19981113084536.006e4c00@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > >-z- has taken place. I would have guessed (from my position of overall >ignorance about the Italian dialects) that this probably also reflected >roughly the situation in dialectal Italian south of the Spezia-Rimini line >(north of the line, -s- is also voiced, of course, but so are the stops). It's not that simple. Standard Italian has plenty of cases of voiced stops (Rohlfs gives: ago, drago, lago, sugo, spiga, briga, segale, luogo, lattuga, aguzzo, pagare, segare, pregare, sfogare, affogare, fregare, annegare, frugare, intrigare; spada, strada, malgrado, contrada, rugiada, contado, scudo, lido, padella, badia, badessa, badile, gradire, scodella, gridare, medaglia, stadera, mortadella, podesta`, budello, mudare, podere; povero, vescovo, arrivare, ricevere, cavezza, ricoverare, rimproverare, sceverare, scovolo). As to -s-, it is pronouned voiceless /s/ in asino, cosa, casa, mese, fuso, peso, naso, the suffixes -oso, -ese (except cortese, francese, marchese, palese, paese), the verbal endings -esi, -isi, -osi, -usi, -eso, -iso, -uso. Voiced /z/ is found in: base, battesimo, bisogno, caso, chiesa, crisi, cristianesimo, deserto, dose, fantasia, fase, fisica, lasagna, lesina, medesimo, misero, musica, osare, pausa, posa, paradiso, rosa, quaresima, quasi, spasimre, sposo, scusare, te`si, uso, usare, vaso, ventesimo, viso (from : esame, esatto, esempio, esemplare, esiguo, esigere, esercito, esente, eseredare). According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in approxiamte order of frequency) are not native Tuscan forms, but words imported from the north. What *is* native Tuscan is the development -k- > -h- ("gorgia toscana"), and also (in a smaller area) -t- > -T- and -p- > -P-. As in the Goidelic lenitions, this also occurs across word boundaries. To the south of Tuscany, an equally "sandhinista" area is Lazio (as well as Corsica and Sardinia), but this time the parallel is with Brythonic (-p- > -b- [here we can distinguish between Northern imports and local lenitions], -t- > -d-, -k- > -g-). Apparently -s- participates in Corsica (/u zale/ "the salt"), but I'm not sure about Lazio. There is no "aspiration" of /s/ in Tuscany, whatever that might mean. So these *are* counterexamples. However, the Eastern Romance non-lenition (south of La Spezia-Rimini) applies both to the stops and to -s-. Of course the imbalance in the system (3/4-way for the stops /d/, [/dd/,] /t/, /tt/, 2-way for the sibilant /s/, /ss/) may have favoured the influx of northern words containing /z/. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From hubeyh at montclair.edu Fri Nov 13 17:03:33 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 12:03:33 EST Subject: Doing historical linguistics Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > Well, why don't you write (or get a student to write) a brute-force > program to explain it? Students are not what they used to be :-) > We *should* use information technology to assist research in > historical linguistics. If only to nip in the bud attempts to relate > Bq. "dry" with something in Caucasian ("Warning 23: oldest > attested semantics: "barren, sterile"") or Bq. with words > for "woman" sounding like /kwVn-/ ~ /kwVm-/ or Sum. with > words for "woman" sounding like /em(e)/ ("Error 09: operation not > commutative"). As they say "it all comes out in the wash". That is why averages are used in statistics. The fact is that if there are lots of errors in the present method, there's no way to find out, exactly as in the case of using stats. For example, without Sumerian, no book like Tuna's would have been possible. That means that many Turkic words could have always been attributed to Tocharian, Iranian etc where now one may be able to point to roots earlier than IE as in Hittie etc. That means the comparative method also suffers from the same problems as any stat method, but that cannot be discovered. It may be that looking at thousands of languages employing brute-force methods might create results which themselves could be useful for standard practice. But at the same time clear models are needed. After all, what we want in the final analysis is if there exists any pattern that is not likely due to chance. > Certainly there are people using programs to paint and make music. > But the value of the result depends entirely on the human input > parameters, and on human selection/rejection/editing of the output. > Even computer art is a craft. It takes a specialist (or a gifted > person) to get interesting results. There's no doubt that humans created computers and wrote the programs. There's also no doubt that computers can now do things that great mathematicians used to do. So there should also be no doubt that they can do what linguists do. > A program (brute-force or otherwise) to assist in reconstruction of > proto-languages can be made, but surely to be useful it should > incorporate existing knowledge about linguistics in its programming > or configuration parameters (likely phonological developments and > semantic shifts) and it should be fed reliable and complete data, > including morphological information, otherwise it's GIGO. Existing knowledge, yes, existing dogma no. There needs to be clear models not lumping things constantly ignoring possible errors. If nothing else familiarity with some aspects of probability theory will restrain linguists from making the kinds of blanket/certain statements which no respectable scientist would ever make. In this case I am referring particularly to kinds of statements which Larry Trask often makes, probably thinking that he is discussing chemistry probably via analogy. > It is an illusion to think that we can throw away two centuries of > comparative linguistic practice, and let "logic, probability theory > and fuzzy set theory" do all the work. Mathematicians aren't going > to be replaced by theorem proving automata any time soon, and neither > are historical linguists scheduled to be replaced with > "proto-language construction programs". Mathematicians might be replaced by computers. About a year ago a program which was more or less brute-force proved a theorem and when people (mathematicians) examined the theorem, they thought it was rather ingenious. Let us not forget about Big Blue. I will post the music article separately. The reason all this happens is because people make intuition errors. Intuition (creativity, etc) is not all what it is cracked up to be, and it is mostly wrong to rely on it. > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From jhewson at morgan.ucs.mun.ca Fri Nov 13 17:02:28 1998 From: jhewson at morgan.ucs.mun.ca (John Hewson) Date: Fri, 13 Nov 1998 12:02:28 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, Larry Trask wrote: > On Thu, 12 Nov 1998, John Hewson wrote: > > > Actually there is an algorithm that a Systems Analyst and I > > developed to produce my _Computer-generated Dictionary of > > Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be > > stated in a single sentence, as on p.iv of the above: From the data of the > > daughter languages generate all possible protoforms, then sort > > alphabetically, and examine all sets of identical protoforms collocated by > > the sort. > > Very interesting, but I'm amazed. What particularly catches my > attention is the expression "all possible protoforms". How on earth can > this label be fleshed out? How do you know that your program generates > all possible protoforms? Let me try to answer with a simple example. An /s/ in Fox can come from 6 different Proto-Algonkian sources. The program generates all 6 from every /s/ in a Fox word. Let us call this the productivity of Fox /s/. The total of all possible protoforms for a Fox word is based on the productivity of all the phonemes in the word. Since Algonkian words often contain several morphs, the number of possible protoforms can be very high. In the first batch we did, the average number of possible protoforms for every item of input data was over 20. Given the length of Algonkian words we were able to simplify by deleting the vowels and working from the consonant frameworks only. Otherwise the program would probably have bombed, even on the mainframe on which we were working at that time. Lowe and Mazaudon, by contrast, worked on monosyllables that had tone. Such systems are for the reconstruction of the vocabulary once all the correspondences and reflexes of the relevant daughter languages have been worked out. It saves the worker the donkey work of pounding through dictionaries looking for items that may not exist. It cannot replace the linguist, it is merely a tool to aid the linguist. Algorithms may operate mechanically, but they are not created mechanically. To create them requires imagination, and to use them requires a professional level of understanding of the comparative method. John Hewson, FRSC tel: (709)737-8131 Henrietta Harvey Professor of Linguistics fax: (709)737-4000 Memorial University of Newfoundland St. John's NF, CANADA A1B 3X9 From scott at math.csuohio.edu Fri Nov 13 17:00:24 1998 From: scott at math.csuohio.edu (Brian M. Scott) Date: Fri, 13 Nov 1998 12:00:24 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > Ross Clark wrote: > > This is not circularity. Mathematical methods can be wrong. A number > > This is worse than pseudoinduction. Which mathematical method is wrong? > No axiomatic system can ever be wrong by definition. It may, however, be inapplicable. If a mathematical method produces a historically incorrect conclusion, then obviously the assumptions on which it is based do not obtain, and it is wrong, i.e., the wrong method. Brian M. Scott Dept. of Mathematics Cleveland State Univ. From cravens at macc.wisc.edu Sat Nov 14 15:15:45 1998 From: cravens at macc.wisc.edu (Thomas D. Cravens) Date: Sat, 14 Nov 1998 10:15:45 EST Subject: s > z (etc.) (Italy, Spain) Message-ID: ----------------------------Original message---------------------------- Comments below (Tom Cravens) (Note: I [TC] mistakenly sent this directly to Miguel, rather than the list. Perhaps the discussion of weakening by class is of some general interest.) At 04:26 PM 11/13/98 EST, you wrote: >----------------------------Original message---------------------------- >"Alan R. King" wrote: > >>In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > >>-z- has taken place. I would have guessed (from my position of overall >>ignorance about the Italian dialects) that this probably also reflected >>roughly the situation in dialectal Italian south of the Spezia-Rimini line >>(north of the line, -s- is also voiced, of course, but so are the stops). > >It's not that simple. Standard Italian has plenty of cases of voiced >stops (Rohlfs gives: ago, drago, lago, sugo, spiga, briga, segale, >luogo, lattuga, aguzzo, pagare, segare, pregare, sfogare, affogare, >fregare, annegare, frugare, intrigare; spada, strada, malgrado, >contrada, rugiada, contado, scudo, lido, padella, badia, badessa, >badile, gradire, scodella, gridare, medaglia, stadera, mortadella, >podesta`, budello, mudare, podere; povero, vescovo, arrivare, >ricevere, cavezza, ricoverare, rimproverare, sceverare, scovolo). >As to -s-, it is pronouned voiceless /s/ in asino, cosa, casa, mese, >fuso, peso, naso, the suffixes -oso, -ese (except cortese, francese, >marchese, palese, paese), the verbal endings -esi, -isi, -osi, -usi, >-eso, -iso, -uso. Voiced /z/ is found in: base, battesimo, bisogno, >caso, chiesa, crisi, cristianesimo, deserto, dose, fantasia, fase, >fisica, lasagna, lesina, medesimo, misero, musica, osare, pausa, >posa, paradiso, rosa, quaresima, quasi, spasimre, sposo, scusare, >te`si, uso, usare, vaso, ventesimo, viso (from : esame, esatto, >esempio, esemplare, esiguo, esigere, esercito, esente, eseredare). Unless I'm mistaken (I don't have the volume at hand), Rohlfs is talking about Tuscan in that passage, not Standard Italian. The two are distinct, although they match up fairly well for the examples he gives (but not always across the lexicon: Mod. Florentine /letikare/ vs. It. /litigare/ 'argue, fight'). In any case, data pulled out of Standard Italian (assuming we can identify what that is) are not the most felicitous for constructing or testing theories, given that the Standard was presumably no one's native language until relatively recently, and even in the far-off days of trying to derive a standard from Tuscan, that standard was based on literary models, rather than actual speech, and ultimately was and has been subject to a good amount of winnowing. [It's of interest how much Tuscan detail ("tutte helle hosine" someone once told me) is absent from Italian.] >According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in >approxiamte order of frequency) are not native Tuscan forms, but >words imported from the north. Yes, Rohlfs says so, but this bit isn't that simple, either. Restructured voicing of /p t k/ is not lexically consistent within Tuscany. More importantly, it affects forms which are not northern (e.g. codesto), some which did not voice in N. Italy or even in Spain (PAUCU > W. Tuscan /pogo/, AVICA > W. Tuscan /oga/), and lots of local toponyms, presumably not borrowed. The latest work on this suggests strongly that surface-level (allophonic) voicing was once much more widespread in Tuscany than it is today, and that much of the /p t k/ > /b-v d g/ now found may be the result of reinterpretation of surface forms during rule competition (gorgia coming in to compete with voicing). It's still going on. Relevant to the current thread is that /p t k/ and /s/ appear to have been affected by most of the same complex of rules, i.e. that Tuscan doesn't provide a good counterexample to the hypothesis that /p t k/ and /s/ are expected to undergo restructuring en bloc. The gorgia mentioned below does, though, and draws attention to the expectation that sound change (accepting that introduction of allophony is a change in sound) will proceed member by member within classes, with nothing to suggest that it must spread outside the innovating class. It can, of course, as Corsican /s/ --> [z] along with voicing of /p t k/ shows, but it needn't be the case. Alto Aragones illustrates this as well with regard to degemination, or at least did fifty years ago when Badia Margarit was investigating the speech of Bielsa. Working on memory again, but I seem to recall that in the midst of the massive degemination that one would expect in that part of the Romance world, there was maintenance of length not only for the expected /rr/, but also for /ll/ and /nn/ (and possibly even /mm/, but there were very few examples, one of which was [tammjen], suspect as a recent, and perhaps surface-only, assimilation). In any case, Belsetan by that time had degeminated all of /pp tt kk/, /bb dd gg/, etc, but was hanging on to the last remnants of /nn/, /ll/, /rr/. Marie-Jose' Dalbera-Stefanaggi has found similar phenomena in Northern Corsica: liquids and nasals are the last to lose the possibility of manifesting surface length. In neither type did degemination of the stops imply (immediate) degemination outside their class. >What *is* native Tuscan is the development -k- > -h- ("gorgia >toscana"), and also (in a smaller area) -t- > -T- and -p- > -P-. As >in the Goidelic lenitions, this also occurs across word boundaries. >To the south of Tuscany, an equally "sandhinista" area is Lazio (as >well as Corsica and Sardinia), but this time the parallel is with >Brythonic (-p- > -b- [here we can distinguish between Northern >imports and local lenitions], -t- > -d-, -k- > -g-). Apparently -s- >participates in Corsica (/u zale/ "the salt"), but I'm not sure about >Lazio. There is no "aspiration" of /s/ in Tuscany, whatever that >might mean. So these *are* counterexamples. Yes. I think Lazio is typically [s], but the voicing of /p t k/ is extensive enough to cause spelling problems a` la ~ in Am. E. >However, the Eastern Romance non-lenition (south of La Spezia-Rimini) >applies both to the stops and to -s-. Of course the imbalance in the >system (3/4-way for the stops /d/, [/dd/,] /t/, /tt/, 2-way for the >sibilant /s/, /ss/) may have favoured the influx of northern words >containing /z/. Maybe. But there's still the gap of no /zz/, and no phonologically "breve" counterpart of long palatal laterals and nasals, i.e. no *[aljo] possible to contrast with [alj:o] 'garlic', no *[onji] can contrast with [onj:i] 'each' (apologies for the fake IPA). Tom Cravens University of Wisconsin-Madison cravens at macc.wisc.edu >======================= >Miguel Carrasquer Vidal >mcv at wxs.nl >Amsterdam From fcosw5 at mail.scu.edu.tw Sat Nov 14 20:33:59 1998 From: fcosw5 at mail.scu.edu.tw (Steven Schaufele) Date: Sat, 14 Nov 1998 15:33:59 EST Subject: concerning the algorithmizability of historical linguistics Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > I also have very small doubts that I can write a program that can take > n words from language A, and m from language B and write a program that > can change one set into the other (at least enough of them to dumbfound > the skeptics and force them to having N new looks into the "comparative > method") using only regular sound changes. The only problem is that I > don't have the time or energy to put into such a useless demonstration. > It will probably be done by someone (like Starostin) or someone else > who is upset at the way linguists spurn statistics and math. After all, > it the only thing that will make people sit up and take notice there is > no better way than to demolish their toy :-) This is precisely where i feel we are taking leave of scientific responsibility. If you can write a program that can take n words -- *any* n words -- from language A -- *any* A -- and m words -- *any* m words -- from language B -- *any* B -- and change one set into the other, how is such a result falsifiable? Such a program sounds to me like one that is guaranteed to find evidence of affiliation no matter what data are fed to it. And i'm quite sure that is not desirable. Should we ever discover extraterrestrial, or otherwise non-human, language, such an approach as Mark seems to be suggesting here would surely be able to `prove' that such a language is nevertheless affiliated somehow with any given human language. Which would definitely be a major step in the wrong direction, since at least IMHO the discovery of such a language ought precisely to provide us with an opportunity to get beyond studying merely human language (presumably a manifestation of species-specific characteristics of the human brain, etc.) and onto the path of examining the nature of Language itself in the abstract. I'm concerned that Mark's approach to developing AI programs for historical linguistics might also lend itself to developing programs that would quite happily `demonstrate' that, e.g., fish, ichthyosaurs, and whales are all more closely related to each other than any of them are to salamanders, turtles, or cows, or that dragonflies, eagles, and bats are more closely related to each other than any of them are to spiders, penguins, or shrews. As, e.g., Larry and Miguel have noted, AI systems can be extremely helpful in doing a lot of the comparative work our science is based on, and perhaps even draw our attention to relationships and affiliations we might not otherwise have considered. But the results of such analyses still need to be subjected to expert peer review. Maybe a certain amount of that sort of thing can be automated too, but i very much doubt that all of it can be, at least in the rather elementary, simple-minded manner of reducing it to an algorithm. There's a lot of judgment involved, which judgment needs to be based on a lot of experience which is definitely not algorithmizable. To what extent that kind of experience can be taught to a computer, or a computer programmed to learn it as a living, flesh-and-blood historical linguist does, i confess i have no idea. I am also bothered, as is Larry, by Mark's continued (apparent) tendency to equate scientific research in a field such as (historical and comparative) linguistics with artistic endeavours such as painting pictures, composing music, or playing chess. A freshly-painted picture, a new piece of music, a game of chess is a new creation and derives some of whatever value it has from that novelty; the same is true, of course, of a freshly-crafted, previously-unused and -unheard-of sentence. Or a mathematical theorem, for that matter. A reconstructed protoform, on the other hand, is a hypothesis concerning something (a linguistic expression) that is supposed to have really existed in the objective universe at some point in the past. Its validity ultimately stands or falls on how closely it approximates objective, historical reality, not on its solipsistic elegance. Best, Steven -- Steven Schaufele, Ph.D., Asst. Prof. of Linguistics, English Department Soochow University, Waishuanghsi Campus, Taipei 11102, Taiwan, ROC (886)(02)2881-9471 ext. 6504 fcosw5 at mail.scu.edu.tw http://www.prairienet.org/~fcosws/homepage.html ***O syntagmata linguarum liberemini humanarum!*** ***Nihil vestris privari nisi obicibus potestis!*** From bwald at HUMnet.UCLA.EDU Sat Nov 14 20:34:50 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sat, 14 Nov 1998 15:34:50 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- Mark Hubey makes an interesting, visionary and somewhat naive point -- as follows: >If there are people who are writing programs to paint (yes, produce >art), and compose music, it takes no genius to see that even if Starostin only >wrote (or got a student to write) a brute-force, dumb program on a >commodity grade PC, he can uncover relationships that many humans cannot >do, even >if they collaborate. The reason for this will take too long to explain. >But given a set of words (and their meanings) even a brute-force >program can keep cranking 24 hours a day to produce cognates via regular >sound changes, clusters, and things that a typical linguist does not >even know exists. In principle, once someone has established cognate relationships, they can indeed write an algorithm to derive their original data from their reconstruction. But that is the REVERSE of what Mark envisions. Thinking about the nature of the discovery procedure imagined by Mark we take all the words (from dictionaries -- but shouldn't it be morphemes?) in all the languages of the world and run all possible correspondences on them (whatever "all" means in this context, e.g., k =t, k= p , k = a, k = @@ etc etc etc) and then choose the ones that work out best (semantically, phonetically and in terms of number of cognates per language pair -- or whatever, AND apply apppropriate statistical tests to make sure the "best" is above chance, AND fight like hell against anyone who dares claim "borrowing", "wander-words" and all that esoterica, in cases where statistical tests are not overwhelmingly decisive etc etc) and VOILA. I wonder if Mark realises what is involved in such a program and how pitifully simple the chess program that beat the world champion is in comparison (with its evaluation of 3 gazillion positions per minute or whatever it was). Because of what can be done a posteriori, I don't dismiss the vision out of hand -- but in view of what's really involved, we'll be arguing the way we do now for generations to come, before anything remotely resembling Mark's vision emerges for any uses other than cryptography. He ends with the usual premise of science fiction writers, and pop accounts about scientist-heros and their breakthroughs: >It's too bad that the attitude of most linguists is, in fact, the most >damaging to themselves and their own professions. But, that is the way >evolution is. Short term goals and intuition only go so far. In other words, the "establishment" are the usual and familiar small-minded villains that hinder progress, and are too vested in their own positions to recognise a good idea when it is shoved up their .... Not me, Mark. I'm open-minded. So, explain how this brute-force program works better than I just did above, and how it addresses all those issues I compressed into my description. Can it also decipher Linear A etc. (or do I mean "translate")? P.S. Anybody. What is the promise of programmed cryptographic methods for discovering "non-obvious" sound correspondences and-so-on among the lexica of different languages P.P.S. As opposed to cryptography, doesn't the programmed "cognate" search have to start with semantic features (and then the gazillions of brute attempts at sound correspondences)? How should a semantic feature search be organised for big-time time-depths (cf. Johanna's remarks on "five" different meanings for a cognate, e.g., "fly", "wing", "feather", "fur", etc. "five" is a very rough count. Cognates or assumed cognates exhibit different DEGREES of semantic resemblance, from obvious to "far-fetched", e.g., "night", "bump into things in the dark", "get knocked out", "sleep", "lie down", "sexual intercourse", "AIDS", "dirty needles", "gay", "happy", etc. That also has to be modelled for statistical measurement, e.g., "wing" is "closer" to "feather" than to "fur", etc. -- but "fur flies", at least in English.) PPPS. Mark wrote: >...most of the IE words could be due to the substratum which >could have been a family. One can always insist that the reason why IE >words resemble each other is because they are all left over from a >previous language which was spread out over the same region. Isn't that like saying: The "Odyssey" wasn't composed by Homer but by somebody else with the same name (or whom we choose to call Homer -- yeah, I know, but he wasn't blind). If not that, then Mark is proposing an extremely complicated and unlikely hypothesis, in which case in the absence of contrary evidence we choose the simpler one ("science" actually makes decisions like that -- in the absence of contrary evidence, I said.) I'm not even sure that I can imagine what the complicated hypothesis is in this case. Maybe something like, IE had fragmented into Germanic and Slavic etc, but the current languages classified as Gmic, Slvic etc don't descend from them but just happened to borrow most of their vocabulary and grammar from them. So, then, BY WHAT CRITERIA did these languages borrow so much from their IE neighbors rather than give up their own presumably non-IE languages and ADOPT those IE languages, so that they are INDEED descendants of IE. From bwald at HUMnet.UCLA.EDU Sat Nov 14 20:35:18 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sat, 14 Nov 1998 15:35:18 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- I welcome Alan King's message. On his first point: >B. Wald says: It just happened -- because both directions >>are possible (under certain conditions -- certainly NOT ****z > s /V_V). >That provoked me to try and think of counterexamples to Benji's latter >assertion here. Good. If I'm wrong, I'd like to know as soon as possible, so that I don't keep on thinking that and misleading other people. That's my best pay-off for my interventions on the list. He continues: >I couldn't come up with an example of z > s /V_V, but I >did find one of +voice > -voice /V_V. Just as good. I would have thought that if such a change occurs it can't be due to the intervocalic position (just as in the z > s "reversal" case), and would expect it to occur in that position only as part of a more general voicing shift (which the intervocalic position is not able to resist at the time). A short digression. He asked: >(Okay, I'm going to be a pain and say it: just what is "Central Spanish" >supposed to mean? The phenomenon in question includes the WHOLE of >Castilian (i.e. Spanish) in its territorial extension. I don't know. In my insecurity as a non-expert, and without checking, I wanted to rule out some non-Central (northern) dialects, without being sure which ones, maybe Leonese or Aragonese or whatever, where maybe z > s did not occur, at least at first (since I remembered it is said to have been observed first in the "Central" dialects like Castille -- and not to have taken effect until, say, after the Jews were expelled, so that Ladino still has-z- in "casa" etc.) To get back to the main point, Alan proposes an example from some (southern) Welsh dialects. I'm not sure I understood the entire discussion, but the situation seems to be as follows: >I have put quotes around "voiced" and "voiceless" because >phonetically all stops in Welsh tend to be voiceless (or at least voicing >is not critical) and the contrast is realized principally in terms of >tenseness and aspiration. To simplify the exposition I shall henceforth >largely ignore that fact in the transcriptions and terminology used.) Alan is referring to STANDARD Welsh here, and that by "critical" he must mean relevantly that the degree of voicing of the "tense" stops varies. The issue then becomes whether this is a novel situation or reflects an older situation in which there were consistently voiced stops that have now come to vary in voicing as they shift to tenseness as their primary (or distinctive) feature. The issue is also whether this shift, if it is so, is ONLY intervocalic (post-stress) or whether it is general to the stop paradigm (e.g., pre-stress, initial, etc.). He continues: >... in Northern spoken Welsh, this system has >been altered, in that nearly all consonants in the intervocalic post-stress >position are pronounced geminate (and correspondingly, all stressed vowels >followed by a consonant in non-final syllables are short - but I'm going to >focus on the consonants here). So corresponding to standard ['ka:dajr] we >will find ['kaddar] in northern Welsh, while ['etto] shows no change since >the consonant is already geminate. Very interesting. I would assume that the gemination of voiced stops represents an interruption of voicing between vowels, and that even the shortening of a long vowel before a geminate voiced stop anticipates such an interruption. In any case, that does seem to me a mechanism by which intervocalic voiced stops could devoice -- and to the extent that gemination is strictly an intervocalic phenomenon in Welsh the stage is indeed set for devoicing of the intervocalic stop. Finally, he notes: >Now: apart from this gemination and shortening, but possibly related to it, >we find some varieties of Welsh in which, corresponding to consonants like >the /d/ in , if the consonants in question are "voiced" stops in >standard Welsh, we find "voiceless" stops instead: ['katar] etc. Assuming >once again that the standard form is diachronically prior, we then have d > >t (or possibly d > dd > t) in post-stress intervocalic position, and >similarly b > p and g > k. Irritatingly, this phenomenon has been noted in >the south, in an area not contiguous to the northern "geminating" varieties. In which case, it is commendable that Alan suggests that the route assumed may have been a continuation of the one still evident in the north. I need to reflect further on this case, but it seems plausible to me on the face of it. What gives me most pause is whether this "counts" as STRICTLY intervocalic position, if there is the assumed geminate intermediate stage, and it just so happens that geminates only occur intervocalically -- so that it might be interpreted as a shift of voicing in GEMINATES, regardless of their position -- which just happens to be intervocalic. I'm also not sure I followed the entire thing. For example, since there were also geminate VOICELESS intervocalic stops in these dialects of Welsh, was there a merger, or did the "voiced" geminates de-geminate before "fully" devoicing? i.e., -tt- vs. d > dd > *td* > t (where somehow *td* is something that avoided merger with -tt- as it devoiced.) The phenomenon is certainly interesting in its own right, and deserves further attention, possibly further dialectal research to examine what can be discovered about the details of the path. This is particularly important if the result is -tt- vs. -t- (< -tt- vs. -d-), since Alan's suggestion seemed to be that gemination had something to do with the devoicing. So far the evidence in the south does not indicate that, but only that, for some reason, -d- came to move into a space parallel to -tt- by devoicing. If that's the case, it is indeed a counter-example to the notion that stops cannot devoice intervocalically without first going through some other changes (like gemination and associated metric effects on syllabification). I appreciate Alan's example, and hope that further clarity will emerge on the historical changes represented in it. PS. I forgot whether Alan mentioned this, but if a form like "chair" had a long vowel before the stop originally, as in the standard, and shortened it before gemination in the north, then the shortening in the south may also suggest prior gemination. Obviously we need to know a whole lot more about the history of the relevant Welsh dialects before we can draw secure conclusions. From larryt at cogs.susx.ac.uk Sat Nov 14 20:35:44 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Sat, 14 Nov 1998 15:35:44 EST Subject: the Trask-Hubey debate In-Reply-To: <364BC8D4.743917C0@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, H.M.Hubey wrote: > Any piece of paintwork does not count as painting. The painters have > their own black magic guild which decides what is art and what is > not. Does that ring a bell? Nope. What we try to reconstruct is not individual miscellaneous items, but rather the whole system of a vanished language, or as much of that as we can. A proposed reconstruction is accepted by other linguists to the extent that it (a) is internally coherent and (b) accounts successfully for the data. Aesthetic factors like symmetry and economy are not negligible, but they must take a back seat to our major concern: how well does the proposed reconstruction account for the data in a principled manner? > I also have very small doubts that I can write a program that can take > n words from language A, and m from language B and write a program that > can change one set into the other (at least enough of them to dumbfound > the skeptics and force them to having N new looks into the "comparative > method") using only regular sound changes. The only problem is that I > don't have the time or energy to put into such a useless demonstration. I think "useless" is a very apt description. No doubt one could write a cutesy program that would convert, say, a set of English words into Zulu words, but such activity bears no discoverable resemblance to comparative reconstruction or to any other aspect of historical linguistics. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Sat Nov 14 20:36:08 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Sat, 14 Nov 1998 15:36:08 EST Subject: Stabilized languages In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, Roger Wright wrote: > Change is at some times slower than at other times, though. Agreed, of course, but it never vanishes in a living language. > And it just isn't true that "dead" languages don't change. Medieval > Latin varied astonishingly widely in time and space, Renaissance > Latin was recognizably different again, and even the modern Latin > used by the Vatican has been continually acquiring new vocabulary. Agreed, but I wasn't claiming that a dead language can never change, but rather that absence of change is only possible in a dead language. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mccay at redestb.es Sat Nov 14 20:36:37 1998 From: mccay at redestb.es (Alan R. King) Date: Sat, 14 Nov 1998 15:36:37 EST Subject: help with Finnish, please? Message-ID: ----------------------------Original message---------------------------- This may or may not be deemed a suitable question for HISTLING, but I'm going to try anyway. I am translating/editing an article on Basque language standardization, and would like to double-check on the accuracy of a few statements by the original author concerning Finnish words and their etymologies. Anyone want to (and able to) help? I will acknowledge help (if ultimately used) in a footnote to the translated version. The author says that: Finnish PEILI "mirror" is the assimilated form of a loanword SPEKLI; Finnish ROUVA "lady" similarly < FROUWA; Finnish ROPAKANTA "propaganda" < PROPAGANDA; Finnish RANTA "?" < STRAND; Finnish STRUKTUURI < STRUKTURA. The author does not state what language(s) the second forms given came from, but I would like to know (I'm thinking here of SPEKLI and FROUWA in particular; is this some sort of proto-Germanic, or what? should they be starred? and of course, are they accurate?). Kiitos! Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From mccay at redestb.es Sat Nov 14 20:38:59 1998 From: mccay at redestb.es (Alan R. King) Date: Sat, 14 Nov 1998 15:38:59 EST Subject: intervocalic devoicing in Welsh (?) Message-ID: ----------------------------Original message---------------------------- In response to Benji Wald in response to me on a possible example of intervocalic devoicing in certain varieties of Welsh, I unfortunately cannot add very much as that is about the limit of my knowledge on the particular subject (and I do hope the information I gave is largely accurate). From here on one would need the help of a Welsh *specialist* (I just happen to know some Welsh and have an inquisitive mind, but have not studied the subject in detail). So regarding Benji's PS at the end: >PS. I forgot whether Alan mentioned this, but if a form like "chair" had a >long vowel before the stop originally, as in the standard, and shortened it >before gemination in the north, then the shortening in the south may also >suggest prior gemination. Obviously we need to know a whole lot more about >the history of the relevant Welsh dialects before we can draw secure >conclusions. I can only agree with the second sentence, and point out with regard to the first that I don't know, off-hand, what length the vowel has in the first syllable of the word for "chair" (standard CADAIR) in varieties with the devoicing phenomenon; if the transcription I gave implied that it is short, that was a slip-up. Since southern dialects in general maintain vowel quantity distinctions, my initial assumption must be that it is probably long. Furthermore, when a Welsh specialist is found, she or he should be questioned on the exact dialectal distribution of the devoicing phenomenon. I know it does occur in the southeast, but can't remember whether or not it is also found further north, in which case there might be a dialect with both devoicing and vowel shortening. Bearing in mind my disclaimers, I'll be glad to give any related information about Welsh if needed and I feel competent to do so. Hwyl, Alan From mcv at wxs.nl Sat Nov 14 20:39:38 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Sat, 14 Nov 1998 15:39:38 EST Subject: s > z (etc.) (Italy, Spain) In-Reply-To: <28111321133870@vms2.macc.wisc.edu> Message-ID: ----------------------------Original message---------------------------- Tom Cravens wrote: >Comments below (Tom Cravens) >(Note: I [TC] mistakenly sent this directly to Miguel, rather than the list. >Perhaps the discussion of weakening by class is of some general interest.) And this is how I replied, and what Tom replied to my replies while we were sorting out where the message should have gone... I trust it's allright with Tom to compress things this way, as his final reply to me (though private in principle) does not contain information more sensitive than that he attended to a party in Perugia in his student days (who hasn't?), and does contain a couple of useful references of general interest. [TC:] >>Unless I'm mistaken (I don't have the volume at hand), Rohlfs is talking >>about Tuscan in that passage, not Standard Italian. [mcv:] >For v, d, g he's talking about "la lingua nazionale" (and then >separately about Toscana, N, C and S). For s, he starts discussing >Tuscan right away. Since the written language does not distinguish >/VsV/ from /VzV/, both written , in a sense there is no >"standard" pronunciation. [TC:] Well, there is, sort of. At least prescriptivists try to enforce it (more or less vainly). But you're right, it's certainly not consistent. A striking anecdotal-but-real indication that (some) speakers don't worry about it, though: many years ago while a student in Perugia, I was at a party where there was a young woman addressed by some as Lui[s]a, others as Lui[z]a. My Italian was very elementary at the time, but I had a good ear, so I finally asked her if her name was Lui[s]a or Lui[z]a. The answer was the classic Italian "E` lo stesso." [mcv:] >At least the spelling doesn't enforce any, >like it does in the case of the stops, so there's logic in that. [TC:] Right. [mcv:] >>>According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in >>>approxiamte order of frequency) are not native Tuscan forms, but >>>words imported from the north. [TC:] >>Yes, Rohlfs says so, but this bit isn't that simple, either. [mcv:] >I thought as much. I didn't stop to re-read the whole chapter called >"Ricapitolazione critica dello sviluppo de -k-, -p-, -t-, -s- in >Toscana", but I noted a touch of controversy. Usually happens when >things are not that simple. [TC:] Yes, Rohlfs and a few others were once pitted against Merlo and very few others, in trying to enforce a strict Neogrammarian all-or-none interpretation. The effects linger, with the Rohlfs school having gained the upper hand. But a detailed investigation suggests that something is very much awry. Martin Maiden has a nice, if very brief, treatment in his Linguistic history of Italian, and Luciano Giannelli and I address this rather superficially in Maiden and Parry (eds). 1997. The dialects of Italy. Routledge. [TC:] >[Alto Aragones geminate /nn/, /ll/...] [mcv:] Interesting. I had completely overlooked that (it is mentioned, although very briefly, in my copy of "Dialectologma espaqola" by Alonso Zamora Vicente). [and I have nothing further to add at this time] ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From ratcliff at fs.tufs.ac.jp Sat Nov 14 20:47:39 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Sat, 14 Nov 1998 15:47:39 EST Subject: reconstruction methodology Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > ----------------------------Original > message---------------------------- > Robert R. Ratcliffe wrote: > > > > How do we test it then?-- By > > implication. Each reconstruction (of a proto-phoneme for example) > has > > implications for the whole system of the proto language (the whole > > phonological system, eg), for the development path leading from the > > proto-language to the attested languages (the sequence of sound > changes, > > eg), and for the forms of the reflexes in the descendant languages. > > Only the last is directly observable, of course, and only this real > data > > can be used to rule out a proposed reconstruction absolutely. > > But here is where the iteration comes in. The first attempt at > reconstruction of a protolanguage *X will be based on N languages. > If we add the (N+1)st language then *X might have to be changed. That's right. It happens all the time. I'll quote something to the point:"We find in this development an exact parallelism to the manner in which scientific ideas generally arise, develop and change. They are created to point out the common part in a variety of observed phenomena... At first almost any idea will do.. afterwards, the inconsistencies of the first trial make themselves felt; the first idea is then changed to meet better the new requirements. For a shorter or longer time the facts and ideas may remain in accord, but the uninterrupted increase in empirical knowledge involves sooner or later new fundamental alterations of the general idea, and in this way there is a never-ending process of adaptation of ideas to facts." (Wilhelm Ostwald, from the article on chemical element in the Encyclopedia Brittanica, 1911) > > We might find another language y to add to the family. How many > correspondences do we need? > > Even worse, if the similarity of the language to other languages is > not considered, it will be added to the most established, largest > family, and it will continue to snowball. I don't believe this happens very often. It generally seems to take a long time before specialists in an established family are willing to accept newly described or newly discovered languages as members of the family. It was a while before Hittite was accepted as IE. And it has certainly taken a long time for Semitists to recognize that their family is simply part of a larger family whose other members are found in Africa. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From DISTERH at UNIVSCVM.SC.EDU Sat Nov 14 21:24:33 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Sat, 14 Nov 1998 16:24:33 EST Subject: discussion ended Message-ID: Dear Colleagues, By now the discussion about phonological reconstruction and mathematical models has reached the limits of its usefulness to the entire list. I will no longer post discussions on this topic. However, the few people that are involved in this discussion may wish to continue it off-list. I apologize for any inconvenience this may cause you. Dorothy Disterheft Moderator, HISTLING From bwald at HUMnet.UCLA.EDU Tue Nov 17 00:18:57 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Mon, 16 Nov 1998 19:18:57 EST Subject: intervocalic devoicing in Welsh (?) Message-ID: ----------------------------Original message---------------------------- In agreement with Alan King's last message on Welsh intervocalic devoicing, I hope that some Welsh experts will indeed enlighten us further on the situation. Meanwhile, the possibility of the phenomenon raised a question of a more general nature in my mind. I embedded it in my last message, but I would like to restate it more elegantly. It concerns the phonological conditions on sound change in the following way. Suppose, for the sake of argument, that one or more intervocalic voiced stops geminate. Let's express this as: [+voiced] > [+geminate] / V [+stop,__] V Maybe that's already odd, if [+geminate] implies an INTERRUPTION of voicing ONLY in intervocalic position. Presumably, however, this change is attested in Northern Welsh. Thus, odd or not, it becomes an interesting fact about possible sound change under some conditions. It's really the next stage that raises the issue that I'm most interested in. Suppose that next, the voiced geminate devoices (somehow avoiding merger with a previously existing voiceless geminate -- that raises all kinds of questions, but that's not my main concern here). My main concern is that since (or if) geminates only occur in intervocalic position, then we can express the sound change simply as: [+geminate] > [-voiced] The point is that there is no need for a condition. This could be a trick of economical notation, OR it could be a claim that if there WERE non-intervocalic geminates they would also devoice. I suggest that the sound change is the latter (it's not a notational trick). In that case, it is not "intervocalic" devoicing, where "intervocalic" implies a conditioning factor. That is, intervocalic position has nothing to do with the sound change, even though the change only has the OPPORTUNITY to occur in intervocalic position. There are plenty other cases of such a sequence of sound changes, abstractly: 1. X > Y /W__Z conditioned 2. Y > Q unconditioned But what other cases are there of sound changes where a. Y only exists in context W__Z AND b. X > Q / W__Z is an UNEXPECTED sound change. From mccay at redestb.es Tue Nov 17 13:01:29 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 17 Nov 1998 08:01:29 EST Subject: intervocalic devoicing in Welsh (?) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- >In agreement with Alan King's last message on Welsh intervocalic devoicing, >I hope that some Welsh experts will indeed enlighten us further on the >situation. I indeed hope so, but if nobody volunteers and the matter is of sufficient interest I *could* dig into my books and/or bother a colleague or too to get some more information. I would rather not put so much time into this, though. >Meanwhile, the possibility of the phenomenon raised a question >of a more general nature in my mind. I embedded it in my last message, but >I would like to restate it more elegantly. It concerns the phonological >conditions on sound change in the following way. > >Suppose, for the sake of argument, that one or more intervocalic voiced >stops geminate. Let's express this as: [+voiced] > [+geminate] / V >[+stop,__] V Actually if we're looking at the northern Welsh model, the [+voiced] bit is strictly unnecessary. The change can be represented simply as: [+stop] > [+geminate] / V __ V, which I should think is intuitively more plausible. Of course, this occurred under the circumstance that voiceless stops in the specified position were already geminate (for diachronic reasons that we've already been into). Again, relevant to the Welsh phenomenon is the fact that the gemination probably only occurs when the preceding vowel is stressed, so another modification is necessary: [+stop] > [+geminate] / [V, +stress] __ V. I would suggest there are two ways of "understanding" this change. One is just in the terms implied by the foregoing notation: northern Welsh speakers "decided" to geminated all stops between a preceding stressed and a following (unstressed) vowel. "For fun", so to speak. The other view might be that since voiceless stops in that position were already always geminate, they "decided" to *simplify* their phonological system by extending the "gemination habit" to all stops, regardless of voicing. Actually they went further than that. Gemination in the said position was extended to *most* consonants; notably including /n/, /l/ and /r/. At this point we might say that they were really getting "carried away", since length in these particular consonants (only) had until that time been phonemic, as it still is in southern Welsh (although phonetically most of the work to distinguish short and long liquids is probably performed by the preceding stressed vowel through a compensatory length contrast). So far from doing such a strange thing as geminating intervocalic voiced stops for no apparent reason, we could say that they pretty much had a "gemination party", geminating nearly everything they found in the position / [V, +stress] __ V. Why? I don't know if it's a motivation or merely an effect, but the outcome of this development is a very characteristic *staccato, almost syncopated* rhythm to northern Welsh speech, since most stressed syllables are pronounced with a short, rather energetic vowel followed by a long and also energetically articulated consonant which seems to go implosion-interruption-explosion. Interestingly, in some speakers at least, the intensity of the stressed and the posttonic unstressed vowels doesn't seem to contrast as much in this "articulatory style" as in, say, English or even Spanish (so that foreign ears may have difficulty interpreting which is the stressed syllable), and it now occurs to me that this may be explained if we assume that the "staccato-ey" features I've described have taken over the function of identifying the position of the stress. In particular it is common for the pitch of the posttonic syllable to be higher than that of stressed syllable. As I already said previously, stops in general tend to be strictly voiceless in pronunciation, with other features such as aspiration doing most of the work of differentiating the voiced and voiceless series. It seems to me that in this form of speech intervocalic geminate stops are *particularly* voiceless: i.e. voicing is interrupted along with everything else in their intervocalic articulation. I believe that the voiced and voiceless series nonetheless remain phonologically fully differentiated in this pronunciation (although some foreign speakers might be excused, again, for getting the signals wrong). But *phonetically* one might argue that you already have intervocalic devoicing here, at least as a subsidiary effect - assuming of course that some voicing is there in the voiced stop series to begin with. And if that is the case, it appears to me that we might describe the phenomenon as one of dissimilation. There are surely enough precedents for that is phonology? When intervocalic stops get voiced, that is assimilation to the voicing of the surrounding vowels, I take it; why then, when the opposite happens it is surely dissimilation. The motivation might be to heighten the contrast between adjacent segments: vowels voiced, consonants unvoiced. Maybe we should look for other instances of *that* phenomenon? Please remember, I repeat, that I am mostly describing my "non-rigorous" *impressions* in the preceding description. I am not by training a phonetician (nor a historical linguist, for that matter). Caveat emptor. Regards, Alan From eska at vtaix.cc.vt.edu Tue Nov 17 16:32:31 1998 From: eska at vtaix.cc.vt.edu (Joseph F. Eska) Date: Tue, 17 Nov 1998 11:32:31 EST Subject: On Welsh voicing Message-ID: ----------------------------Original message---------------------------- I don't have the time at the moment to enter the discussion fully, but one thing to bear in mind is that the phonetic realisation of phonemes is not identical across languages. In Welsh, broadly speaking, phonemic /p b/, for example, are [p^h p] phonetically, i.e., the contrast is one of phonetic aspiration, rather than voicing. There is no process of intervocalic devoicing in Welsh, but there is one of intervocalic de-aspiration for voiceless stops, which, in many dialects, is accompanied by partial voicing. Joe Eska eska at vtaix.cc.vt.edu From P.G.Honeybone at newcastle.ac.uk Wed Nov 18 21:06:33 1998 From: P.G.Honeybone at newcastle.ac.uk (Patrick Honeybone) Date: Wed, 18 Nov 1998 16:06:33 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- In connection with the recent discussion on the possibility of intervocalic devoicing in general, and whether this happened in some dialects of Welsh in particular, Joe Eska wrote: > I don't have the time at the moment to enter the discussion fully, > but one thing to bear in mind is that the phonetic realisation of > phonemes is not identical across languages. In Welsh, broadly > speaking, phonemic /p b/, for example, are [p^h p] phonetically, > i.e., the contrast is one of phonetic aspiration, rather than > voicing. There is no process of intervocalic devoicing in Welsh, but > there is one of intervocalic de-aspiration for voiceless stops, > which, in many dialects, is accompanied by partial voicing. I think that this is an important point, and although Alan King referred to it a couple of times in some of his contributions (e.g... > As I already said previously, stops in general tend to be strictly > voiceless in pronunciation, with other features such as aspiration > doing most of the work of differentiating the voiced and voiceless > series. ... the implications of it weren't taken up. So I thought I'd add my thoughts. If this is right (and I'm afraid I know woefully little about the phonology of Welsh myself, so can't comment) then it fits in with an approach to laryngeal features like [voice] and [aspiration] that's been firming up recently. This is the idea that in a language with a two-way distinction in stops (and other obstruents, too, probably, but most of the work seems to centre on stops), that distinction could be made in two ways phonologically. Either the language uses [voice] *or* it uses [aspiration]. This fits in well with phonological theories that use privative features, but could also be understood as using [+/- voice] *or* [+/- aspiration]. (I should perhaps say at this point that there is some variation among proponents of the idea in terms of the names that they give to the features; some use [aspiration], some [spread glottis], some [tense] and some (in Government Phonology) 'H' but I think this is not really a substantial difference). So, using a system like this, people tend to say that most standard Germanic languages (but not Dutch or even all the dialects of English or German) characterise the difference using [aspiration] (or whatever), whereas Romance languages and Slavic tend (again - not necessarily all of them) to use [voice] instead. The idea also has an obvious connection with the long tradition of using [fortis/lenis] as a feature (-pair) rather than [+/- voice] in the Germanic literature. (A good reference for all this is: Iverson, G. & Salmons, J. (1995). 'Aspiration and laryngeal representation in Germanic'. Phonology 12. 369-396). This means that the conventional transcriptions like /t/ and /d/ can be misleading, because in, say, German, /t/ involves [aspiration] and /d/ is unspecified (in a privative system) whereas in, say, French, /d/ involves [voice] and /t/ is unspecified. So /t/ vs. /d/ in German is not the same opposition as /t/ vs. /d/ in French. And if what Joe Eska and Alan King wrote about Welsh stops is right, then Welsh looks like it uses [aspiration] and not [voice]. Welsh /t/ > /d/ would not be quite the same kind of change as Romance /t/ > /d/, nor would a putative Welsh /d/ > /t/ be quite the same as a putative Romance /d/ > /t/ (which is sort of where we came in). If this is all on the right lines then the different kinds of stops in the different kinds of languages should have different phonological effects (and they do seem to) and might well allow different predictions for the kinds of lenition that the languages could undergo (and I think they do...). The point of what I'm trying to say is: maybe we should be careful about comparing changes in different languages - the symbols we use to transcribe sounds could be pulling the wool over our eyes. But I do have more - if I could just go on for a bit longer... This discussion on types of lenition has been very interesting and one key point is evident in the following exchange between Alan King and Benji Wald: >>B. Wald says: It just happened -- because both directions >>are possible (under certain conditions -- certainly NOT ****z > s >>/V_V). That provoked me to try and think of counterexamples to >>Benji's latter assertion here. > >Good. If I'm wrong, I'd like to know as soon as possible, so that I >don't keep on thinking that and misleading other people. That's my >best pay-off for my interventions on the list. > >He continues: > >>I couldn't come up with an example of z > s /V_V, but I >>did find one of +voice > -voice /V_V. > >Just as good. I would have thought that if such a change occurs it >can't be due to the intervocalic position (just as in the z > s >"reversal" case), and would expect it to occur in that position only >as part of a more general voicing shift (which the intervocalic >position is not able to resist at the time). I've just read an article by Ernst Jahr which seems relevant to this point [Jahr, E. H. (1989) 'Language planning and language change' In: Breivik, L. E. & E. H. Jahr 'Language Change. Contributions to the Study of its Causes'. Berlin: Mouton de Gruyter]. In it, Jahr reports on a change that looks like [+voice] > [-voice]/ V_V, or at least, [+voice] > [-voice]/V:_V, in 'upper-class' Oslo Norwegian. The basic details are: in 1880, in words such as the following, with a long vowel, stop, vowel sequence (and I think it basically really is just this environment), the stop was voiced in the pronunciation of this group of people, whereas now the stop is voiceless: 1880 now 'creator' [ska:b at r] [ska:p at r] 'know' [vi:d@] [vi:t@] 'realm' [ri:k@] [ri:k@] ('@' = schwa) So this looks like it might be an example of the 'unnatural' change that Benji Wald was looking for. But Jahr connects this with the language planning movement in Norway and the connected change in the spelling of the words (they used to have , and , now they have

, and ), so it wouldn't count as a 'normal', 'real' sound change. *And anyway* Norwegian, as far as I'm aware, has aspiration in 'voiceless' stops and little or no voicing in 'voiced' stops, so the symbols might be confusing, and ... I think what I'm trying to say is, even if we did find an example of [+voice] > [-voice]/ V_V it might be explicable in other ways and not be a counter-example to lenition tendencies. From manaster at umich.edu Thu Nov 19 18:08:38 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Thu, 19 Nov 1998 13:08:38 EST Subject: Arm. targal 'spoon' (fwd) Message-ID: ----------------------------Original message---------------------------- Dear historical linguists, I have as some of you may know proposed an Armenian sound law whereby inter alia PIE *w ends up ultimately as j^ after *r(H) or *l(H) in Armenian, e.g., olj^ is then simply *solwo-. The only counterexample that I am aware of is targal if this is really, as many think, from *drwaHlaH or the like. So the obvious question is can anyone think of an alternative etymology. One (maybe even the) possibility would be some proposal whereby targal is from *tarigal or *tarugal with the regular Armenian syncope of the relevant vowels. Any takers? AMR From mccay at redestb.es Thu Nov 19 17:55:05 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 19 Nov 1998 12:55:05 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- Having read Patrick Honeybone's latest contribution to the discussion I (rather unwittingly) triggered off, I have probably little of substance to add, but just a couple of related questions and thoughts. Just for clarification, if anyone cares: I am not a specialist in the subject of this discussion (as I've said several times already), and furthermore I haven't got a theory to defend, I'm more of an onlooker sitting on the sidelines and throwing in an idea or two, for what (if anything) it's worth, from time to time. And as such, if I make any silly mistakes, I expect to be corrected. On the question of aspiration versus voicing in Welsh: >I think that this is an important point, and although Alan King >referred to it a couple of times in some of his contributions (e.g... > >> As I already said previously, stops in general tend to be strictly >> voiceless in pronunciation, with other features such as aspiration >> doing most of the work of differentiating the voiced and voiceless >> series. > >... the implications of it weren't taken up. I confirm once again, that according to every phonological description I've ever seen, as well as my own acoustic impressions (the effect is unmistakable, at least if you know a minimum about phonetics and your ear is accustomed to hearing a language with a *voicing* opposition, as is my case): in Welsh stops there is definitely a contrast, and no doubt an opposition, in terms of aspiration, while voicing would seem to be secondary (if the contrast is there at all in phonological terms). In most contexts all stops sound pretty voiceless. The proof of the pudding is that there is, or was, an important colony of (originally, anyway) Welsh speakers, dating from a hundred years ago, in the Patagonia region of Argentina. Believe it or not (this is such a fascinating story, I can't resist telling it), the original emigrants are said to have moved in an organised attempt to ensure that the Welsh language remained alive *somewhere* in the world even if it disappeared in Wales, and had the permission and complicity of the Argentinean government to colonise a region of pampa (and some indigenous inhabitants, it should be added) to this end. Life being what it is, nowadays Spanish seems to have taken over most functions in Welsh Patagonia, which has been largely Argentineanized by now, but there is still a significant residue of maintenance of group identity and even some linguistic maintenance. Predictably, the Welsh back in Wales find all this very interesting, and some of the Patagonians are drawn to their ancestral homeland, so some ties exist. Okay, back to Welsh stops. I've never actually met a Patagonian, but from what I've heard, the Welsh in Wales have a notion of what a "Patagonian" accent is, and one of the most salient features by which "you can tell them apart immediately" is their pronunciation of the stops, since the more recent generations, who are Spanish speaking, apparently assimilated the Welsh system to the Castilian one. Presumably the younger Patagonians have no difficulty with proverbial obstacles to the acquisition of Welsh pronunciation such as its famous voiceless lateral fricative, but are unable to aspirate a /p/! (Or more likely, are unable to *not* voice a /b/ while still keeping it distinct from a voiceless aspirated /p/.) By several reports, native speakers of Romance languages who learn to speak Welsh fairly well all sound like "Patagonians" to the native Welsh. >And if what Joe Eska and Alan King wrote about Welsh stops is right, >then Welsh looks like it uses [aspiration] and not [voice]. Welsh /t/ >> /d/ would not be quite the same kind of change as Romance /t/ > /d/, >nor would a putative Welsh /d/ > /t/ be quite the same as a putative >Romance /d/ > /t/ (which is sort of where we came in). > >If this is all on the right lines then the different kinds of stops in >the different kinds of languages should have different phonological >effects (and they do seem to) and might well allow different >predictions for the kinds of lenition that the languages could undergo >(and I think they do...). Maybe we should try to be more explicit about the kinds of change we are predicting to be possible in each kind of phonological system. Let's take this set of four logical possibilities for the kind of distinction in stops and the theoretical possibility of intervocalic "lenition" (admitting, for convenience, that "lenition" might, until demonstrated otherwise, be either voicing or deaspiration) and "de-lenition" (with apologies: I just made the word up, to denote the reverse change): AC: voicing distinction in stops intervocalic stops may lenite AD: voicing distinction in stops intervocalic stops may de-lenite BC: aspiration distinction in stops intervocalic stops may lenite BD: aspiration distinction in stops intervocalic stops may de-lenite Which of these conditional relations between the type of distiction in stops and the possibility of a certain type of change are we hypothesizing to be true? If I understand him, Patrick Honeybone suggests the following: +AC, -AD, +BD (i.e., in voicing systems, stops may lenite but cannot de-lenite, while in aspiration systems, stops may de-lenite). He doesn't say anything about BC: in aspiration systems is lenition also admitted? I would think it both theoretically and methodologically interesting to encompass this question in such a hypothesis: it would certainly give you more predictions to test. On the other hand, I think there is a danger in being excessively categorical in matters like this, things may not be so black and white. One way in which this is true was pointed out by Patrick Honeybone in the last point in his message: >The basic details are: in 1880, in words such as the following, with a >long vowel, stop, vowel sequence (and I think it basically really is >just this environment), the stop was voiced in the pronunciation of >this group of people, whereas now the stop is voiceless: > > 1880 now >'creator' [ska:b at r] [ska:p at r] >'know' [vi:d@] [vi:t@] >'realm' [ri:k@] [ri:k@] > >('@' = schwa) > >So this looks like it might be an example of the 'unnatural' change >that Benji Wald was looking for. But Jahr connects this with the >language planning movement in Norway and the connected change in the >spelling of the words (they used to have , and , now they >have

, and ), so it wouldn't count as a 'normal', 'real' >sound change. *And anyway* Norwegian, as far as I'm aware, has >aspiration in 'voiceless' stops and little or no voicing in 'voiced' >stops, so the symbols might be confusing, and ... I think what I'm >trying to say is, even if we did find an example of [+voice] > >[-voice]/ V_V it might be explicable in other ways and not be a >counter-example to lenition tendencies. Precisely. Examples such as this one, which is very nice, give me much pause for thought. However, I would emphasise that this sort of problem is by no means limited to cases where language planning as such has intervened. I am reminded of the (only apparent) enigma in the evolution of modern English vowels. I don't remember the details, but around about he seventeenth (?) century there is every indication that the vowels in "sea" and "say" had merged, while that in "see" remained distinct; whereas in present-day English, the first two have miraculously become differentiated once again, and it is "sea" and "see" that are homophonous. Now as far as I know the only way to solve this paradox must involve the fact that languages are heterogeneous systems at many levels; in other words, historical sociology must be invoked of necessity. It seems fairly obvious that what happened is that there were several phonological varieties of English in coexistence and, at some point, competing for acceptance as the prestige and/or "official" model. My own sobering conclusion is that, in any language at any time, beneath the "calm" surface of a language's "general" description, there are undoubtedly undercurrents that may be far more turbulent (and which seriously affect the fishing, I expect!). So one question I'd like to ask is: okay, present-day Welsh stops (Patagonian Welsh excepted) have an "aspiration" system. But for how long has this been the case in the history of Welsh? I myself have no idea. What I do know, partly thanks to Patrick Honeybone, is that, for example, some Germanic languages are "aspirating", but some are "voicing"; it follows that in the history of the various Germanic languages, some must have shifted from one type to the other. In which languages, when, and what type do we think Proto-Germanic was? Furthermore, since changes occurred, some languages at some stages must have been in transition. More interestingly (or worryingly), the "transition" may have consisted of periods of heterogeneity and internal competition within the language. Extending that principle: if Welsh, most Germanic, etc. are "aspirating", but Romance, most Slavic, etc. are "voicing", then some of all these languages must also have shifted type at some stage or other. So my general question is: how confident can the historical linguist be in (a) asserting and (b) demonstrating constraints on language change such as these? Granting that the constraints on changes are true per se at a micro-level, it would seem that the cumulative factors that may have influenced the final outcome of an observable change in a language at the macro-level are always potentially capable of scrambling things to the extent that prediction cannot be confidently maintained. There is the "Patagonian factor": articulation habits may be altered quite suddenly, in structurally unpredictable ways. There is the "see-sea-say" factor: owing to "shifting undercurrents", a language can perform apparent about-turns, even in violation of historical linguistic "laws". There is the heterogeneity factor which explains this: generalizations may not be as generalizable as they appear across dialectal or sociolinguistic varieties. There is also the notational factor observed by Patrick Honeybone: "we should be careful about comparing changes in different languages - the symbols we use to transcribe sounds could be pulling the wool over our eyes". And so on. It is because I perceive such a mass of potentially conflicting factors in real-life linguistic evolution that I tend to balk at assertions in historical linguistics like that of Benji's which sparked off this debate: >It just happened -- because both directions >are possible (under certain conditions -- certainly NOT ****z > s >/V_V). and, when confronted by such, am usually led to a position of devil's advocate: >That provoked me to try and think of counterexamples to >Benji's latter assertion here. Perhaps, strictly speaking, I failed to come up with an instance of intervocalic DEVOICING. Nonetheless, I feel that the ensuing debate has brought to the forefront to what extent a priori diachronic laws probably require a great deal of qualification and specification, so much so that, I suspect, their empirical verifiability could be doubtful. Alan From bwald at HUMnet.UCLA.EDU Thu Nov 19 17:16:11 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Thu, 19 Nov 1998 12:16:11 EST Subject: On Welsh voicing Message-ID: ----------------------------Original message---------------------------- Joe Eska writes: > In Welsh, broadly >speaking, phonemic /p b/, for example, are [p^h p] phonetically, i.e., >the contrast is one of phonetic aspiration, rather than voicing. >There is no process of intervocalic devoicing in Welsh, but there is >one of intervocalic de-aspiration for voiceless stops, which, in many >dialects, is accompanied by partial voicing. I feel I am not risking too much to read into this message that he is suggesting that Welsh fits into "natural" sound change with respect to voicing in intervocalic position, i.e., that the Welsh varieties at issue are variably indulging in an EXPECTED sound change: [-asp] -> <[+voi]> / V_V (if this is indeed restricted to intervocalic position). Perhaps there are Welsh varieties for which the sound change is completed as well, but Patrick Honeybone's message warns against taking transcriptional practices or changes in orthography at face value (with regard to culling the literature on Welsh without listening to the sounds themselves). A deeper issue, then, implicates the situation in Proto-Celtic (whether there were true intervocalic voiced stops), and perhaps even ultimately whether I-E "glottalic" theory is significantly reflected in Celtic (or Proto-Celtic). That is, either voiced stops existed at some point intervocalically BEFORE Welsh emerged, in which case there was devoicing under some conditions (not made clear), or, as far back as we can go into the ancestry of Welsh there were NEVER (consistent) intervocalic voiced stops. Alan King's message is more complex, and contains many interesting ideas, some of which intersect with Honeybone's, where Patrick emphasises the issue of distinctive or "primary" vs. secondary features. To take Patrick's point first, the issue of the nature of sound change remains. As far as I know, the idea that [+stop] > [-voiced] /V_V (and ONLY in V_V) would be an "unnatural" change, REGARDLESS of whether [voiced] was a distinctive feature of Welsh or not. That seems to be because the common change, [+stop] > [+voiced] / V_V has a PHONETIC explanation which leads to unexpectedness of the reverse change on PHONETIC grounds alone. It is not clear to me to what extent distinctive vs. "redundant" features has ever been admitted among practicing historical linguists as a mitigation of the phonetic explanation. It's clearly an issue of basic and general interest. Can phonology (patterns of phonetic organisation in language) mitigate phonetically regular (and Neogrammarianly "blind" phonetic) changes? Without pausing to think of examples, I suppose so, but I don't know about such a case as intervocalic DEvoicing, with respect to this issue. Now, for Alan's comments and suggestions. >Actually if we're looking at the northern Welsh model, the [+voiced] bit is strictly unnecessary. The change can be represented simply as: [+stop] > [+geminate] / V __ V, which I should think is intuitively more plausible. Of course, this occurred under the circumstance that voiceless stops in the specified position were already geminate (for diachronic reasons that we've >already been into). This feeds into the issue of phonological mitigation of "blind" or strictly phonetically motivated change. It did give me pause in my last message, but I accepted it as a fact (provisionally). Next, Alan writes: >Again, relevant to the Welsh phenomenon is the fact that the gemination probably only occurs when the preceding vowel is stressed, so another >modification is necessary: [+stop] > [+geminate] / [V, +stress] __ V. Off the top of my head, the +stress condition may be important, since the following vowel is then less prominent, and hence its features, such as voicing, may play less of a role in inhibiting a devoicing motivated by other factors (such as "pattern pressure"). That is, devoicing following a stressed vowel may be more "natural" than preceding a stressed vowel (due to anticipation of its maximally prominent voicing). Still, I don't know that the thesis of "naturalness" of intervocalic voicing of stops has ever considered such a distinction in offering its explanation in terms of strictly phonetic motivation. Alan again: >I would suggest there are two ways of "understanding" this change. One is just in the terms implied by the foregoing notation: northern Welsh speakers "decided" to geminated all stops between a preceding stressed and >a following (unstressed) vowel. "For fun", so to speak. I don't get the point of the personification. I'll ignore it. Alan continues: > The other view might be that since voiceless stops in that position were already always geminate, they "decided" to *simplify* their phonological system by extending the "gemination habit" to all stops, regardless of voicing. That seems more plausible to me. The change is [stop] > [+geminate]. NO CONDITIONING. For maintaining distinctions, however, that seems to necessitate accepting that there were previously voiced stops, which goes against what Esko seemed to be suggesting. Actually, it doesn't if the distinction affects only an inventory of [+/- aspirate] stops. The aspirate dimension remains after the change, e.g., -tth- vs. -tt-. The further change -tt- > -dd- (> -d-), or whatever, can be seen as maximal differentiation of +asp and -asp, but intervocalic position might be unnecessary and remains problematic (so far) as a CONDITIONING factor. Next Alan writes: >Actually they went further than that. Gemination in the said position was extended to *most* consonants; notably including /n/, /l/ and /r/. At this point we might say that they were really getting "carried away", since length in these particular consonants (only) had until that time been phonemic, as it still is in southern Welsh (although phonetically most of the work to distinguish short and long liquids is probably performed by the >preceding stressed vowel through a compensatory length contrast). Generalisation to resonants (or whatever you call them) detracts from the "maximal differentiation" suggestion, UNLESS there were previously resonant aspirates, which do exist in some languages I'm familiar with, and therefore is not a priori implausible. (I suppose Welsh /ll/ might even count as a liquid "aspirate' -- I'm unfamiliar with its historical sources. It only now occurs to me that "devoicing" (aspiration?) of Spanish /rr/ in some varieties as opposed to /r/ may be another case of maximal differentiation, but I'm accustomed to viewing this tendency, where it exists, as part of the general Spanish tendency to devoice "fricatives", cf. Buenos Aires variable /zh/ > /sh/, significantly from /y/ (?and /ly/?), which becomes /dzh/ in much of Spanish, where prior existence of /ch/ might play a role in keeping devoicing of the voiced affricate at bay -- there's also common Spanish /ch/ > /sh/, in many areas, but that's another matter) Alan again: >So far from doing such a strange thing as geminating intervocalic voiced stops for no apparent reason, we could say that they pretty much had a "gemination party", geminating nearly everything they found in the position >/ [V, +stress] __ V. I need to be assured that the conditioning factor is indeed necessary. I had previously suggested (on the basis of incomplete data) that the gemination was unconditioned but, perhaps, has only been recognised in intervocalic position, i.e., it reflects a more general paradigmatic shift among stop series. Alan's next comments are most interesting: >Why? I don't know if it's a motivation or merely an effect, but the outcome of this development is a very characteristic *staccato, almost syncopated* rhythm to northern Welsh speech, since most stressed syllables are pronounced with a short, rather energetic vowel followed by a long and also energetically articulated consonant which seems to go >implosion-interruption-explosion. Implicit in this suggestion is a METRIC explanation for the change in terms of syllabic rhythm in words. This is certainly worthy of further serious discussion. There's a chicken-and-egg problem involved. It also somewhat (but not completely) overlaps with my earlier suggestion that Welsh stressed vowels may be much more prominent and influential on (preceding) consonants than unstressed vowels, relevantly with respect to voicing. It would indeed be interesting (and relevant) if Welsh unstressed vowels had any tendency to devoice. I've never heard anything along those lines for Welsh. Variable devoicing is a well known feature of Swahili (and some adjacent languages) POST-stress vowels (which are necessarily word-FINAL, and voicing alone is not a distinctive feature of Swahili consonants -- the stops written as voiced in Latin or Arabic letters are imploded in first-language Swahili, truly voiced stops are prenasalised -- some other languages in the vicinity borrow words with voiced stops by prenasalising them, e.g., English "soda" > Gikuyu "sonda", Swahili just implodes the "d" in such words. Gikuyu doesn't have imploded or simple "voiced" stops. Meanwhile Gikuyu has variable denasalisation of the prenasals, so the actual pronunciation can be "so(n)da" without reference to the English original, or the faintly possible Swahili intermediary.) Alan continues: >Interestingly, in some speakers at least, the intensity of the stressed and the posttonic unstressed vowels doesn't seem to contrast as much in this "articulatory style" as in, say, English or even Spanish (so that foreign ears may have difficulty interpreting which is the stressed syllable), and it now occurs to me that this may be explained if we assume that the "staccato-ey" features I've described have taken over the function of identifying the position of the stress. In particular it is common for the pitch of the posttonic syllable >to be higher than that of stressed syllable. That does not support what I just had in mind concerning the difference between stressed and post-tonic vowels, but continues the theme of metricity as a crucial feature of the relevant Welsh sound changes. Alan again: >As I already said previously, stops in general tend to be strictly voiceless in pronunciation, with other features such as aspiration doing most of the work of differentiating the voiced and voiceless series. It seems to me that in this form of speech intervocalic geminate stops are *particularly* voiceless: i.e. voicing is interrupted along with everything else in their intervocalic articulation. I believe that the voiced and voiceless series nonetheless remain phonologically fully differentiated in this pronunciation (although some foreign speakers might be excused, again, for getting the signals wrong). But *phonetically* one might argue that you already have intervocalic devoicing here, at least as a subsidiary effect - assuming of course that some voicing is there in the voiced stop >series to begin with. Again, Alan is presupposing an earlier change FROM voiced stops. For the moment this has to weighed against the contrary implication of Esko's remarks, which seem to leave open the possibility that where there is no voicing in Welsh there historically never was. Alan again: And if that is the case, it appears to me that we might describe the phenomenon as one of dissimilation. Exactly. But in the way I described it above as "maximal differentiation" [+asp] = [-voice] vs. [-asp] > [+voice]. I continued to question the conditioning factor, but Alan's comments on metricity leave me hesitant as to what is a more likely (and even more "natural") account. Alana again: >There are surely enough precedents for that is phonology? When intervocalic stops get voiced, that is assimilation to the voicing of the surrounding vowels, I take it; why then, when the opposite happens it is surely dissimilation. The motivation might be to heighten the contrast between adjacent segments: vowels voiced, >consonants unvoiced. The whole issue of dissimilation is complex and requires a lot of thought (I vaguely remember that a paper or a few have addressed this subject at length in general). To be sure, if the PHONETIC explanation of intervocalic voicing is (voicing) ASSIMILATION, then it is most expected that any attempt at explanation of a reverse phenomenon would begin with the notion of DISSIMILATION. I have my reservations, but the "maximal differentiation" notion fits into a general notion of dissimilation on the paradigmatic level. "Maximal differentiation" is a most well-known concept in accounting for the tendencies of VOWEL systems, but I know at least a few examples where this can be posited as an underlying motivation for changes in certain CONSONANTAL systems (even beyond dissimilation of coarticulation effects for consonants). EG, among "Swahili" varieties Bajuni exhibits the change t > ch and nd > ndr, where "ndr" is what English ears hear (and have commented on), but is simply an effect of the release of the retracted consonant and is not heard as an "r" release by Bajunis. The t > ch is a bona fide palatal affricate, but arose from the same retraction . Dissimilation is involved to the extent that these two sounds have become more distinct from the opposing dental stops t. and nd. (which, ironically, are reflexes of earlier *ch and *nj, still maintained in Southern and standard Swahili). Finally, "dissimilation" alone is not sufficient explanation for (problematic) intervocalic devoicing, but I hasten to add that Alan went a lot further than to suggest that it could be. From mcv at wxs.nl Fri Nov 20 12:08:23 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 20 Nov 1998 07:08:23 EST Subject: Arm. targal 'spoon' (fwd) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- manaster at umich.edu wrote: >I have as some of you may know proposed an Armenian >sound law whereby inter alia PIE *w ends up ultimately as >j^ after *r(H) or *l(H) in Armenian, e.g., olj^ is then simply >*solwo-. The only counterexample that I am aware of is targal >if this is really, as many think, from *drwaHlaH or the like. >So the obvious question is can anyone think of an alternative >etymology. One (maybe even the) possibility would be some >proposal whereby targal is from *tarigal or *tarugal with the >regular Armenian syncope of the relevant vowels. Any takers? We could postulate a form like *deru'wa(:)l- or doru'wa(:)l-, with *e or *o > a and regular development of w- before the stressed syllable, but in such a derivation (or compound) I would sooner expect *druwa(:)l-, which would have given **artugal. If the the etymology is, as I suspect, *derwa(:)l-, have you considered the possibility that it's not a counterexample at all? Armenian *w behaves very differently depending on whether it's syllable initial or syllable final and on where the stress used to be. I gather it was lost after the stress in intervocalic position (erkan "millstone" < 'gwra:wen-), it remains as -w in absolute final position after a vowel (naw "ship"), and becomes g- syllable initially, at least if the (PIE) syllable was stressed. The case of , at least, fits none of these patterns, as it occurs after the stress, not intervocalically, and, if the sound law operated late enough, maybe even in syllable-final position (*[h]olw). And on seeing *[h]olw, I wonder if there might be a connection between your proposed *(r/l)w > j^ (by way of -y-?) and what we see in the o- and a:-stems in -i, where the oblique cases have -w- (hogi, hogw-oy "spirit", maybe from *hogw, *hogwosio; and aygi, aygwoy "vineyard" from *aygw, *aygwosio, and thus from plain PIE *oiwa: instead of Pokorny's dreadful *oiwiia:)? That would of course require explaining why *gw < *w was not delabialized here, but seems otherwise a neat solution: after a consonant, syllable final (?) -w > -y, and further develops into -j^ after liquids, and -i after stops. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From manaster at umich.edu Fri Nov 20 12:08:50 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Fri, 20 Nov 1998 07:08:50 EST Subject: Arm. targal 'spoon' (fwd) In-Reply-To: <365dbbf8.116934866@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- On Fri, 20 Nov 1998, Miguel Carrasquer Vidal wrote: > We could postulate a form like *deru'wa(:)l- or doru'wa(:)l-, with *e > or *o > a and regular development of w- before the stressed syllable, > but in such a derivation (or compound) I would sooner expect > *druwa(:)l-, which would have given **artugal. > This does seem a dead end. > > If the the etymology is, as I suspect, *derwa(:)l-, have you > considered the possibility that it's not a counterexample at all? > Armenian *w behaves very differently depending on whether it's > syllable initial or syllable final and on where the stress used to > be. I have, but there are too few examples to hang anything like that on. > And on seeing *[h]olw, I wonder if there might be a connection > between your proposed *(r/l)w > j^ (by way of -y-?) and what we see > in the o- and a:-stems in -i, where the oblique cases have -w- (hogi, > hogw-oy "spirit", maybe from *hogw, *hogwosio; and aygi, aygwoy > "vineyard" from *aygw, *aygwosio, and thus from plain PIE *oiwa: > instead of Pokorny's dreadful *oiwiia:)? That would of course > require explaining why *gw < *w was not delabialized here, but seems > otherwise a neat solution: after a consonant, syllable final (?) -w > > -y, and further develops into -j^ after liquids, and -i after stops. > I was just thinking about this. Wow! Alexis From Roger.Wright at liverpool.ac.uk Fri Nov 20 12:13:42 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 20 Nov 1998 07:13:42 EST Subject: intervocalic devoicing In-Reply-To: <3.0.3.32.19981119115151.006f2a38@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- Alan King says: >The proof of the pudding is that there is, or was, an important colony of >(originally, anyway) Welsh speakers, dating from a hundred years ago, in >the Patagonia region of Argentina. ... Life being what it is, nowadays >Spanish seems to have taken over most functions in Welsh Patagonia, >which has been largely Argentineanized by now, but there is still a >significant residue of maintenance of group identity and even some >linguistic maintenance. Predictably, the Welsh back in Wales find all >this very interesting, and some of the Patagonians are drawn to their >ancestral homeland, so some ties exist. There is, or was, a scheme whereby two such Patagonians a year had a scholarship to attend Bangor University (North Wales). Part of the point of this, for the linguists, is that Welsh is now spoken almost only by bilinguals, and perhaps the "pure" Welsh is that which can be deduced to be the highest common factor of the Welsh-Spanish and Welsh-English bilinguals. > .... is their pronunciation of the stops, since the more >recent generations, who are Spanish speaking, apparently assimilated the >Welsh system to the Castilian one. Argentinian Spanish isn't very Castilian, but, yes, it still has the voicing opposition, normally without aspiration. (Although some Castilian specialists prefer to think the contrast usually called one of voicing is essentially a tense-lax opposition). RW From tbd at mailserv.waikato.ac.nz Mon Nov 23 12:08:52 1998 From: tbd at mailserv.waikato.ac.nz (T. F. Baer-Doyle) Date: Mon, 23 Nov 1998 07:08:52 EST Subject: help dorothy Message-ID: Dear Collegues, Here is some sad news about one of our colleagues which I'm passing on to you. -Dorothy Disterheft ----------------------------Original message---------------------------- Dorothy I'm a complete stranger to linguistics but I have sad news of one of your members. Michael John Edmonds from University of Auckland, New Zealand, died 10 days ago. His death had been expected for some time, sadly, but he went out with all flags flying in Chomsky'sk fashion in a protest against the colonisation of disabled students by able bodied bureaucrats. Should any members of the list wish to send condolences to his Mum, Karen Shaw, they may send emails to my address and I shall pass them on. She would be glad to hear from any friends especially, as most of Mike's communications were via this medium. Many thenks - and sorry I can't negotiate the listserver instructions that were provided. Dr T. F. Baer-Doyle University of Waikato. Dr T F Baer-Doyle University of Waikato Private Bag 3105 Hamilton, New Zealand Ph (direct line) 64 7 838 4907 (w) Ph 64 7 824 1991 (h) Fax 64 7 838 4434 email. tbd at waikato.ac.nz From mccay at redestb.es Mon Nov 23 12:10:47 1998 From: mccay at redestb.es (Alan R. King) Date: Mon, 23 Nov 1998 07:10:47 EST Subject: intervocalic devoicing: language nomenclature In-Reply-To: Message-ID: ----------------------------Original message---------------------------- I wasn't going to bother, but as there isn't much excitement here at the moment I'll go for it, and defend myself on a "correction" of one of my last posts. Roger Wright "pulled me up" on my referring to "Spanish" as "Castilian". I said: >> .... is their pronunciation of the stops, since the more >>recent generations, who are Spanish speaking, apparently assimilated the >>Welsh system to the Castilian one. to which Roger remarks: >Argentinian Spanish isn't very Castilian, but, yes, it still has the >voicing opposition, normally without aspiration. For the record, it is perfectly commonplace, particularly in Latin America, to refer to "Spanish" as "castellano". Agreed, in English the most widespread term is "Spanish", but the two can be considered synonyms (although some other usages have sometimes been suggested). The point is that "Castilian" does *not* normally refer to a certain dialect of the Spanish language, either traditionally or in present usage (although it perfectly well could do, of course). Rather, it refers to the Romance *language* which first sprang up in Cantabria and Castile, became the tongue of the medieval kingdom of Castile, was spread from there by conquest to other parts of the peninsula and thence elsewhere around the world (Argentina included), has also been progressively but still incompletely imposed on other linguistic communities within the state of Spain itself (namely Galicia, Asturias, the Basque Country and the "Paisos Catalans"), and is now known as either the Castilian or the Spanish language. As for the preferred Politically Correct designation, there are different viewpoints (on what isn't there?). Among those with strongly aware of the existence, and indeed official recognition by the present constitutional regime, of languages other than "Spanish" in the "Spanish State" (another PC expression in such circles), there are (at least) two conflicting usages. Many prefer the designation "Castilian" to emphasise that it is not, after all, the only "Spanish" language. A politically more radical position (at least in the Basque Country) says that calling it "Spanish" is perfectly PC. Why? Well, because Galicia, the Basque Country and the "Catalan Countries" aren't really *Spain* anyway.... I offer these remarks, naturally, not to start an argument but as documentation of current alternative usages on a subject perfectly tangential and immaterial to the ongoing discussion, but pertinent for linguists and perhaps of passing interest to the present readership, if only out of curiosity. Alan Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From bwald at HUMnet.UCLA.EDU Mon Nov 23 12:15:05 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Mon, 23 Nov 1998 07:15:05 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- Alan King's most recent message includes the following statement: >There is the "see-sea-say" factor: owing to "shifting undercurrents", a language can perform apparent about-turns, even in violation of historical linguistic "laws". I think the reference is to the same thing usually exemplified with "meet - meat - mate". It is indeed an interesting historical problem, but not one to be treated as casually as the above statement does. The "mystery", well-known in the history of English, is that in Early Modern English, during the Great Vowel Shift, there was a point at which orthoepists and the like reported a merger of "meat" and "mate", more generally, for earlier /E:/ and /ey/, as "meet", and more generally /e:/ rose to /i:/ position. Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A small residue of the apparent more general merger remains in the words, "great", "steak", "break" and a few others (like "drain" < "drean") and many words before "r" in closed syllables, e.g., "tear" (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, etc.) There have been many attempts to solve the problem. Halle is (in)famous for suggesting that the merger took place on the phonetic level but that the earlier *phonemes* remained distinct. He did not produce morphophonemic alternations for the "ea" words that presumably detached themselves from /e:/ (or /ey/) and rose to /i:/, leaving mysterious and mystical the support for the maintenance of an underlying (morpho)phonemic distinction in face of the phonetic merger. A more traditional attempt at solution is to hem and haw about "dialect mixture" (esp in London), such that there were "meet" vs. "meat/mate" merger dialects (earlier prestigious) in contact with "meet/meat" merger vs. "mate" dialects. Somehow the latter dialects replaced the former (in the relevant area) but the "exceptional" words I mentioned above from the former dialects survived the replacement process (as if constituting a substratum). Though complicated, this solution is not totally unreasonable, and was obviously aimed at defending the notion of the regularity of sound change at the expense of seeking any control over lexical borrowing across dialects. More recent theories, aimed at greater reconciliation of the two traditional approaches, involve lexical diffusion of sound change. This does not really require two dialects in contact but a gradual diffusion of a sound change through the lexicon. A variant could propose (I don't know if anyone has made this proposal in print) something like: a lot of "ear" words and the other exceptions were the first to rise from /E:/ to /e:/ and somehow they merged with /ey/. However, as the process of /E:/ > /e:/ continued they did NOT merge with /ey/ but /e:/ caught up with the "ee" words either at or on their way to /i:/. More radical, and, unsurprisingly, my preferred solution, is the one of "near-merger" (Labov). According to this theory, /e:/ and /ey/ did not merge as reported, but remained distinct along a dimension of "tenseness" ("frontedness") that the orthoepists did not recognise (even though they themselves may have unconsciously made the distinction). So, the fronter (tenser) /e:/ remained distinct from the front (but less front, laxer) /ey/ and continued to rise to /i:/, to merge with the "ee" words. There was phonetic conditioning holding back many of the "ea" words before same-syllable /r/, and they indeed did merge with "-are" and "-air". A similar laxing effect is still operant for most vowel before (consonantal) -r in current dialects. Similarly, pre-vocalic r (as in great, break, drean = drain) and post-vocalic k (as in steak, break) retarded further raising (although "creak", "streak" etc), a statement about "mini" phonetic conditioning effects during the variable stage of a sound change. None of the above solutions are totally satisfying or above criticism (and further research for testing), but I don't think one should throw up one's hands and cavalierly talk about "apparent about-turns, even in violation of historical linguistic "laws". Instead, I think one should recognise the challenge to investigate the matter further. It is the next frontier to be crossed in understanding the nature of sound change, and indeed the nature of linguistic "laws". (I guess "laws" has a double-meaning, first, man-made/artificial and ultimately fallable attempts to explain/account for a phenomenon, which I take to be the point of Alan's scare quotes, and, second, the actual principles governing linguistic behavior and change, which I take to be the purpose of linguistic research and the way linguists channel and DISCIPLINE their curiosity about the facts of language.) From mccay at redestb.es Tue Nov 24 13:09:59 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 24 Nov 1998 08:09:59 EST Subject: Apparent violations of historical linguistic laws In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Benji Wald says: >None of the above solutions are totally satisfying or above criticism (and >further research for testing), but I don't think one should throw up one's >hands and cavalierly talk about >"apparent about-turns, even in violation of historical linguistic "laws". >Instead, I think one should >recognise the challenge to investigate the matter further. It is the next >frontier to be crossed in understanding the nature of sound change, and >indeed the nature of linguistic "laws". (I guess "laws" has a >double-meaning, first, man-made/artificial and ultimately fallable attempts >to explain/account for a phenomenon, which I take to be the point of Alan's >scare quotes, and, second, the actual principles governing linguistic >behavior and change, which I take to be the purpose of linguistic research >and the way linguists channel and DISCIPLINE their curiosity about the >facts of language.) On a first reading, I agree with much of what this paragraph seems to be saying, which could even be taken as a summary of what I was trying to say. Being a complex statement, however, there is room for many subtle changes of emphasis among the various points, and some of Benji's wording is perhaps a little unfair. I don't actually remember "throwing up my hands". And I would certainly be in favour of "recognising the challenge to investigate the matter further". But the real crux of my message was in these words, reproduced by Benji but perhaps not read quite in the way I had intended: "apparent about-turns, even in violation of historical linguistic 'laws'". The above "summary" focuses on the word *laws*, and indeed defends the concept against what is seen as my attack. But in fact, I was more interested in the word *apparent*. The *interesting* thing about problems such as the one I was recalling (which I didn't claim to have discovered, even though I seem to have made the mistake of calling it the "see-sea-say" problem rather than the "meet - meat - mate" problem - silly me!) is that they serve to exemplify the possibility that what *apparently* happens may be different from what really does. We notice that the *apparent* reversal of the merger of "meet" and "mate" CAN'T be exactly what happened BECAUSE it would violate the laws which we believe to operate in such cases. IF on the other hand the apparent event had, by chance, NOT violated our law, we would not be aware of a "problem" and in consequence would probably not question the reality or exact nature of the said event. All the same, I don't think the "laws" (hence, in the last resort, my scare quotes) ought to be seen as infallible and unassailable, since they do rest for their validation on observed events (or *apparently* observed ones, anyway!). And at the risk of repeating a cliché: we are in a privileged position when discussing the history of a language as well documented as English, which makes it easier to "catch ourselves out". Where most languages and language families are concerned, it is presumably easier to get away with our mistakes. As for the emphasised word DISCIPLINE, I am all for it, but there is more than one kind. I think that true intellectual discipline should also encompass a constant preparedness to criticise one's own theories, even looking for *possible* cracks in the foundations, rather than blindly building higher and higher. But then, I happen to think that the ultimate goal is not to finally discover "all" the laws, but in the process to arrive at a better acquaintance with the phenomena they attempt to synthesize. Fortunately, sometimes this goal is achieved. From DISTERH at UNIVSCVM.SC.EDU Tue Nov 24 14:16:27 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Tue, 24 Nov 1998 09:16:27 EST Subject: reversal of merger In-Reply-To: Message-ID: ----------------------------Original message---------------------------- This message was originally submitted by larryt at COGS.SUSX.AC.UK to the HISTLING list at VM.SC.EDU. If you simply forward it back to the list, using a mail command that generates "Resent-" fields (ask your local user support or consult the documentation of your mail program if in doubt), it will be distributed and the explanations you are now reading will be removed automatically. If on the other hand you edit the contributions you receive into a digest, you will have to remove this paragraph manually. Finally, you should be able to contact the author of this message by using the normal "reply" function of your mail program. ----------------- Message requiring your approval (74 lines) ------------------ On Mon, 23 Nov 1998, Benji Wald wrote: > I think the reference is to the same thing usually exemplified with > "meet - meat - mate". It is indeed an interesting historical > problem, but not one to be treated as casually as the above > statement does. The "mystery", well-known in the history of > English, is that in Early Modern English, during the Great Vowel > Shift, there was a point at which orthoepists and the like reported > a merger of "meat" and "mate", more generally, for earlier /E:/ and > /ey/, as "meet", and more generally /e:/ rose to /i:/ position. > Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A > small residue of the apparent more general merger remains in the > words, "great", "steak", "break" and a few others (like "drain" < > "drean") and many words before "r" in closed syllables, e.g., "tear" > (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, > etc.) > There have been many attempts to solve the problem. For the record, I've recently been collecting published proposals for accounting for apparent cases of reversal of merger. So far I have six, as follows: (1) The merger occurred at the phonetic level, but speakers retained different phonological representations in their heads, allowing later reversal (Halle). (2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had this role from those which did not (Michelena). (3) The merger occurred in the prestige variety but not in less prestigious varieties; since only prestige varieties tend to be well recorded, a shift in prestige shows up in the record as an apparent reversal of merger (Weinreich, Labov and Herzog). (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). (5) The merger never really occurred; instead there was only a near-merger, resulting in the usual inability of speakers to recognize the difference (Labov). (6) The merger occurred for most speakers, but a handful of influential conservative speakers succeeded in reversing the merger by semi-official action (Jahr). Of these, (1) is now dismissed, I think; (2) is unchallenged for the case for which it was proposed; (3) is still widely accepted; (4) is taken seriously but is sparsely documented; (5) is widely defended; (6) is documented for a particular case. I do not claim that these six exhaust the proposals, especially since I still have a small folder of relevant articles to read. My final list may be somewhat longer. Of course, only some of these proposals can account for the cases of "residue" noted by Benji. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From DISTERH at UNIVSCVM.SC.EDU Wed Nov 25 13:43:27 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Wed, 25 Nov 1998 08:43:27 EST Subject: reversal of merger In-Reply-To: Message-ID: ----------------------------Original message---------------------------- ----------------------------Original message---------------------------- This message was originally submitted by larryt at COGS.SUSX.AC.UK to the HISTLING list at VM.SC.EDU. If you simply forward it back to the list, using a mail command that generates "Resent-" fields (ask your local user support or consult the documentation of your mail program if in doubt), it will be distributed and the explanations you are now reading will be removed automatically. If on the other hand you edit the contributions you receive into a digest, you will have to remove this paragraph manually. Finally, you should be able to contact the author of this message by using the normal "reply" function of your mail program. ----------------- Message requiring your approval (74 lines) ------------------ On Mon, 23 Nov 1998, Benji Wald wrote: > I think the reference is to the same thing usually exemplified with > "meet - meat - mate". It is indeed an interesting historical > problem, but not one to be treated as casually as the above > statement does. The "mystery", well-known in the history of > English, is that in Early Modern English, during the Great Vowel > Shift, there was a point at which orthoepists and the like reported > a merger of "meat" and "mate", more generally, for earlier /E:/ and > /ey/, as "meet", and more generally /e:/ rose to /i:/ position. > Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A > small residue of the apparent more general merger remains in the > words, "great", "steak", "break" and a few others (like "drain" < > "drean") and many words before "r" in closed syllables, e.g., "tear" > (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, > etc.) > There have been many attempts to solve the problem. For the record, I've recently been collecting published proposals for accounting for apparent cases of reversal of merger. So far I have six, as follows: (1) The merger occurred at the phonetic level, but speakers retained different phonological representations in their heads, allowing later reversal (Halle). (2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had this role from those which did not (Michelena). (3) The merger occurred in the prestige variety but not in less prestigious varieties; since only prestige varieties tend to be well recorded, a shift in prestige shows up in the record as an apparent reversal of merger (Weinreich, Labov and Herzog). (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). (5) The merger never really occurred; instead there was only a near-merger, resulting in the usual inability of speakers to recognize the difference (Labov). (6) The merger occurred for most speakers, but a handful of influential conservative speakers succeeded in reversing the merger by semi-official action (Jahr). Of these, (1) is now dismissed, I think; (2) is unchallenged for the case for which it was proposed; (3) is still widely accepted; (4) is taken seriously but is sparsely documented; (5) is widely defended; (6) is documented for a particular case. I do not claim that these six exhaust the proposals, especially since I still have a small folder of relevant articles to read. My final list may be somewhat longer. Of course, only some of these proposals can account for the cases of "residue" noted by Benji. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at WXS.NL Wed Nov 25 16:20:27 1998 From: mcv at WXS.NL (Miguel Carrasquer Vidal) Date: Wed, 25 Nov 1998 16:20:27 GMT Subject: reversal of merger In-Reply-To: Message-ID: Larry Trask wrote: [on "reversal of merger":] >I do not claim that these six exhaust the proposals, especially since I >still have a small folder of relevant articles to read. My final list >may be somewhat longer. The one that's certainly missing is "the distinction was reintroduced from outside (a different dialect)". This may be seen as a "spatial" supplement to the socially motivated (3) [change in prestige dialect] and (6) [reintroductioon by conservative minority]. Sorry, I don't have a specific source. I'm fairly sure this is the explanation that Leskien or Brugmann would have given. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From bwald at HUMNET.UCLA.EDU Thu Nov 26 04:45:53 1998 From: bwald at HUMNET.UCLA.EDU (bwald) Date: Wed, 25 Nov 1998 20:45:53 -0800 Subject: reversal of merger Message-ID: I appreciated Larry Trask's enumeration of distinct cases which resemble reversal of merger. Among the cases he mentioned, the following is particularly interesting for the meet:meat:mate case: (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). More examples come to light almost daily, and they tend to be just on the other side of an isogloss of an innovative merger. Speakers report the reference dialect, not their own. According to the Milroys (if I remember correctly -- and I think it is mentioned in Labov 1994 as well), Belfast maintains a separate "meat" class, distinct from both "meet" and "mate" in that the "meat" class varies between both the "meet" and "mate" pronunciations (stressed vowel, that is) while the "meet" and "mate" classes are invariably distinct. If I got this straight -- and I think I have the gist right if not all the details, and it exemplifies Larry's (4) above (though it shows a class which merges with two distinct other classes (or vowels) and has no unique vowel of its own except that it varies between two distinct vowels), then the "residue" in most current varieties of English may have arisen through collapse of the variable class into the two invariant classes. That is, FIRST, Class/Vowel 1 = i: as in "meet", Class/Vowel 2 = e: as in "mate" and the problematic Class/"Vowel" 3 = i: OR e: as in "meat", THEN, Class/"Vowel" 3 breaks up into Classes 1 and 2 but eventually no words remain in both. This is not neogrammarian sound change, but reassignment of the words of a distinct (variable) word class to two distinct word classes. The problem remains why most of the "ea" words were reassigned to /i:/, but much fewer of "ear$" words were. Again, there seems to be phonological conditioning of a familiar type in English here, such that the centralising tendency of a following tautosyllabic r inhibited reassignment of "ear$" words to /i:/ (rather a more mid vowel). How often in practice can we distinguish phonologically conditioned lexical partition of a variable phoneme from sound change? Of course, the Belfast situation may not reflect the situation of the earlier reported London merger of "meat" and "mate" and then unmerger and merger with "meet", but, in the absence of compelling evidence to the contrary, it remains a possibility to the extent that such a linguistic possibility as the one described for Belfast is attested. It also goes without saying that it is quite interesting to contemplate how such a variable phoneme could arise in the first place X -- collapse of two "dialects" into one? (more on this later.) I also remember that when I read Labov's (1994) account of the Belfast situation I was struck that he did not report phonetic measurements (of vowel formant structure) for the three word classes at issue, so even if my characterisation of the Milroy's remarks on the Belfast variable class is accurate, it is still not clear that there is indeed absolute merger with the two other distinct word classes, rather than a near-merger (for the mid-level "meat" and "mate"). This problem is aggravated for extrapolating to historical London by the problem of whether or not /ey/ as in "hair", "pain", etc. had merged with /e:/ as in "hare", "pane" etc., in the relevant time period -- it did later, by the end of the 18th c., and they are still distinct in many British areas. That's why I preferred the "meat"-"mate" example to the "sea" - "say", since "say" descends from a diphthong and may have never become a monophthong -- again the "reversal" problem of whether [ay] > [ey] OR [ay] > [a:] > [e:] > [ey] (the "reversal" if the diphthong ever became a monophthong and then diphthongised again). With regard to Belfast, I do not know if the Milroys ever performed acoustic measurements to determine whether Class 3 shows mergers with Classes 1 and 2, or whether they rely simply on ear and self-report (neither of which is sufficient for near-merger). I wonder if I should assume that they did do measurements, given that Labov (1994) does not discuss Class 3 as a possible case of near-merger. With regard to the whole concept of a "variable" phoneme (one that is phonemically but not phonetically distinct from one (or more?) other phonemes), I was struck in the mid 1970s by "unmerger" in second language phonology. The particular case I examined was the distinction between /ch/ and /sh/ in the fluent English of L1 Spanish (Mexican) speakers. Although /ch/ and /sh/ had the same range of allophones (namely, [ch] and [sh] and something in-between), the patterns of realisation of English /ch/ and /sh/ words was different, such that /ch/ words were more often realised as [ch] for the same speaker and vice-versa for /sh/ -- with the same phonetic conditioning for both (the sound [ch] was more favored word initially for both classes and [sh] more favored elsewhere -- but still affected in frequency by the English word class). Further work has been done on such L2 processes since then. Certainly, investigators generally speak of "developmental" processes in such cases, and L1 acquisition is somewhat similar, .but the speakers I looked at were adults, some quite old, and had spoken English for decades (since their early 20s), and their systems seemed to be stable -- as if they had distinct phonemes /ch/ and /sh/, and a single pronunciation rule which variably merged them, with different frequencies in different phonetic positions. Trask's case (2) piqued my curiosity: >(2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had >this role from those which did not (Michelena). I am not familiar with the Michelena example. Is the general case the one that, say, Kiparsky used to explain the general case of Yiddish final voiced stops. That is, for verbs and nouns the lexical morpheme may appear finally or intervocalically due to inflection, e.g., hob-en 'have' but 'hop' '(I) have. The phonological rule is: devoice all final stops. Intervocalic position preserves the word class distinct from the word class of final voiceless stops, which remain voiceless in intervocalic position. Generally, Yiddish has revoiced the final stops when they are voiced in intervocalic position, i.e., Yiddish has withdrawn the rule: devoice final stops. Kiparsky proposed that the former rule can be detected where there was no alternation, e.g., with the adverb 'avek' ("away", historically from 'a-veg'). This kind of "unmerger" is apparently what Halle was hoping he could establish for the shift in "ea" words, but it turned out to be an inappropriate example. But, now, what is the Michelena example? Last but, most emphatically, not least, I want to apologise to Alan King if I misrepresented the thrust of his remarks on the complexity of linguistic, even phonological, even phonetic, change. As I've said before, I am grateful to Alan for his frequent insightful examples and comments. The Welsh problem deserves to become a classic, and I do not doubt that lasting fame will come to the linguist (or group of linguists) who manage to make the linguistic community at large aware of the general theoretical significance and challenge posed by the problem (and to who/mever "solves" the Welsh problem). From larryt at COGS.SUSX.AC.UK Thu Nov 26 15:51:35 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Thu, 26 Nov 1998 15:51:35 +0000 Subject: reversal of merger, proposal (2) Message-ID: People are asking me about this, so I'll spell it out a bit. The case is made for a development in the Gipuzkoan variety of Basque, by Luis Michelena, in his magnum opus Fonetica Historica Vasca and elsewhere. Note the following orthographic conventions in Basque: = laminal [s] = English = the affricate [ts] = English represents the glide [j] (= US [y]) historically, but today represents a variety of consonants, according to dialect, derived from this glide by strengthening. Basque possesses a set of palatal and palato-alveolar consonants which, historically, never occur in ordinary lexical items, but only in expressive variants of these items and in expressive formations generally. For example, `bull' has an expressive variant `little bull', and `corner' has the expressive variant or `nook, cosy little place'. Such expressive palatalization was formerly pervasive in Basque, and it survives today in some varieties. Originally, therefore, the consonants and (among others) occurred *only* in such expressive forms, and nowhere else at all. But then the language acquired new instances of in ordinary lexical items, by borrowing, as in `soap' and `care, attention, concern', both borrowed from Romance. These were not expressive variants or forms, since they had no other form. Meanwhile, the historical glide /j/ was undergoing fortition in most dialects to some kind of consonant. The result differed according to region, but, in much of the country, it became a voiced palato-alveolar fricative [ezh], similar to French . This is still the state of affairs in the eastern extremity of the country. In the center of the country, however, and particularly in the Gipuzkoan dialect, this fricative underwent devoicing to [esh} (English ), thus merging with the historical . This is still the state of affairs today in some southern parts of Gipuzkoa: complete merger. In most of Gipuzkoa, however, there was a later change: (that is, [esh]) was backed to velar [x] or to uvular [X], a change which was apparently borrowed from Castilian Spanish, which underwent the same development. Now, we might have expected *all* instances of to undergo this backing, but that's not what happened. Instances of derived from earlier /j/ underwent backing. Hence native `lord', `eat' and `owner', which had historically contained the glide, underwent backing, and their modern Gipuzkoan forms are <[x]aun>, <[x]an> and <[x]abe>, with velars or uvulars. Instances of borrowed also underwent backing, and so borrowed and became <[x]aboi> and , also with velars or uvulars. But instances of *original* (that is, expressive) did *not* undergo backing, and so expressive , for example, has remained , with the sound of English and no backing. Hence Gipuzkoan has apparently first merged historical /j/ with and then reversed the merger, so that only non-historical instances of underwent backing. Now, one might suppose that the merger never really took place, that the two sounds somehow remained distinct until after the backing of one of them had occurred. But Michelena argues against this as inadequate. For one thing, the backing observed in loan words shows that these words must have contained the same sound that developed from original /j/. For another, those varieties that failed to undergo the backing invariably exhibit a complete merger. But there's more. In one or two cases, the expressive value of had apparently been lost -- and *these* instances of *did* undergo backing. Consider the word `man'. This has the regular combining form , and this in turn has the regular expressive variant . >From this , Basque formed a derivative `poor fellow', with the rare diminutive suffix <-xo> (diminutive suffixes, being intrinsically expressive, always contain the special consonants). And this has developed in Gipuzkoan to the unexpected , in which the first instance of has failed to back while the second instance has undergone backing to the velar/uvular [x]/[X], notated . Michelena's explanation is therefore the following. Instances of which were expressive in nature clearly retained their expressive function, and hence they remained palato-alveolar and did not undergo backing -- because they were part of a wider system of palatal and palato-alveolar consonants confined to expressive forms. All other instances of -- that is, those derived from /j/ and those in loan words -- lacked this expressive value and hence could and did undergo backing. In the case of , the expressive value of the first remained obvious, and hence this did not back. But, perhaps because of the rarity of diminutive <-xo>, the expressive value of the second was lost, and so this one underwent backing. This account seems to explain the facts admirably, and, to my knowledge, it has never been challenged. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Thu Nov 26 14:31:23 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Thu, 26 Nov 1998 14:31:23 +0000 Subject: reversal of merger In-Reply-To: <365f2e22.604986038@mail.wxs.nl> Message-ID: On Wed, 25 Nov 1998, Miguel Carrasquer Vidal wrote: > Larry Trask wrote: > > [on "reversal of merger":] > >I do not claim that these six exhaust the proposals, especially since I > >still have a small folder of relevant articles to read. My final list > >may be somewhat longer. > > The one that's certainly missing is "the distinction was reintroduced > from outside (a different dialect)". This may be seen as a "spatial" > supplement to the socially motivated (3) [change in prestige dialect] > and (6) [reintroductioon by conservative minority]. Sorry, I don't > have a specific source. I'm fairly sure this is the explanation that > Leskien or Brugmann would have given. Yes. I've been looking for a good explicit source for a proposal along these lines, but so far the only mentions I've dug up have been too inexplicit to constitute a formal proposal. Indeed, the ones I've seen have amounted to little more than mouthing mantras. I would be grateful for a reference to a source that offers a fully explicit proposal of this type. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Fri Nov 27 15:50:40 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Fri, 27 Nov 1998 15:50:40 +0000 Subject: reversal of merger, proposal (2) In-Reply-To: Message-ID: On Fri, 27 Nov 1998 Alexis Manaster Ramer wrote: > I think that there is at least one alternative. Truly > expressive forms are often pronounced somewhat differently > from non-expressive forms with the same phonemes. As Labov > points out, for example, the f-word in English often has > the /f/ made with the lower lip tucked in. Sapir's > paper on "abormal speech" in Nootka contains several > examples. > What would fit > all the facts cited is the theory that there were two > somewhat different articulations of what is written > as , which I will write as and , and that > only one of these, (x2), went to . Note that the strikingly > anomalous development of to , *, > can be explained if we assume that, once the force of the > dim. suffix - . I agree that this interpretation of the Basque case is perfectly possible. However, it seems less economical than Michelena's proposal, in requiring speakers of the Gipuzkoan dialect to construct and maintain a contrast between two different kinds of [esh], a contrast which is not observed or recorded in any variety of Basque. Moreover, such a contrast would have produced an immensely crowded sibilant system. Quite apart from the further contrast proposed here, Basque already has a contrast among three voiceless sibilants: a laminal (notated ), an apical (notated ), and a palato-alveolar (notated ), and it also has the three corresponding affricates, notated , and . This is already a surprisingly crowded system for a language which has only a very modest number of phonemes to start with. And note further that this proposal would have required speakers to change the articulation of to in words in which the expressive value of had been lost: hence it requires that speakers should always have carefully distinguished between expressive and non-expressive occurrences of , which is precisely what Michelena was proposing in the first place. > Another possibility is that the expressive 's that do > not change to are due to interference from other dialects. > There is in general little doubt that many expressive > features do come from external sources. E.g., in Polish > a few words which normally have get in expressive > forms, at least for some speakers, e.g., , an > obsolescent word for 'young female servant' vs. > roughly 'whore', 'poverty' vs. , used to > express sympathy or the like when talking about some > particular people's poverty, etc. In this case, we > happen to know that and are different dialect > realizations of older /e:/, which earlier in the century > were in competition for standard status. In general, > won this fight, thus making eminently suitable to > an expressive function. Yes, but we are not here dealing with a handful of exceptional words. With just one or two exceptions, like the second in , *every* instance of original (expressive) remained unchanged, while *every* instance of non-original underwent backing. The process was completely systematic, not merely sporadic. And no known dialect of Basque exhibits a contrast between two types of , nor does any dialect have a reflex of expressive which is different from , except that word-initial develops regularly to the affricate in some varieties. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Fri Nov 27 15:26:03 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Fri, 27 Nov 1998 15:26:03 +0000 Subject: reversal of merger, proposal (2) Message-ID: Alexis Manaster Ramer is temporarily cut off from this list because of a fault, and so he's asked me to forward the posting below. If you want to reply, please reply to him and not to me. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk ---------- Forwarded message ---------- Date: Fri, 27 Nov 1998 08:30:21 -0500 (EST) From: manaster at umich.edu To: Larry Trask Subject: Re: reversal of merger, proposal (2) I think that there is at least one alternative. Truly expressive forms are often pronounced somewhat differently from non-expressive forms with the same phonemes. As Labov points out, for example, the f-word in English often has the /f/ made with the lower lip tucked in. Sapir's paper on "abormal speech" in Nootka contains several examples. What would fit all the facts cited is the theory that there were two somewhat different articulations of what is written as , which I will write as and , and that only one of these, (x2), went to . Note that the strikingly anomalous development of to , *, can be explained if we assume that, once the force of the dim. suffix -. Another possibility is that the expressive 's that do not change to are due to interference from other dialects. There is in general little doubt that many expressive features do come from external sources. E.g., in Polish a few words which normally have get in expressive forms, at least for some speakers, e.g., , an obsolescent word for 'young female servant' vs. roughly 'whore', 'poverty' vs. , used to express sympathy or the like when talking about some particular people's poverty, etc. In this case, we happen to know that and are different dialect realizations of older /e:/, which earlier in the century were in competition for standard status. In general, won this fight, thus making eminently suitable to an expressive function. On Thu, 26 Nov 1998, Larry Trask wrote: > The case is made for a development in the Gipuzkoan variety of Basque, > by Luis Michelena, in his magnum opus Fonetica Historica Vasca and > elsewhere. [snip] > Hence Gipuzkoan has apparently first merged historical /j/ with and > then reversed the merger, so that only non-historical instances of > underwent backing. [snip] > Consider the word `man'. This has the regular combining form > , and this in turn has the regular expressive variant . > >From this , Basque formed a derivative `poor fellow', > with the rare diminutive suffix <-xo> (diminutive suffixes, being > intrinsically expressive, always contain the special consonants). And > this has developed in Gipuzkoan to the unexpected Michelena's explanation is therefore the following. Instances of > which were expressive in nature clearly retained their expressive > function, and hence they remained palato-alveolar and did not undergo > backing -- because they were part of a wider system of palatal and > palato-alveolar consonants confined to expressive forms. All other > instances of -- that is, those derived from /j/ and those in loan > words -- lacked this expressive value and hence could and did undergo > backing. In the case of , the expressive value of the first > remained obvious, and hence this did not back. But, perhaps because of > the rarity of diminutive <-xo>, the expressive value of the second > was lost, and so this one underwent backing. From henryh at LING.UPENN.EDU Fri Nov 27 14:28:28 1998 From: henryh at LING.UPENN.EDU (Henry M. Hoenigswald) Date: Fri, 27 Nov 1998 09:28:28 -0500 Subject: Merger reversed Message-ID: Reversed merger is not a matter for empirical confirmation or disconfirmation. In their haphazard but productive fashion, the nineteenth-century scholars devised a framework (still uncritically used, and, incidentally, uncritically used by them in the first place) in which there is no room for reversal of merger by 'sound-change' (='a replacement, along a line of descent, such that it can be stated without naming particular morphs'). By their lights, there was either no merger, or there is more than one line of descent. There are no doubt other possible frameworks but they have not been explored; maybe for good reasons maybe not. HMH From bwald at HUMNET.UCLA.EDU Sat Nov 28 02:05:36 1998 From: bwald at HUMNET.UCLA.EDU (bwald) Date: Fri, 27 Nov 1998 18:05:36 -0800 Subject: reversal of merger, proposal (2) Message-ID: Larry Trask's example of Basque "unmerger" is quite interesting, and not unexpected for anyone who recognises the reality of synchronic phonological(= morphophonemic) processes. He explains the gist as: > For example, `bull' has an expressive variant > `little bull', and `corner' has the expressive variant > or `nook, cosy little place'. Such expressive >palatalization was formerly pervasive in Basque, and it survives today >in some varieties. Similar (and probably related) expressive palatalisation is used in Spanish, e.g., in talking to babies and forming nicknames, e.g., Jo*s*e > *ch*e, vamo*n*os > vamo*ny*o etc. The relevant (part of the) Basque expressive rule seems to be: s > sh/ch (by /s/ I mean the phoneme written "z") If I understand right, the rule remains in the relevant area as an alternation between s and sh/ch in affected lexical items, keeping it transparent and alive. This souce of (t)sh, on a synchronic level, is completely distinct from the historical phonemes sh and ch, which do NOT alternate with s (and NEVER did). [Larry implied above that it is not STILL a productive affective rule in some Basque areas but was not clear on whether it is in the relevant area -- and, in any case, the issue would only be relevant at the time that sh-backing arose; he seems to take it for granted, so I will too.] I did not quite get the following out of Larry's account of Michelena, who seemed not to have any tool of analysis that goes beyond surface phonetic changes, but it seems clear from the data that the sound change of sh > x is easily expressed as arising as an ordered rule which precedes the affective rule. The affective rule is not affected because its INPUT is /s/ not /sh/. If you cling to the phonetic surface for sound change (the traditional concept), well, the affective rule operates on the surface as a MORPHOPHONEMIC RULE FOLLOWING THE SOUND CHANGE. IE, both the affective rule and the sound change operate on the "surface", and the sound change operates FIRST. Larry notes: >Consider the word `man'. This has the regular combining form >, and this in turn has the regular expressive variant . As we expect. Then he notes: >>From this , Basque formed a derivative `poor fellow', >with the rare diminutive suffix <-xo> (diminutive suffixes, being >intrinsically expressive, always contain the special consonants). And >this has developed in Gipuzkoan to the unexpected , in >which the first instance of has failed to back while the second >instance has undergone backing to the velar/uvular [x]/[X], notated . OK. So we get historical confusion for this particular word, so that the expressive origin of the second /sh/ became obscure for some reason. If I understand the above account right, -xo is RARE, and it does NOT alternate with -zo. So, the origin of its /sh/ in an *affective rule* is not obvious. It's a crap-shoot whether the affect of the rest of the word will save this /sh/ from backing -- and the lexical rarity doesn't help build up a case one way or the other. (Still, if the -xo morpheme generally tended to back, then that would support the idea that it is the morphophonemic alternation per se that protected /sh/ from backing, and NOT the "affect" of the /sh/.) According to Larry: >Michelena's explanation is therefore the following. Instances of >which were expressive in nature clearly retained their expressive >function, and hence they remained palato-alveolar and did not undergo >backing -- because they were part of a wider system of palatal and >palato-alveolar consonants confined to expressive forms. As I said, it's worth explicitly mentioning that such forms ALTERNATE with non-palatals. The palatals subject to backing do NOT have such alternations. That's ALL we need to know to make this case similar to other cases of "unmerger" in which a morphophonemic alternation, captured in a phonological rule, distinguishes one phoneme from another but allows them to overlap phonetically in pronunciation. The important fact is that the rule is MORPHOPHONEMIC. The fact that the rule is "affective" is basically IRRELEVANT, as far as I can see, except for the point I make below. With regard to affective rules, something I discounted as relevant above in favor of the alternation, there is something of further interest about it. If in the relevant dialects, there is NO palatal except by the affective rule, then it is NOT like Spanish affect. Spanish affective palatalisation rule does NOT create segments that do not exist independently as lexical phonemes, i.e., /ch/ and /ny/ (also f > p and other irrelevancies which are part of the same affect). Presumably, the Basque rule does. All I can suggest about that is that languages vary as to whether affective (and onomatopoetic) words are constrained by canonical phonological rules. It is notable that Spanish onomatopoetic words for sounds (as represented, for example, in cartoons) can end in stops, while canonical words cannot. At this point we come to the boundary between words and conventionalised sounds. Forms like "bop!" and whatever in Spanish certainly seem to be composed of phonemes, but violate word canonicity. Forms like English "whew!" and "tsk tsk", when pronounced, are NOT composed of phonemes, and there is not much gained by calling them "words", though you can call them what you like, as long as you recognise what's going on. P.S. Strange things happen in Spanish. Speakers from Mexico City who cannot recognise the difference between "ch" and "sh" in English, nevertheless have at least one word with [sh], "Xola", the name of a street in DF, that they pronounce quite distinctly from "chola" (with [ch]). In classical phonological theory this is a phonemic minimal pair, but is quite clear that the pronunciation "Xola" is exceptional for the speakers, as if some kind of "affect" on [ch] (actually assumed to be an Aztec word by the speakers and thus treated with respect -- affect?) From mcv at WXS.NL Sat Nov 28 03:48:35 1998 From: mcv at WXS.NL (Miguel Carrasquer Vidal) Date: Sat, 28 Nov 1998 03:48:35 GMT Subject: reversal of merger, proposal (2) In-Reply-To: Message-ID: bwald wrote: >Similar (and probably related) expressive palatalisation is used in >Spanish, e.g., in talking to babies and forming nicknames, e.g., Jo*s*e > >*ch*e, vamo*n*os > vamo*ny*o etc. The relevant (part of the) Basque >expressive rule seems to be: > > s > sh/ch (by /s/ I mean the phoneme written "z") The phoneme written (apical sibilant) behaves in exactly the same manner (sagu "mouse" ~ xagu "little mouse"). >If I understand right, the rule remains in the relevant area as an >alternation between s and sh/ch in affected lexical items, keeping it >transparent and alive. This souce of (t)sh, on a synchronic level, is >completely distinct from the historical phonemes sh and ch, which do NOT >alternate with s (and NEVER did). Such historical phonemes *x and *tx (/S/, /tS/) do not seem to exist. There are a couple of other facts that need to be taken into account. Basque does have a non-expressive palatalization rule, to the effect that in a sequence -(V)iCV- the consonant C tends to palatalize, at least if it's one of: /t/ /dit,ut/ "I have them" /s/ /giSon/ "man" /s_/ /iSil/ "quiet" /ts/ /itSal/ "shadow" /ts_/ /itSu/ "blind" /n/ /ban,o/ "but" /l/ /ol,o/ "chicken" (/d/, /b/, /p/, /g/, /k/, /r/ and /rr/ do not palatalize). The palatalization of /t/, /s/ and /ts/ seems to be optional, or less general (Larry?). This may of course also be relevant in the case of the word . Another aspect is the rule (in Gipuzkoan) that initial x- /S/ > tx- /tS/ (and /tS/ does not back, as didn't in Castillian). The fact that "soap" (< OCast. xabon(e)) becomes [xaBoi] would suggest that the change /S/- > /tS/- occurred after the change /S/ > /x/. On the other hand, I notice that this rule too seems to be subject to the effects of the affective/non-affective sense of the word. Words like "dog" (originally "little dog") and "calf" (from "calf", in disuse) are affected, but if my Gipuzkoan doesn't fail me, "torito" (as opposed to "toro") and "mouslet" (as opposed to "mouse") are not. I'm not sure what to make of that. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From manaster at UMICH.EDU Sat Nov 28 18:40:29 1998 From: manaster at UMICH.EDU (Alexis Manaster-Ramer) Date: Sat, 28 Nov 1998 13:40:29 -0500 Subject: Undoing a merger? Message-ID: On Fri, 27 Nov 1998, Larry Trask wrote: > I agree that [AMR's] interpretation of the Basque case [as involving > two kinds of [esh] is perfectly possible. However, it seems less economical than Michelena's proposal, > in requiring speakers of the Gipuzkoan dialect to construct and maintain > a contrast between two different kinds of [esh], a contrast which is not > observed or recorded in any variety of Basque. > > Moreover, such a contrast would have produced an immensely crowded > sibilant system. I have to agree, barring new evidence. > > Yes [re AMR's proposal that we are dealing with dialect interference], , but we are not here dealing with a handful of exceptional words. > With just one or two exceptions, like the second in , > *every* instance of original (expressive) remained unchanged, while > *every* instance of non-original underwent backing. Here, I think I did not make myself clear. What I am saying is that exceptional-looking expressive phonology is often easily explained as involving dialect interference. The Polish example I cited involves only a few examples in my speech, to be sure, but there are many other examples that are more productive, and if fact even in my speech a much more general e -> i rule exists, along with several other rules, as an expressive device for mocking "peasant" speech. There are plenty of expressive phonological devices in lgs and many must surely come from dialect or even foreign-lg contact. One last point: the Basque example is NOT an example of a merger that was undone, because by assumption (Michelena's and Larry's, not mine) the /s^/ which changed to /x/ and the one did not were the same to begin with. Rather, if valid, this would be an example of something else, namely, the resistance of expressive vocabulary to otherwise regular sound changes, of which many examples have been proposed over the years. AMR From mccay at redestb.es Sun Nov 1 14:39:52 1998 From: mccay at redestb.es (Alan R. King) Date: Sun, 1 Nov 1998 09:39:52 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- Yet another instance of s > r occurs dialectally in Galician (also Galego, Gallego; Romance, nearest relation Portuguese). It occurs with phonological regularity as a (substandard) geographically localized dialect feature (for isogloss details see Fernandez Rey), and is mentioned as such in several modern manuals of Galician linguistics, according to which: etymological (and standard Galician) /s/ (generally alveolar in Galician, with regional variants) is realized in the varieties in question as: [r], more exactly a "voiced simple alveolar liquid vibrant" (see Freixeiro Mato), in the following context: "in word internal position, or as a sandhi phenomenon, PRECEDING A VOICED CONSONANT, and to a lesser extent (Fernandez Rey: "exceptionally"), before the voiceless consonants /f/ and /th/ ["voiceless interdental fricative"], and even before /ch/ ["voiceless prepalatal affricate"]" (translated-paraphrased from Freixeiro Mato). Examples from Freixeiro and from Fernandez Rey [format: "standard spelling" + /broad transcription of realization, reinterpreted by me avoiding phonetic symbols not available/ + 'meaning']: "desde" /derde/ 'since, from' "as vacas" /arbakas/ 'the cows' "lesma" /lerma/ 'kind of mollusk' "as mans" /armans/ 'the hands' "escindir" /erthindir/ 'separate, split' "os zapatos" /orthapatos/ 'the shoes' "as flores" /arflores/ 'the flowers' "todos xuntos" /todor shuntos/ 'all together' Just for the record: Both sources also mention another, more "sporadic" tendency in some dialects of Galician to aspirate /s/ when "implosive" (i.e. when followed by a consonant), e.g. "desde" /dehde/ 'since, from' "disgusto" /dihhusto/ 'displeasure' (in many parts of Galicia, /g/ in most positions is regularly realised as [h]) "espantallo" /ehpantalyo/ 'scarecrow' "a escoba" /ahkoba/ 'the broom' Finally, I should no doubt mention that in Galician, /s/ is normally voiceless, but is regularly voiced when it precedes a voiced consonant, so in standard pronunciation "desde" is [dezde], "as vacas" is [azbakas], etc. Sources: Fern?ndez Rei, Francisco: Dialectolox?a da lingua galega [second edition]. Vigo: Xerais, 1991. (see page 57) Freixeiro Mato, Xos? Ram?n: Gram?tica da lingua galega. I: Fon?tica e fonolox?a. Vigo; A Nosa Terra. (see page 161) Alan Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From manaster at umich.edu Sun Nov 1 14:40:29 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Sun, 1 Nov 1998 09:40:29 EST Subject: h- in Turkic In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Sasha raises important points to which I respond briefly below. I do not strongly disagree with anything he says, but I do feel that the subject should not be closed. On Sat, 31 Oct 1998, Alexander Vovin wrote: > ----------------------------Original message---------------------------- > Alexis, > > I think your message regarding initial h- raises one important > methodological issue. Namely, can we allow a reconstruction of a segment > for a proto-language that is preserved in a single language of otherwise > big language family, and for which there is no second independent > evidence? It seems that you would answer that question in the affirmative > in this particular case, My answer is: It depends. Some syllable-final consonants in Uto-Aztecan survive only in Tubatulabal. Some IE laryngeals survive only in Hittite. > although I remember that once you yourself were > bashing (quite justifiable, in my opinion, a person X from Moscow > Nostratic school for search of IE and Nostratic accent distinctions > uniquely preserved in Bengali). But that is not the only or main reason I was bashing them! > I would hate to disagree with you on > Khaladj h-, but I think I have to. But I have not taken a strong enough position on Khalaj h- for anyone to be able to disagree with. > I would answer in the negative to the > question I posed above, although I think that some exceptions could be > allowed when a language that unikely preserves segment X, is on the top of > the branching. In all other cases it is much safer to reconstruct > something, especially something radical, like PT *h- on the basis of two > independent pieces of evidence. Khaladj is probably *close* to the root of > Turkic tree, but it does not represent primary branching, I think. It will > be dangerous enough to reconstruct PT *h- on its sole evidence (although I > think that this might eventually turn out to be true -- let us see), but > looking for the traces of something Nostratic in khaladj *only*, does not > seem to be very realistic. I did not say that this IS so, only that it is a hypothesis I would like to investigate. You should anyway be happy because you are the one who made me see that it could not just be *p-. But anyway it may simply be that Khalaj, in addition to what I call Ataturkish (haha), is the only Turkic language I am fairly comfortable with, but it sure FEELS to me like the h- is a real well-established feature here, and I would hate to have to assume that it is purely one of those mysterious that so many language make ex nihilo. Anyway, you are right that all by itself the whole thing would be somewhat improbable, but I do have some other ideas re Altaic *w-. Moreover, while no living Turkic language has anything that looks like a regular reflex of Altaic *p- or *w-, I am not yet ready to dismiss some of Doerfer's claims for older stages of the Turkic. In sum, if I choose to waste my time pursuing the possibility that the Khalaj h- is a reflex of something old, that should not worry you too much(:-). Of course, I have cagily enough NOT been doing so. Rather I keep trying to get someone else to waster THEIR time checking this out(:-). Alexis From bwald at HUMnet.UCLA.EDU Sun Nov 1 14:42:38 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sun, 1 Nov 1998 09:42:38 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Jim Rader's message on "rolled" /r/ (spelled "rr" intervocalically in Spanish orthography) in Latin American dialectology prodded my memory on some Bolivian Spanish I've heard where intervocalic "rr" indeed sounded a lot like "z" to me. In many areas of Latin America either intervocalic rolled (NOT tapped) or final "r" has a varying amount of local friction accompanying the liquid articulation producing a sibilant-like sound. When I first heard this Bolivian z-like intervocalic "rr", it reminded me of the Mexico City final "r", which also (most) often has a lot of local friction, but more "palatal"-sounding, like "zh" (even often devoiced finally to a "sh"-like sound). My general thought was that since Spanish historically lost voiced lingual fricatives (by devoicing and further shifts, no "z", "zh"), the liquid articulation was encroaching on phonetic voiced lingual fricative territory with impunity (since there were no previous voiced fricatives for it to threaten with merger). The interesting thing to me was that there doesn't seem to be any extra-segmental conditioning involved. That is, where palatalised "r" > "zh/sh" (e.g., in Polish or Czech) is heavily contextually conditioned (by an immediately preceding stop and/or following a historically palatal vowel), the development of local friction with rolled "r" among Latin American Spanish dialects does not show such coarticulation motivation, and, thus, seems to be part of a paradigmatic ("unconditioned") shift among consonants. In contrast, the change exemplified in (Parisian) French "chair(e)" to "chaise" does cause merger with previous -z-, but I am unclear on how regular this change actually was, and to what extent sociolinguistic affectation distorted its progress toward a bona fide neogrammarian sound change. Conclusion: r > z may be possible as an unconditioned change depending on aspects of the larger consonantal system in which the original "r" was positioned. Degree of aperture may vary for voiced continuants (phonetic fricatives or liquids) when it does not support a phonemic contrast. (z > r may be less dependent on position in a consonantal system, contributing to Campbell's impression that it is a more "natural" change. This relates to the general issue of directionality of degree of aperture in consonantal sound changes, "weakening", i.e., "opening", apparently being more commonly observed than "strengthening", i.e., "closing", but with enormous qualifications. The qualifications may relate more to stop/(af)fricative changes than to changes within continuants.) In any case, such considerations of consonantal system type (e.g., what are the phonetically contrastive dimensions?) have obvious implications for reconstruction and plausibility of reconstructed systems. (Oblique comment: the Chagga dialects of Northeast Bantu, Kilimanjaro area, have had varied unconditioned outcomes of *t, among them /d/, /r/, /zh/ and /R/ (i.e., velar /r/). The /zh/ variant, which varies with a palatalised /r/ in some dialects, e.g., Moshi, Vunjo, seems to show the closing process from an earlier and more widespread /r/ reflex of *t, where *t > r surrounds the /zh/ reflex,, and stems from the post-alveolar articulation of the original *t -- as preserved in most Bantu languages. The *t > r process also occurs in various other East Bantu areas as part of a lenition process that variously affects the voiceless stops *p, *t or *k -- but most often *p. In Chagga, the /R/, Machame dialect, is also clearly a further evolved form of /r/, and in turn evolves into /h/, cf. Brazilian Portuguese, in the Siha dialect. The *t > d change is most eastern, e.g., Rombo dialect, and not clearly related to the *t > r change, but is shared with the distinct language to its east, Dabida (NB often called Taita in the literature). The unconditioned change of a post-alveolar into a palatal /zh/ rather than a /z/ is paralleled independently by the Bajuni change of *t > ch by a variety of well-motivated steps which reflect the same progression as Chagga *t > zh, but without loss of the voiceless closure component. *t > r starts out as a tap, but then gets rolled and eventually, in Chagga, develops local friction on its way to zh. The central dialects demonstrate the progression. Most likely the r > zh change has nothing to do with the origin of r in *t, and is unusual, but in a less direct way the overall directions of the consonantal shifts in East Bantu may be involved, making r > zh more likely in this larger context than in general -- but cf. Mexico City final r > "zh") P.S. Trask's admonition on probability of chance resemblances reminds me of the classic probability problem: how many people do you need in a room before there is a more than chance (p > .5) probability that TWO will have the same birthday? I forgot the answer, something around 30 (cntr. 366 possible birthdays). The answer says NOTHING about which date this will be. The probability for that remains 1 in 365 and a 1/4 (p < .003). Koestler's fallacy is alive and well in popular American culture in the "strange-but-true" folklore about how many "famous" Americans were either born or died on July 4. There are many other variants of such folklore, e.g., Cabalistic-like algorithms about probability of deaths in office for US presidents and such, depending on year, their ordinal rank as presidents, the number of letters in their names, etc etc. I guess we all know some of the "other" great people who were born the same day as us. Conclusion: most people hate mathematics but love numbers. From mcv at wxs.nl Mon Nov 2 01:40:04 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Sun, 1 Nov 1998 20:40:04 EST Subject: r and s: Galician In-Reply-To: <3.0.3.32.19981031235447.006dbc18@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Yet another instance of s > r occurs dialectally in Galician (also Galego, >Gallego; Romance, nearest relation Portuguese). And in dialectal Mallorqui' too. Same rules as in Galego and Sardo (es bisbe "the bishop" > er birbe). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From dyen at hawaii.edu Mon Nov 2 21:57:12 1998 From: dyen at hawaii.edu (Isidore Dyen) Date: Mon, 2 Nov 1998 16:57:12 EST Subject: Cladistic language concepts In-Reply-To: Message-ID: ----------------------------Original message---------------------------- The social conditions are temporary because they are replaced by other conditions. The effects on the language can be either temporary or permanent. The effects are often permanent. As for the spread of socially dominant languages, like it or not, that is going on at a great pace and there is little that one can see in local resistance that is a real obstacle to it. On Thu, 15 Oct 1998, bwald wrote: > ----------------------------Original message---------------------------- > Roger Wright quotes from my last message. > > >Benji Wald says: > > > >>On the contrary, up > >>to the present, history indicates the continual long-term fragmentation of > >>languages and destruction of mutual intelligibility. > > >Yes; nicely put. "Up to the present" -- > >There's a good case for saying this may never happen again. > > I'm glad that Roger appreciated the qualification I put on what I was > saying. I deleted a further paragraph on how I invite such views as he is > proposing about a sharp discontinuity between the past and the future, > since I see little reason to suppose that such a discontinuity has come > about in the 20th c. or will in the foreseeable future, despite impressive > advances in communication technology (at least impressive to us current > beings) and increasing sharing of various kinds of literacies. Meanwhile, > the same old problems of miscommunication that have always existed (and > have occasionally been reported in the past) persist. (Would it be > surprising if "human nature" exists not only in our linguistic devices but > in how we use them?) > > I think the main thing that would interest me among Roger's proposals is > what changes in the nature of human society (as a whole or in its various > various parts) he envisages or suggests to have relatively recently arrived > which will override the steady and unrelenting effects of social localism > that in the past (presumably) have been the major cause of linguistic > fragmentation and loss of mutual intelligibility. I am skeptical, and > suspect he is underrating the long-term cumulative effects of localism, and > overrating the stability of centralised power, but I am open to hearing > interesting proposals about the "changing" relation of social change to > linguistic change. > > With regard to Isidore's latest message, I share his appreciation for > Hubey's comments, but paused at the following passage: > > >The point is, as I see it, that linguistic change > is built into the way the community interacts with its language, whereas > some aspects of linguistic change are conditioned by the social changes > that are going on in the community. The latter type of change, since it is > local and temporary I thought could be excluded from being regarded as > a 'main factor', but I suppose it gets to be a matter of defintion. > > I'm not sure I understand the intent of "local and temporary". The > "temporary" part seems to suggest that the local changes are eventually > undone, as if afterwards they seem to have never occurred (i.e., no > *lasting* harm done to mutual intelligibility). If that is not what is > meant, then they have had their effect in changing the local language AWAY > from other local varieties. This seems more than a matter of definition > (of "main factor"?) to me, but of the cumulative consequences of local and > temporary (temporally bounded?) changes. > > Of course, most changes do spread beyond the temporary local interest group > that intiates them, so their effects don't go away but continue, and > establish the typical mosaic patterns we commonly find in dialect > geography. I cannot address all fronts at once, so when I emphasised > fragmentation and loss of mutual intelligibility as the traditional and > still usual focus of historical linguistics (at least on the elementary > level), I did not complicate that traditional picture with diffusion and > convergence which takes place across languages as well as in them, due to > bilingualism, bidialectalism, register/stylistic complexity, or whatever > level of analysis is appropriate at the time. I am continually struck by > cases in which it is not clear which family or (even more commonly *in the > literature*) branch of a family some language (group) or other belongs to > because of convergence and sharing of features. Such things most > strikingly occur at the margin of isogloss bundles. Isidore (and no doubt > Roger) is certainly right that channels and interests promoting > communication ("mutual" intelligibility) *across* local groups is also a > factor in change. > > The issue remains: will languages continue to fragment and produce mutual > unintelligbility, or will the whole world eventually abandon its local > variety in favor of some "homogeneous commercial or totalitarian English" > or whatever (no doubt incomprehensible to us current beings). I think not. > From nbvint at nessie.mcc.ac.uk Mon Nov 2 17:23:32 1998 From: nbvint at nessie.mcc.ac.uk (nigel vincent) Date: Mon, 2 Nov 1998 12:23:32 EST Subject: s > r in Sardinian Message-ID: ----------------------------Original message---------------------------- Presumably the development of s > r in Sardinian is a rather different phenomenon from that in Latin which started this discussion off since the Latin environment is intervocalic while the Sardinian one is preconsonantal. Moreover, Michael Jones notes that in Logudorese-Nuorese there is an interesting complementarity, with r > s before voiceless consonants (thus 'batos kanes' '4 dogs' beside 'bator gatos' '4 cats' and s > r before voiced consonants (see his chapter on Sardinian in M. Harris & N. Vincent (eds) The Romance Languages, 1988, p. 323). For a lot more (fascinating) detail, see Michel Contini 'Etude de geographie phonetique et de phonetique instrumentale du sarde' Alessandria, Edizioni dell'Orso, 2 vols, 1987. [NB on a terminological note people in the Anglophone world these days usually call the language 'Sardinian' not 'Sard'. Posner's usage here is decidedly 'arcaizzante'.] Nigel Vincent Tel: +44-(0)161-275 3194 Department of Linguistics Fax: +44-(0)161-275 3187 University of Manchester e-mail: nigel.vincent at man.ac.uk Manchester M13 9PL http://lings.ln.man.ac.uk/Html/NBV/ UK Visit our web-page: http://lings.ln.man.ac.uk/ From Roger.Wright at liverpool.ac.uk Mon Nov 2 16:26:53 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Mon, 2 Nov 1998 11:26:53 EST Subject: s > r In-Reply-To: <19981029.173548.12590.0.TonyBreed@juno.com> Message-ID: ----------------------------Original message---------------------------- And in Spanish, including the most pompous and high-class registers in which dropping final /s/ would in general be seen as unacceptable, a word-final -/s/ before word-initial rolled /r/- just isn't pronounced at all: "los reyes" as [lor'ejes]. (This is also normal in the rare cases of what looks from the spelling like word-internal /sr/; thus "Israel" as [ira'el]). In the light of this continuing discussion, I'm now wondering if this /s/ has assimilated rather than dropping; before any other voiced consonant it would indeed assimilate, by voicing, "los malos" as [lozm'alos] (but [lozr'ejes] with [z] isn't normal). RW >I don't know if this transformation exists throughout Sard or in limited >areas. In that section, Posner talks about the general weakness of /s/, >transforming, variously, to /j/, /h/, /S/, and /x/ via /S/, as well as >dropping completely in certain cases (in French), in addition to the Sard >example. From lsa at lsadc.org Mon Nov 2 16:18:47 1998 From: lsa at lsadc.org (LSA) Date: Mon, 2 Nov 1998 11:18:47 EST Subject: October Bulletin Message-ID: ----------------------------Original message---------------------------- The October 1998 LSA Bulletin is now available on the LSA web site: www.lsadc.org From mccay at redestb.es Tue Nov 3 13:20:12 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 3 Nov 1998 08:20:12 EST Subject: s > r (Spanish) Message-ID: ----------------------------Original message---------------------------- Roger Wright wrote: >----------------------------Original message---------------------------- > > >And in Spanish, including the most pompous and high-class registers in >which dropping final /s/ would in general be seen as unacceptable, a >word-final -/s/ before word-initial rolled /r/- just isn't pronounced at >all: "los reyes" as [lor'ejes]. (This is also normal in the rare cases >of what looks from the spelling like word-internal /sr/; thus "Israel" >as [ira'el]). In the light of this continuing discussion, I'm now >wondering if this /s/ has assimilated rather than dropping; before >any other voiced consonant it would indeed assimilate, by voicing, "los >malos" as [lozm'alos] (but [lozr'ejes] with [z] isn't normal). > RW I don't have any systematic argument to back this up, but as a "second-language" Spanish speaker of over twenty years' standing (and a linguist, though not a phonetician, residing in a largely Castilian-dominated country), it had never occurred to me before to "hear" these pronunciations as anything other than assimilations or /s/ to the following /r/. And while I only speak from intuition and impressions, I am almost certain that the pronunciation of such /sr/ phoneme sequences as those mentioned by Roger Wright is something different from (and phonetically more complex than) a single trilled /r/; the phonetic representations [lor'ejes], [ira'el] etc. feel wrong (to me): something's missing in them. Or is this merely because I know how they are spelt?? Would I be saying something totally absurd if I suggested that the assimilated /s/ in these contexts is more like a "rolled [z]" than a normal Spanish /r/? Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From cwilhelm at ucla.edu Tue Nov 3 13:20:29 1998 From: cwilhelm at ucla.edu (Chris Wilhelm) Date: Tue, 3 Nov 1998 08:20:29 EST Subject: XI UCLA Indo-European Conference Message-ID: ----------------------------Original message---------------------------- I would like to announce the Eleventh UCLA Indo-Eurpoean Conference, to be held 4-5 June, 1999, on the UCLA Campus. The Abstract deadline is 1 March, 1999. Contact: Christopher Wilhelm, Coordinator 100 Dodd Hall, UCLA 405 Hilgard Ave. Los Angeles, CA 90095 Tel: (310) 825-4171 (daytimes) (310) 473-4223 (eves. & weekeends) Fax: (310) 206-1903 e-mail: iesa at ucla.edu cwilhelm at ucla.edu From larryt at cogs.susx.ac.uk Tue Nov 3 20:38:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 3 Nov 1998 15:38:09 EST Subject: Q: non-IE terms Message-ID: ----------------------------Original message---------------------------- As many of you know, I'm compiling a dictionary of historical and comparative linguistics, which I have to submit in a couple of months. At present, I have pretty good coverage of general terms from HL, comparative linguistics, philology and language change. And I also have coverage of IE terms which is about as nearly exhaustive as I think I can justify. But I don't have terms from language families other than IE. I've been avoiding those, because of worries about space. However, I've just had a meeting with my editor, and she agrees that we should go for coverage of these as well. So I'm posting a request for assistance. What I'm looking for is named "laws" and processes from families other than IE -- Semitic, Bantu, Japanese, Algonquian or whatever. That is, I'm looking for equivalents of IE terms like `Verner's Law', `First (Slavic) Palatalization' and `ablaut'. If you can suggest any such terms, it would be very helpful if you could provide as much as possible of the following information: * the term * its domain * a definition * an example or two * a *complete* reference to the first published use of the term I'm thinking of things like `Meussen's Law' in Bantu and `rendaku' in Japanese. All assistance will be gratefully received. But please note that terms must be explicitly historical in nature: non-historical terms cannot be included. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 3 19:23:52 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 3 Nov 1998 14:23:52 EST Subject: rhotacism from Ray Hickey In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Sun, 1 Nov 1998, bwald wrote: > P.S. Trask's admonition on probability of chance resemblances reminds me > of the classic probability problem: how many people do you need in a room > before there is a more than chance > (p > .5) probability that TWO will have the same birthday? I forgot the > answer, something around 30 (cntr. 366 possible birthdays). The answer > says NOTHING about which date this will be. The probability for that > remains 1 in 365 and a 1/4 (p < .003). Koestler's fallacy is alive and > well in popular American culture in the "strange-but-true" folklore about > how many "famous" Americans were either born or died on July 4. There are > many other variants of such folklore, e.g., Cabalistic-like algorithms > about probability of deaths in office for US presidents and such, depending > on year, their ordinal rank as presidents, the number of letters in their > names, etc etc. The answer is in fact 23. Put 23 arbitrary people in a room, and the probability that two of them will share a birthday date (not an actual date of birth) exceeds 50%. Put 40 people in the room, and the probability exceeds 90%. Most people won't believe this, and you can win a few bets this way. I've done so, with first-year groups of around 45 students. I've never lost, though I must lose eventually. One year, when I announced my bet, there was widespread giggling. It turned out that, unknown to me, the class contained a pair of twins. So I magnanimously agreed to count the twins as one person, and I won anyway. As Benji points out, human beings are woefully bad at estimating probabilities. Mostly, I think, we tend to interpret `random distribution' as `disperse distribution', meaning that we tend to assume that independent events have a tendency to avoid one another. They don't, or they wouldn't be random. The linguistic consequences of this failing are all too obvious. Ancient Greek for `honey' was , and Hawaiian for `honey' is . Wow! I can hear Arthur Koestler telling us that Something Deeply Significant is going on here. But, of course, neither the Greeks nor the Hawaiians had any interest in ensuring that their words for `honey' were different, nor any means of doing so. Collect enough languages, and enough words, and you're going to be drowning in such coincidences. Standard Italian `two' is a lot more similar to Malay `two' than it is to Neapolitan Italian `two'. In fact, looking at that list of number names on the Web ( http://www.tezcat.com/~markrose/numbers.shtml ) can be quite an illuminating experience. English /tu:/ and German /tsvai/ do not look to be closely related, nor do the Kashmiri dialect variants /zi/ and /do:/, nor do the Pashto variants /bu/, /lu/ and /do:v/, nor does Armenian /erku/ look like anything else IE -- yet all are cognate. Historical linguistics 15, miscellaneous resemblances 0. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Tue Nov 3 15:37:26 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Tue, 3 Nov 1998 10:37:26 EST Subject: s > r (Spanish) In-Reply-To: <3.0.3.32.19981103011411.00729d68@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >I am >almost certain that the pronunciation of such /sr/ phoneme sequences as >those mentioned by Roger Wright is something different from (and >phonetically more complex than) a single trilled /r/; the phonetic >representations [lor'ejes], [ira'el] etc. feel wrong (to me): something's ([lo'rejes] of course) >missing in them. Or is this merely because I know how they are spelt?? > >Would I be saying something totally absurd if I suggested that the >assimilated /s/ in these contexts is more like a "rolled [z]" than a normal >Spanish /r/? I'm afraid so... Now I do agree that there may be someting more than just /rr/ in /irrael/ . I'm not absolutely sure, because my native language is Castilian with some interference from Catalan (and Dutch), and if I say /irrael/ now it's the result of a conscious effort to mend my native spelling-pronunciation ways (I suppose the Castilian spoken in the Basque country may be similarly affected). But it's my impression that the initial part of [rr] in and similar words is voiceless, making it [r.rr] or [hrr], reflecting the voicelessness of the [s] that used to be there. Spanish /s/ is very very reluctant indeed to become voiced [it rather becomes [h] or [r] than [z]], and a word like for me is always [asno], never [azno]. The same phenomenon, who am I telling, is even stronger in Basque, where (voiceless!) and assimilate forwards, ezta? (=ez da). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mott at lingua.fil.ub.es Tue Nov 3 14:29:09 1998 From: mott at lingua.fil.ub.es (Brian Mott) Date: Tue, 3 Nov 1998 09:29:09 EST Subject: r and s: Galician In-Reply-To: <363dda7a.112697997@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- The change of s to r before another consonanant is widespread in southern Spain and also in Madrid, so that los dedos may be articulated as lor dedoh and buenos dias as buenor diah. On Sun, 1 Nov 1998, Miguel Carrasquer Vidal wrote: > ----------------------------Original message---------------------------- > "Alan R. King" wrote: > > >Yet another instance of s > r occurs dialectally in Galician (also Galego, > >Gallego; Romance, nearest relation Portuguese). > > And in dialectal Mallorqui' too. Same rules as in Galego and Sardo > (es bisbe "the bishop" > er birbe). > > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam > From martinez at eucmos.sim.ucm.es Wed Nov 4 16:59:31 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 4 Nov 1998 11:59:31 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- > > The change of s to r before another consonanant is widespread in southern > Spain and also in Madrid, so that los dedos may be articulated as lor > dedoh and buenos dias as buenor diah. it is nice to read it, because I never heard a change like "buenor diah" in Madrid, nor in Toledo, Galicia nor in Southern Spain. You say "widespread", but I never, *really never*, met it. You can hear following (sub)standard variants: in Madrid 1) buenos dias 2) bueno dia in the South (from Toledo toward andalucia) 3) buenoeh diah 4) buenoh diah a Galician should use the standard variation, also "buenos dias", but "bos dias" (in Galician). Have you ever heard "buenor diah" ? Where do you get such informations? I would like to read about it. j.m. From mccay at redestb.es Wed Nov 4 16:28:35 1998 From: mccay at redestb.es (Alan R. King) Date: Wed, 4 Nov 1998 11:28:35 EST Subject: s > r: Iberian miscellanea Message-ID: ----------------------------Original message---------------------------- Referring to my subject line "r and s: Galician", Brian Mott wrote: >----------------------------Original message---------------------------- >The change of s to r before another consonanant is widespread in southern >Spain and also in Madrid, so that los dedos may be articulated as lor >dedoh and buenos dias as buenor diah. True, and quite a well-known fact (to me and presumably to some others in this discussion), but the language spoken in southern Spain and also in Madrid is not Galician, but Spanish (Castilian). I thought I was broadening the discussion. This is not to deny (indeed, it's surely all part of the fun) the potential relevance of the fact that both Galician and Castilian, besides sharing a common inheritance from Latin, are undoubtedly co-members of a linguistic area, that of the Iberian Peninsula, which arguably even includes Basque, a totally unrelated language, as Larry Trask will testify. But in an "areal" context it is all the more to the point to label the particular languages (or even dialects) we're talking about. The point of my posting was that (a part of) GALICIAN has rhotacism TOO. Also in reply to my posting, Miguel Carrasquer wrote: >And in dialectal Mallorqui' too. Same rules as in Galego and Sardo >(es bisbe "the bishop" > er birbe). Agreed. What strikes me as saliently different between the Galician and Mallorcan cases is not the similar rhotacism "rule" per se, but the respective overall phonologies that are the contexts, and which contrast fairly radically. Je m'explique. In Galician words can normally only end in a vowel, /s/, /N/ (velar except under assimilation), /l/, /r/ or /z/ (voiceless interdental fricative or voiceless sibilant, depending on dialect); for word-internal syllables the situation is more complex. But my point is that the range of CC sequences is somewhat limited, especially across word boundaries. In Catalan (of which Mallorcan is a dialect), on the contrary, word-final consonants are notoriously frequent (for a Romance language), resulting in a wide range of CC sequences across word boundaries. Now within Catalan, one of the salient phonological hallmarks of Mallorcan is the strong tendency to assimilate, delete or otherwise modify the first C in just such CC sequences, to a much greater extent than in peninsular Catalan (and inevitably still greater than in non-Catalan Ibero-Romance, where most such sequences are not found - at least across word boundaries - in the first place). Thus the s > r rule in Mallorcan is merely one of a long list of such sandhi rules, a mere leaf on the tree, so to speak. In Galician the same s > r process occurs, but attracts more attention, if you will, because there's no corresponding "tree". Relevant or not to the present discussion, I'm not sure; but maybe it's better to have too much information than to risk missing a significant insight because of incomplete data. Miguel also responded to my more recent posting, this one concerning Castilian and itself in response to Roger Wright's observation about what we could call, without prejudice for the discussion and our possibly different analyses, the sr > rr issue (in "Israel", "los reyes" etc.). Firstly, Miguel, the transcription [lor'ejes] is Roger Wright's (see his message), I just copied it, non mea culpa. Secondly, my intention was to support Wright's hypothesis (that here we may have, not dropping of /s/, but assimilation, complete or partial, of /s/ to following /r/); and I even coincide with his strategy in arguing for this, namely to build on the fact that in such forms [s] is not pronounced even in those varieties (and registers) of Spanish where wholesale loss or modification of syllable- or word-final sibilants is NOT the general norm in the first place. Apart from agreeing, I attempted to up the ante by expressing my surprise that anyone could suggest otherwise, since in my (non-specialist, impressionistic and spontaneous) perception the /s/ segment in question, in the Spanish varieties in question, is not dropped, it's just realized as something other than [s]. I repeat that I can't swear by this, nor prove it, it's just a subjective feeling. If you like I can restate my feeling by saying that I intue that in pronunciation, "Israel" still contains a consonant cluster of some sort. So the /s/ has partially assimilated, in my opinion; it isn't simply lost. I agree. If the textbooks say it's lost here, I wasn't aware (I do not specialize in Castilian, it's just a language I speak every day). The Basque Country is one area (partly) in the Iberian Peninsula where, in contrast to large regions of the peninsula, we can generalize (at least in synchronic terms) that final or preconsonantal consonants in general and sibilants in particular do NOT drop, nor do they get aspirated, rhotacised, or changed into anything other than sibilants. This statements applies to both the indigenous language, Basque, and to the Castilian spoken in the part of the Basque Country that is also part of the Iberian Peninsula, but for the present discussion I am focussing on the latter. I shall not go into whatever qualifications might be in order for Basque, which could include comment on Miguel's point about ezta = ez da, because it would only lead us up a side track. However I must qualify the Spanish side of my statement in a way that may be pertinent: in the Castilian spoken here, and as far as I know in most other places where Castilian is spoken and sibilants generally maintained, and for that matter in the Basque spoken on the southern side of the Pyrenees too, _pace Carrasquer_, sibilants are generally VOICED preceding any voiced consonant. I also believe this to be true of the other Romance languages in the Peninsula. Consequently I am surprised, and indeed puzzled, by Miguel Carrasquer's statement: >Spanish /s/ is very >very reluctant indeed to become voiced [it rather becomes [h] or [r] >than [z]], and a word like for me is always [asno], never >[azno]. In this respect, what is no doubt special about Castilian (together with Basque) vis-?-vis perhaps most other Romance languages in or out of the Iberian Peninsula (exceptions are Galician and Romanian) is that these languages have no PHONEME /z/ (voiced sibilant), hence Castilian speakers do not PERCEIVE voiced preconsonantal /s/ as a "z"; it is a mere conditioned allophone which, in my experience, normal untrained native speakers are absolutely unable to perceive (just as they can't perceive the difference between fricative and plosive allophones of their voiced stop phonemes). They say, indeed, [aZno] (with an "apical" Z in some varieties, naturally) for /asno/ "donkey", just as southern Basque speakers will normally say [eZne] for /eSne/ "milk". (On the other hand I have heard [eSne] from northern Basque speakers, providing a very neat "control group" as far as the Basque data is concerned.) So, back to rhotacism. Whatever may happen in other parts of Iberia (particularly in regions where syllable-final /s/ is in general unstable), in "Basque Country Spanish" where sibilants are stable (and, I should point out to many of our readers, the native dialect of a couple of million souls), at least, we should initially expect the sequence /sr/ to be realized phonetically as [Zr], with voicing of the sibilant preceding the voiced consonant, /r/, although still phonemically /sr/ (given that there is no /z/ phoneme and the voicing contrast is non-distinctive and predictable). The big point is that we DON'T find [Zr], we find something else which some of us in this discussion have transcribed [r], others [rr], and I for my part must for now transcribe as [?r] where I cannot state the nature of the segment represented by [?], but I feel there is something there, i.e. I am explicitly rejecting the suggestion that "Israel" is (in this kind of Spanish) phonetically and phonologically [irael] - /irael/ (where [r] and /r/ are the rolled or trilled r as in Spanish "irreal" - /ireal/ - [ireal]. Actually, I'm saying two things. That "Israel" is not pronounced, here, as if it were "Irrael". And that, in the phonology of (this form of) Spanish, "s" is never dropped, not even in "Israel". In which case, it is probably rhotacised. But only when an r follows. With apologies for a lengthy message, which I hope is useful. Alan From hubeyh at montclair.edu Wed Nov 4 16:26:42 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 4 Nov 1998 11:26:42 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > As Benji points out, human beings are woefully bad at estimating > probabilities. Mostly, I think, we tend to interpret `random Most people are quite bad at estimating numbers. > distribution' as `disperse distribution', meaning that we tend to assume > that independent events have a tendency to avoid one another. They > don't, or they wouldn't be random. Tendency to avoid one another, in the limit is called "mutual exclusivity", and such events are highly dependent. > The linguistic consequences of this failing are all too obvious. > Ancient Greek for `honey' was , and Hawaiian for `honey' is > . Wow! I can hear Arthur Koestler telling us that Something > Deeply Significant is going on here. But, of course, neither the Greeks > nor the Hawaiians had any interest in ensuring that their words for > `honey' were different, nor any means of doing so. Here you unfortunately are falling for the proof by example. Even induction is not valid in logic in the physical sciences and the social sciences. It is also not impossible for this food which is savored even by bears let alone humans to be from a very old word that belongs to protoworld. What is often forgotten is that diffusion processes which give rise to the GAussian density also have the property that if we divided up the density into discrete intervals and tested the number at various levels, the highest is always at 0 which would correspond in the linguistics case to "no change", in the same way that a drunkard who takes steps at random into any direction will most often be found where he started. There's no law that says that (1) linguistic change is 100% regular and (2) that if a sound X changes to Y it cannot change back to X again. If we are arguing by example, then add this; Turkic for honey is "bal" and also means "mud" and it probably does belong to protoworld. > Collect enough languages, and enough words, and you're going to be > drowning in such coincidences. Standard Italian `two' is a lot > more similar to Malay `two' than it is to Neapolitan Italian > `two'. In fact, looking at that list of number names on the Web > ( http://www.tezcat.com/~markrose/numbers.shtml ) can be quite an Mark Rosenfelder, nice guy that he is, produced it to reproduce some of my results. He got his p=0.001 to do that. But he forgot that what he has done is (1) for independent processes and (2) the 25 matches he got for English uses 100,000 or more words from English, not the Swadesh list. I posted before that there is a list in which quantitative reasoning is not off-limits in linguistics. It is called "language" and you can join it by sending email to majordomo at csam.montclair.edu. I am sure many people on that list will be more than happy to be illuminated by more of your examples. See ya' there. > illuminating experience. English /tu:/ and German /tsvai/ do not look > to be closely related, nor do the Kashmiri dialect variants /zi/ and > /do:/, nor do the Pashto variants /bu/, /lu/ and /do:v/, nor does > Armenian /erku/ look like anything else IE -- yet all are cognate. Don't forget that Armenia sat in an area inundated with Turkic speakers and it is "eki" or "iki" in that language, and unless that word can be found in Armenian circa 1,000 BC or earlier, there cannot be any proof that it was not due to borrowing. There is report of Kashogs (Kazak?) north of the Caucasus many centuries before the common era. Let us not also forget that Sumerian for two is "imma" which is one of the 165 cognates between Sumerian and Turkic, which is "ikki". The next edition of Dr. Tuna's book promises to have even more words. In fact, it is too easy. Besides, perhaps someone should compute the probability that if N alleged cognages are found that they will all display a different sound change. After all, if N presumed cognates are found, and it is highly probable that at least N/2 are repeated, then where does the hallowed "regular sound change" of intuitive historical linguistics go? How about if 300 "matches" can be found by "accident"? What is the probability that none of them are repeated? What if 150 are repeated? Is this "regular sound change" or not? This is what happens if badly made calculations run against other calculations. someone should try calculating these probabilities. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From DISTERH at UNIVSCVM.SC.EDU Thu Nov 5 16:07:21 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Thu, 5 Nov 1998 11:07:21 EST Subject: HISTLING on holiday Message-ID: Dear Colleagues, I will be away from my computer Friday, Nov. 6 through Tuesday Nov. 10. This means that HISTLING will also be on holiday, unfortunately. If you have any announcements that you would like to have posted before I leave tommorrow (Friday, Nov. 6) at 8 a.m. EST, please send them to me before that time. Of course you may send messages to HISTLING anytime while I'm gone, but they won't be distributed until late Tuesday night. My apologies for any conconvenience that this may cause you. Dorothy Disterheft Moderator, HISTLING From sally at isp.pitt.edu Thu Nov 5 16:05:08 1998 From: sally at isp.pitt.edu (Sarah G. Thomason) Date: Thu, 5 Nov 1998 11:05:08 EST Subject: rhotacism from Ray Hickey In-Reply-To: Your message of "Wed, 04 Nov 1998 11:26:42 EST." <363FD1A9.255A5469@montclair.edu> Message-ID: ----------------------------Original message---------------------------- H.M. Hubey's claim that "there cannot be any proof" that Armenian /erku/ is not a loanword from a Turkic language is mistaken: there are at least two, and perhaps more than two, other cognate sets between Armenian and other IE languages (e.g. Ancient Greek) with initial correspondences Armenian erk- : other IE dw-. One is a word for "long", the other a word for "fear/terrible" -- sorry, I don't have the actual forms in my head or by my desk. So this is a genuine recurring correspondence, reflecting PIE *dw- > Armenian (e)rk-, the end result of a series of regular sound changes in Armenian. It's well known; in fact, it's just about everybody's favorite example of a phonetically odd regular correspondence -- the example we tend to trot out to demonstrate that genuine correspondences need not be phonetically similar at all. (There are lots of other such examples; this is just the most familiar one, the one that probably everyone who has taken an introductory historical linguistics course has heard about.) Of course, Sumerian "imma" doesn't look very similar to Turkic "ikki", either; but in the absence of other pairs of words showing -mm- in Sumerian and -kk- in Turkic, no historical linguist would accept it as a promising cognate set, especially in the absence of *systematic* evidence of cognacy elsewhere in the lexicon (systematic, i.e. with recurring correspondences, as opposed to scattered similarities of the sort Trask was warning against). I must have missed part of this thread: has Hubey given a definition of "intuitive historical linguistics"? Is it his view that all historical linguistics that isn't supported by statistics is "intuitive"? -- Sally From mcv at wxs.nl Thu Nov 5 15:59:56 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 5 Nov 1998 10:59:56 EST Subject: s > r: Iberian miscellanea In-Reply-To: <3.0.3.32.19981104120907.00713060@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Consequently I am >surprised, and indeed puzzled, by Miguel Carrasquer's statement: > >>Spanish /s/ is very >>very reluctant indeed to become voiced [it rather becomes [h] or [r] >>than [z]], and a word like for me is always [asno], never >>[azno]. > >In this respect, what is no doubt special about Castilian (together with >Basque) vis-`-vis perhaps most other Romance languages in or out of the >Iberian Peninsula (exceptions are Galician and Romanian) is that these Do you mean Roman (or C/S.Italian in general)? Romanian certainly has /z/ ( < *dj, usually). >languages have no PHONEME /z/ (voiced sibilant), hence Castilian speakers >do not PERCEIVE voiced preconsonantal /s/ as a "z"; it is a mere >conditioned allophone which, in my experience, normal untrained native >speakers are absolutely unable to perceive (just as they can't perceive the >difference between fricative and plosive allophones of their voiced stop >phonemes). They say, indeed, [aZno] (with an "apical" Z in some varieties, >naturally) for /asno/ "donkey", just as southern Basque speakers will >normally say [eZne] for /eSne/ "milk". (On the other hand I have heard >[eSne] from northern Basque speakers, providing a very neat "control group" >as far as the Basque data is concerned.) Now I am pussled/nonpluzzed. As an abnormal, trained native speaker, I certainly *can* distinguish [s] and [z] (or apical [S] and [Z])... Let's see what Navarro-Toma's (my 1926 edition, but usually valid enough) has to say about the subject: S SONORA -- Alveolar fricativa sonora: ort. , fon. [z]. [...] La sonora aparece u'nicamente, en nuestra lengua en posicio'n final de si'laba, precediendo inmediatamente a otra consonante sonora; en cualquier otra posicio'n su presencia es anormal y espora'dica. Es siempre, asimismo, una articulacio'n breve y suave; la pronunciacio'n lenta o fuerte impide su sonorizacio'n, reapareciendo en su lugar la sorda. [...] {There's my error, of course. /asno/ is indeed [asno] pronounced in isolation, as a dictionary entry or object of introspection. It becomes [azno] only in connected speech, "when I'm not watching".} En el grupo (, , ), la se sonoriza como en los casos precedentes; pero la punta de la lengua, arrastrada por la ene'rgica articulacio'n de la [r_] [r stroke-above in the original, trilled r --mcv] siguiente, abandona la forma caracteri'stica de la estrechez redondeada que la punta de la lengua forma en la , haciendo perder a e'sta su timbre sibilante y articulando propiamente, en vez de [z] ordinaria y regular, una [R] [inverted r in the original, fricative r --mcv], o sea una fricativa, $114: [IRr_a at li'tA], [lORr_E'y at s], [dO'Rr_ at ales]; otras veces, en pronunciacio'n relativamente fuerte, la se pierde por completo, aumenta'ndose, a manera de compensacio'n, las vibraciones de la [r_] siguiente. {Rhotacism indeed. I must have picked up my non-native pronunciation [hr] ([initially] voiceless trill) from the speech of S/C parts, where -s > -h in connected speech, leading to voiceless [hB], [hG], [hm], [hn], [hl] and pre-aspirated [hp], [ht], [hk] in cases like , , , , , , , , as discussed by Toma's Navarro in subsequent paragraphs.} ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Thu Nov 5 15:57:59 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 5 Nov 1998 10:57:59 EST Subject: rhotacism from Ray Hickey In-Reply-To: <363FD1A9.255A5469@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: [Armenian erku < *dwo:] >Don't forget that Armenia sat in an area inundated with Turkic speakers >and it is "eki" or "iki" in that language, and unless that word can be >found in Armenian circa 1,000 BC or earlier, there cannot be any proof >that it was not due to borrowing. There's more than sufficient proof that Arm. erk- comes from *dw-. We have "birth pains" (cf. Greek "pain", Alb. "pain", from PIE *dau-, d at u-, du:-), Arm. "long time", "long" < PIE *deu-, *dwa:- (cf. Grk. d(w)a:n "long time"), Arm. "I fear" < PIE *dwei- (Grk. d(w)e(i)os "fear"), and of course Arm. "two" < PIE *dwo: (Grk. du(w)o:) There is also a mountain of evidence that PIE *o: > u in Armenian, so the hypothesis that * became by regular sound change (*dw- > erk-, *o: > u), is vastly superior at interpreting the facts than the hypothesis that the word was borrowed from Turkic, a hypothesis that fails to explain the /r/, the /k/ (Turkic /k/ is aspirated, so should have given Arm. /k`/), and the /u/, and explains the initial /e/ (phonetically [je] in Armenian) only by invoking a much less widespread variant of Turkic *iki. >Let us not forget that Sumerian for two is "imma" The Sumerian for "2" is actually or . ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From DRC at antnov1.auckland.ac.nz Thu Nov 5 15:57:35 1998 From: DRC at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 5 Nov 1998 10:57:35 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > ----------------------------Original message------------------------ ---- > Larry Trask wrote: > > > > As Benji points out, human beings are woefully bad at estimating > > probabilities. Mostly, I think, we tend to interpret `random > > Most people are quite bad at estimating numbers. > > > distribution' as `disperse distribution', meaning that we tend to assume > > that independent events have a tendency to avoid one another. They > > don't, or they wouldn't be random. > > Tendency to avoid one another, in the limit is called "mutual > exclusivity", > and such events are highly dependent. > > > The linguistic consequences of this failing are all too obvious. > > Ancient Greek for `honey' was , and Hawaiian for `honey' is > > . Wow! I can hear Arthur Koestler telling us that Something > > Deeply Significant is going on here. But, of course, neither the Greeks > > nor the Hawaiians had any interest in ensuring that their words for > > `honey' were different, nor any means of doing so. > > > Here you unfortunately are falling for the proof by example. Even > induction > is not valid in logic in the physical sciences and the social sciences. > It is also not impossible for this food which is savored even by bears > let alone humans to be from a very old word that belongs to protoworld. Who can say what is impossible? What is very unlikely is that the Hawaiians had a word for honey 200 years ago, since there were no honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century loanword from Greek, thanks to classically-educated missionaries who needed it to translate the Bible. Ross Clark From ratcliff at fs.tufs.ac.jp Fri Nov 6 01:21:11 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Thu, 5 Nov 1998 20:21:11 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > > It is also not impossible for this food which is savored even by bears > > let alone humans to be from a very old word that belongs to > protoworld. > What is often forgotten is that diffusion processes which give rise to > > the GAussian density also have the property that if we divided up the > density into discrete intervals and tested the number at various > levels, the highest is always at 0 which would correspond in the > linguistics case to "no change", in the same way that a drunkard who > takes steps at > random into any direction will most often be found where he started. > There's no law that says that (1) linguistic change is 100% regular > and (2) that > if a sound X changes to Y it cannot change back to X again. I am quite sympathetic with Mark Hubey's ambition to expand the use of quantitative methods in HL, but it won't do to ignore the EMPIRICAL foundations of the discipline. Mathematical models have to be made to serve the evidence. The proposal that given enough time X>Y>X is plausible is based on a hypothesis that the directionality of sound change is random. It isn't. There are clearly preferred (frequently attested) and disprefered (rarely or not attested) directions for sound change. The only systematic work on this I know of is in Ch. 5 of Labov's 1994 "Principles of Language", which only deals with vowels. But every working historical linguist has probably developed his own practical database. Off the top of my head I might suggest that lenition processes (shift of a stop to a homorganic fricative or approximant) are more common than fortition processes, that shifts to a neighboring point of articulation are more common than shifts to a distant one, that among shifts of the latter type shifts from velar to palatal are more common than shifts from velar to uvular, and that shifts from dental to alveolar, or alveolar to palatal are more common than shifts from dental to labial. Without a substantial body of evidence of actual changes, and a statistical determination of likely and unlikely paths of change, we don't even know what types of correspondences to expect in long distance relationships, or what to look for when trying to discover them. The search for identities or 'similarities' may be right, but could just as well be completely wrong. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From mccay at redestb.es Fri Nov 6 12:19:10 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 6 Nov 1998 07:19:10 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: > >>In this respect, what is no doubt special about Castilian (together with >>Basque) vis-`-vis perhaps most other Romance languages in or out of the >>Iberian Peninsula (exceptions are Galician and Romanian) is that these > >Do you mean Roman (or C/S.Italian in general)? Romanian certainly >has /z/ ( < *dj, usually). Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact that besides Castilian, both Galician and Romanian are also exempt from this general Romance development. In Castilian and Galician (but NOT in Romanian, as you point out), there is no /z/ phoneme. I was confusing these two things. From hubeyh at montclair.edu Fri Nov 6 12:21:37 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:21:37 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Sarah G. Thomason wrote: > > ----------------------------Original message---------------------------- > > H.M. Hubey's claim that "there cannot be any proof" that Armenian > /erku/ is not a loanword from a Turkic language is mistaken: there > are at least two, and perhaps more than two, other cognate sets > between Armenian and other IE languages (e.g. Ancient Greek) with > initial correspondences Armenian erk- : other IE dw-. One is a > word for "long", the other a word for "fear/terrible" -- sorry, I > don't have the actual forms in my head or by my desk. There are different ways to reason; one way with certainty is logic. REasoning under uncertainty is probabilistic or fuzzy-logical. Most people who write about science are philosophers and they mostly use logic. Under this particular type of reasoning, proof is possible only in math. No proof of any laws of physics is possible let alone linguistics. Since borrowings are also regularized and since there is a tendency in language to 'regularize' including 'hypercorrection', it does not matter how many sound correspondences are found. The laws of linguistics are based on probability theory (as are laws of physics). At best we can obtain statements to the effect that such and such occurences are unlikely to be due to chance. This, of course, does not rule out 'borrowing'. > Of course, Sumerian "imma" doesn't look very similar to Turkic > "ikki", either; but in the absence of other pairs of words showing > -mm- in Sumerian and -kk- in Turkic, no historical linguist would > accept it as a promising cognate set, especially in the absence > of *systematic* evidence of cognacy elsewhere in the > lexicon (systematic, i.e. with recurring correspondences, as opposed > to scattered similarities of the sort Trask was warning against). Sure, there are 165 of them in Dr. Tuna's book. And the sound change m > k is one of them. And, of course, there are other examples of this sound change. Once again, if regular sound changes are "proof" then Tuna's work has to be treated consistently as are works are treated. > > I must have missed part of this thread: has Hubey given a > definition of "intuitive historical linguistics"? Is it his view > that all historical linguistics that isn't supported by statistics > is "intuitive"? During the last century engineers used the laws of 'hydraulics' which were derived empirically but could not be obtained from the laws of physics. Things changed during the early part of this century. Historical linguistics is circular, especially as it is based mostly on IE. IT says; 1. The set of languages, {x,y,z...} constitutes a family because of 'regular sound correspondences'. 2. 'Regular sound correspondences' indicate a 'language family'. Surely, this is as circular as; 1. I am Napoleon, and my friend here is General Marat, and he can testify that I am Napoleon. 2. I am General Marat and this is Emperor Napoleon. Any two clowns can memorize and repeat these lines. In the case of linguistics, something else is needed. I noted the basics above. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:21:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:21:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > > There is also a mountain of evidence that PIE *o: > u in Armenian, so > the hypothesis that * became by regular sound change > (*dw- > erk-, *o: > u), is vastly superior at interpreting the facts > than the hypothesis that the word was borrowed from Turkic, a > hypothesis that fails to explain the /r/, the /k/ (Turkic /k/ is > aspirated, so should have given Arm. /k`/), and the /u/, and explains > the initial /e/ (phonetically [je] in Armenian) only by invoking a > much less widespread variant of Turkic *iki. Mountain of evidence is sufficient :-) > >Let us not forget that Sumerian for two is "imma" > > The Sumerian for "2" is actually or . As Tanenbaum writes in one of his books; "if you don't like this year's standards wait until next year." :-) The word is 'imma' according to the books and journal articles used by Tuna. If next year, it is changed from 'min' to 'bobooo', then we can discuss that too. The strange thing is that there are others who are making these changes (one of them on the WWW) and as they change things they keep changing things to other cognate forms :-) There is no escape :-) -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:22:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:22:25 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > Who can say what is impossible? What is very unlikely is that the > Hawaiians had a word for honey 200 years ago, since there were no > honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century > loanword from Greek, thanks to classically-educated missionaries who > needed it to translate the Bible. Well, there is nothing like a real test :-) Probability theory can only tell us if something is likely to be due to chance, and not whether it was borrowed or descended from an earlier form of the same language. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Fri Nov 6 12:22:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 6 Nov 1998 07:22:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > I am quite sympathetic with Mark Hubey's ambition to expand the use of > quantitative methods in HL, but it won't do to ignore the EMPIRICAL > foundations of the discipline. Mathematical models have to be made to > serve the evidence. The proposal that given enough time X>Y>X is > plausible is based on a hypothesis that the directionality of sound > change is random. It isn't. There are clearly preferred (frequently This is something that used to bother me a lot and I used to think that there must be some universality. I am still not sure that there is not but I am more sure now that some of what we see is not really universal but rather due to some specific local phenomena. > attested) and disprefered (rarely or not attested) directions for sound > change. The only systematic work on this I know of is in Ch. 5 of > Labov's 1994 "Principles of Language", which only deals with vowels. This has to be from a specific time period and a specific set of languages. There is a similar problem in probability theory, that of stationarity of a signal. It's impossible to prove. No matter how many languages you look into you can only look at a specific time interval. It's hard to say if the sound changes are due to universals or to a particular combination of sounds and phonotactics. However, what I wrote above referred, in general, to any two sounds. It may be that for some specific X and some specific Y, the sound change X > Y for some specific language (i.e. specific set of phonemes and phonotactics) may be irreversible. But in general I do not see any reason to assume that no sound change is reversible given enough time. > But every working historical linguist has probably developed his own > practical database. Off the top of my head I might suggest that lenition > processes (shift of a stop to a homorganic fricative or approximant) are > more common than fortition processes, that shifts to a neighboring point > of articulation are more common than shifts to a distant one, that among > shifts of the latter type shifts from velar to palatal are more common > than shifts from velar to uvular, and that shifts from dental to > alveolar, or alveolar to palatal are more common than shifts from dental > to labial. Without a substantial body of evidence of actual changes, > and a statistical determination of likely and unlikely paths of change, > we don't even know what types of correspondences to expect in long > distance relationships, or what to look for when trying to discover > them. The search for identities or 'similarities' may be right, but > could just as well be completely wrong. Right. However, it is strange on the other hand to see those consonant clusters and lack of vowels in languages like Abaza, Georgian, or Khoisan and its clicks. The foremost question is this: if there is such a universal trend (say toward lack of cases, or toward voicing, or from stops to fricatives, or approximants) how then did the language (any language) get those stops in the first place? Or how did some language get consonant clusters at all? HOw did a language get so many cases? Etc Etc. That is the main reason I am not so sure anymore about those alleged universals, and instead try to explain them as a part of a greater and long range pattern due to interactions of specific types of languages. The most important thing to remember in these occurrences is to think about phenomena at multiple scales. You can see its workings in physics most clearly and precisely. We have lots of equations for describing things at the sub-atomic level. But most of them are useless for chemistry. At the same time, we have descriptions of phenomena at the level in which we treat the material as continuous things (i.e. we call it continuum mechanics) and in which we do not even think about the atomic composition of matter. It works. Most of what you see around you was designed and created using the equations of continuum mechanics at best. At the next level up, we treat things a lumped objects, and even ignore the continuum aspects. That works too. The same applies to linguistics changes. There are many scales at which changes occur, and if we mix up these levels we create opposing ideas. > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe > Senior Lecturer, Arabic and Linguistics, > Dept. of Linguistics and Information Science > Tokyo University of Foreign Studies > Nishigahara 4-51-21, Kita-ku > Tokyo 114 Japan -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From larryt at cogs.susx.ac.uk Fri Nov 6 12:27:44 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Fri, 6 Nov 1998 07:27:44 EST Subject: Hawaiian In-Reply-To: <31E5B5928AE@antnov1.auckland.ac.nz> Message-ID: ----------------------------Original message---------------------------- On Thu, 5 Nov 1998, Ross Clark wrote: > Who can say what is impossible? What is very unlikely is that the > Hawaiians had a word for honey 200 years ago, since there were no > honey-bees in Polynesia. The Hawaiian word is in fact a 19th-century > loanword from Greek, thanks to classically-educated missionaries who > needed it to translate the Bible. Nice one! I didn't know that, even after watching about 700 David Attenborough programs on TV. A pity, though, since it ruins one of my favorite examples of chance resemblance. But at least it becomes a particularly interesting example of borrowing. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 10 23:15:27 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:15:27 EST Subject: rhotacism from Ray Hickey In-Reply-To: <36428F6E.B04C8B45@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > But in general I do not see any reason to assume that no sound > change is reversible given enough time. I assume we are speaking of phonological change, and not of things like spelling pronunciation and the influence of prestige varieties. That being so, certain sound changes are indeed reversible. For example, PIE */t/ generally changed to theta in Germanic, but this theta has changed back to /t/ in the continental Scandinavian languages. This is `Rueckverwandlung', or `reversal'. It is not particularly common. But some sound changes are quite irreversible. Consider loss. In the ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting /h/ was later lost. I predict confidently that the Greeks will never reverse this change by re-introducing those long-gone /s/s, since no ordinary speaker has the faintest idea where the /s/s used to be, or even that they were ever there at all. And the spontaneous change of nothing to /s/ is not something that has been observed very often, if indeed ever at all. Another type of irreversible change is merger. Once accomplished, a merger cannot be reversed by purely linguistic means -- though it *can* sometimes be reversed by external means, such as by the influence of another variety which has not undergone the merger. For example, the vowels of `toe' and `tow', and of `nose' and `knows', were formerly distinct in English, but they have merged in all varieties except for some rural areas of England (and Wales?). Again, I am confident that the merger cannot now be reversed. > Right. However, it is strange on the other hand to see those > consonant clusters and lack of vowels in languages like Abaza, > Georgian, or Khoisan and its clicks. The foremost question is this: > if there is such a universal trend (say toward lack of cases, or > toward voicing, or from stops to fricatives, or approximants) how > then did the language (any language) get those stops in the first > place? Or how did some language get consonant clusters at all? HOw > did a language get so many cases? Etc Etc. Consonant clusters can and do arise from various sources. Perhaps the most frequent is the loss of intervening vowels. For example, in the English of southern England, `police' is pronounced /pli:s/, `collect' is /klekt/, `correct' is /krekt/, `collapse' sounds like `claps', and so on, producing many new instances of word-initial clusters. In many varieties of English, words like `camera', `chocolate' and `family' have lost their medial vowels, and in England the same is true of `medicine', which is /medsin/ -- all now with consonant clusters which were formerly absent. As is well known, similar things happen in French, in which `small' is pronounced /pti/. Pyrenean varieties of Basque have undergone extensive syncope, so that common `you are' (phonetically [sara]) becomes , `they are' becomes , and so on, producing numerous initial and medial clusters which were formerly absent from the language. But there are other mechanisms, such as unpacking. In some varieties of Basque, the historical palatal nasal /n~/ has been unpacked into a cluster /jn/, and in some varieties of French palatal /n~/ has been unpacked into the cluster /nj/. English has acquired some final clusters by excrescence: `vermin' to regional `varmint', `no' to `nope', `amiddes' to `amidst', `betwix' to `betwixt', and so on. And note also cases like `empty' and `thunder', whose /p/ and /d/ were formerly absent but have been inserted by epenthesis, presumably to ease the transitions between unlike sounds. The rise of cases has been much investigated. Perhaps the single most frequent source of new cases is the reduction of postpositional phrases. For example, both Hungarian and Finnish have many more cases than can be reconstructed for their Proto-Uralic ancestor, and the origins of many of these cases are well understood: they derive from the reduction of postpositional phrases. In the same way, the Basque comitative case-ending <-ekin> `with' derives from the reduction of a postpositional phrase *<-en kidean> `in the company of'. New plosives can arise by simple fortition. For example, original /w/ has been strengthened to /b/ in some American languages, and the glide /j/ (= English ) has been strengthened to a variety of fricatives, affricates and plosives in many varieties of Basque, with some of these being devoiced as well. And some western varieties of Basque have changed the sequence /ua/ first to [uwa] and then to [uba], acquiring new instances of /b/ in the process. But there are other sources of new plosives. For example, Dutch /sx/ has been dissimilated to /sk/ in Afrikaans, resulting in new instances of /k/. So, all of the questions that Mark asks are interesting ones, but they have answers which have been largely worked out in the only way possible: by looking at the evidence. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From delancey at darkwing.uoregon.edu Tue Nov 10 23:16:52 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Tue, 10 Nov 1998 18:16:52 EST Subject: rhotacism from Ray Hickey In-Reply-To: <36428B3F.CF3BFD02@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > Historical linguistics is circular, especially as it is based mostly on > IE. > IT says; > > 1. The set of languages, {x,y,z...} constitutes a family because of > 'regular > sound correspondences'. > 2. 'Regular sound correspondences' indicate a 'language family'. Could you give us a couple of examples of recognized language families that have been established on the basis of this kind of reasoning? There are a few, but it's not the standard methodology at all. Not Indo-European, for sure, which was first proposed and established on the basis of extensive correspondences in morphological paradigms. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Tue Nov 10 23:19:18 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:19:18 EST Subject: The good Dr. Tuna In-Reply-To: <36428C1B.656F3324@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H.M.Hubey wrote: > > >Let us not forget that Sumerian for two is "imma" [Miguel Carrasquer Vidal] > > The Sumerian for "2" is actually or . > As Tanenbaum writes in one of his books; "if you don't like this year's > standards wait until next year." :-) > The word is 'imma' according to the books and journal articles used by > Tuna. If next year, it is changed from 'min' to 'bobooo', then we can > discuss that too. The strange thing is that there are others who are > making these changes (one of them on the WWW) and as they change things > they keep changing things to other cognate forms :-) There is no escape > :-) I do not know Tuna's work, and I do not know if he is a specialist in Sumerian. But apparently he is not, if he has indeed had to extract his Sumerian forms from other people's publications. And relying on secondary sources for languages you are not personally acquainted with is a very dangerous practice: you often wind up merely repeating and propagating other people's errors, as well as perhaps introducing a few more of your own, through misunderstanding. Here's a brief example, which is all too typical of so many of the attempts I have seen at comparing languages carried out by investigators not personally familiar with the languages they are comparing. Please note that I am *not* trying to get at any of the people named below. I am merely pointing out the dangers of such work. In 1985, V. Chirikba published a comparison of Basque with various North Caucasian languages, mainly Abkhaz, arguing that his comparisons constituted evidence for a genetic link. Chirikba's work was cited in part by Shevoroshkin and Manaster Ramer in their article in the 1991 Lamb and Mitchell book. S and M-R conclude that there is "some plausibility to many of [C's] comparisons". By this I suppose they mean that the comparisons look good on the page. Fine, but unfortunately the Basque *data* on the page are mostly wrong. C is a specialist in Caucasian, but he doesn't know Basque, and he has relied upon various secondary sources for his Basque data. Bad move: these sources, whatever they were, were clearly anything but reliable. Moreover, C has introduced some further errors of his own, mainly in his erroneous segmentations. The 34 of C's comparisons repeated by S and M-R are numbered 78-111. On the Basque side, some of the forms are correct. However. (When I say that a Basque word is "arbitrarily segmented", I mean that C has arbitrarily extracted a portion of it which does not match and thrown it away without explanation.) [78] is cited wrongly, badly wrongly. [80] does not exist. [81] is hopelessly misglossed. [84] is slightly misglossed. [89] is a bimorphemic form containing a case-suffix but wrongly cited as monomorphemic, even though the case-suffix itself constitutes item [83]. [90] consists of another bimorphemic (case-suffixed) form wrongly cited as monomorphemic, plus a second truly monomorphemic item unrelated to the first (and wrongly glossed). [91] does not exist, but obviously represents a garbled attempt at citing a stem whose meaning is utterly different from what is suggested. [92] does not exist. [93] does not exist. [95] is wrongly glossed and assigned to the wrong part of speech. [96] is wrongly cited. [98] is cited only in a secondary sense. [99] does not exist, and is moreover arbitrarily segmented. [100] is wrongly glossed. [101] is cited only in a secondary regional form, and is arbitrarily segmented. [102] is arbitrarily (and impossibly) segmented. [104] is a transparent compound wrongly presented as monomorphemic. [105] is cited only in a regional secondary variant, and it is both wrongly glossed and assigned to the wrong part of speech. [108] does not exist. [109] is cited only in a regional variant lacking the initial /g/ found elsewhere, and is glossed only in a secondary sense. [111] is cited as monomorphemic, even though the second half of it is precisely the item cited separately as item [80]. So, of the 34 Basque items compared, 21 are erroneous or non-existent. Of the remaining 13, three are single segments, seven are monosyllables in which only a single segment matches anything in Caucasian, and one more contains a sequence of three segments unmatched in the Abkhaz comparandum. The case therefore rests almost entirely upon the erroneous Basque data, while the real Basque data lend no support. It is very, very dangerous to try to do comparisons on languages you do not have a specialist knowledge of. You have to assume that everybody who has intervened between the native speakers and you is virtually infallible. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From l.campbell at ling.canterbury.ac.nz Tue Nov 10 23:20:01 1998 From: l.campbell at ling.canterbury.ac.nz (Lyle Campbell) Date: Tue, 10 Nov 1998 18:20:01 EST Subject: Hawaiian meli Message-ID: ----------------------------Original message---------------------------- Ross Clark's example of Hawaiian -meli- 'honey' as a loanword from Greek via classically-educated missionaries translating the Bible is a very nice one, as Larry Trask pooints out. Larry notes that it ruins one of his favorite examples of chance resemblance, but becomes a particularly interesting example of borrowing. I suppose you could sort of raise the ante on interestaing cases of chance resemblance and borrowing by throwing Maori -mieri- 'honey' into the mix. A comparison of Hawaiian -meli- and Maori -mieri- (bother the -i- vowel difference) might seem to suggest a Polynesian cognate set (throw in Niuean -meli- 'heney' as well, also apparently from Greek), but the Maori word is actually a French loanword (from French -miel- 'honey'), courtesy apparently of early French Catholic missionizing activities in New Zealand, which soon faded in the country. (There are not many French loans in Maori, but a favorite is -wi:wi:- 'French' < French -oui- 'yes'). As Larry Traks points out, this -meli- / -mieri- false cognate is no longer a case of sheer accidental similarity, in that both are from Indo-European languages, but we still have accident to thank for it in a way, in that by sheer happenstance Hawiian ended up with a Greek form and Maori with a French one (which happen to be related languages), not something that would have been expected. Lyle Campbell From hubeyh at Montclair.edu Tue Nov 10 23:21:07 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Tue, 10 Nov 1998 18:21:07 EST Subject: The good Dr. Tuna Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > I do not know Tuna's work, and I do not know if he is a specialist in > Sumerian. But apparently he is not, if he has indeed had to extract his > 1. You mean you would not believe me if I said 2+2=4 because I am not a mathematician? 2. There's a thing called "Argument from Authority". It is one of the classic fallacies of logic. 3. What exactly is there to not being a Sumerian except that you spent N years more on it than someone else? Either the words published in the "scholarly journals" are close to the sound and semantic shape or are not? I have seen people on the WWW do similar things. I noted one cognate and sent email. he changed it. Now there's another one that I noted, which is a cognate, and if I send email to him, he will probably change that too :-) 4. On the ANE list, scholarly discussion is defined as discussion of ideas published in scholarly journals. Scholarly journals are journals in which scholars publish. Scholars are those who publish their ideas in scholarly journals. What is the point of all this? Either what is written uses commonly accepted reasoning or it does not. 5. The simple facts are that there are simple rules in linguistics and it does not take longer than a few minutes to get the hang of the basic idea, but the justification is often missing, and the practice is usually pretty muddled. So far I have found no book on historical linguistics in which there is a clear algorithm for reconstructing protoforms. And i have asked many people about this. They don't know either. Sometimes they can justify it easily sometimes it is a guess and sometimes it is no more than a concensus. You can't get scientific truth by a vote unless it can be clearly explained and backed up by scientific methodology. Some of the worst linguists are those who merely repeat what they have managed to memorize. It's better to become a gardener than to get a PhD in that manner. Let us also recall some of the things done by amateurs. Gauss, one of the greatest mathematicians of all time, had to publish his own book on Arithmetic and Group theory. You can find similar stories in every field. Either you should get your own copy of Tuna's book or get it from a library and xerox it and then make up your mind instead of using the IBM FUD factor (fear, uncertainty, doubt) in the minds of the readers. > Sumerian forms from other people's publications. And relying on > secondary sources for languages you are not personally acquainted with > is a very dangerous practice: you often wind up merely repeating and > propagating other people's errors, as well as perhaps introducing a few > more of your own, through misunderstanding. That is pretty useless. Nobody can learn every language and nobody can be a mathematician, physicist, engineer, computer scientists, chemist, economist, etc. That is not how it works at all. In fact, there is a thing called "specialization of labor" and has been around for a long time. In fact, it is used by linguists all the time, since they defer to specialists all the time, even to the point of committing the "argument from authority" fallacy. Where are the books on Sumerian? Where are all the books on Hittite? I have a batch of them. It took be 6 months to collect some and xerox the others. Does this mean that I must not trust that the OI has done poor work and that I should not rely "on others". What if I start to look at the original text? What exactly is there that is going to take 12 years to learn? Some symbols arranged in some order with some presumed meanings which can be gleaned from multilingual transcriptions. What is there to gain? Sure, maybe I can spot some things, or maybe I can fit another set of words to the original. But something that has gone on for a long time and which creates such long sentences must be, in the main, correct. One does not need to know much more than simple probability theory to know that. And that is exactly what people use to reach their conclusions, even if they use probability theory only intuitively. > > > Here's a brief example, which is all too typical of so many of the > attempts I have seen at comparing languages carried out by investigators > not personally familiar with the languages they are comparing. Please > note that I am *not* trying to get at any of the people named below. > I am merely pointing out the dangers of such work. > Maybe you are a poor linguist and not Manaster-Ramer :-) That is why so many people fight all the time in linguistics. And there is a way to make sure that things can be done so that they are comparable. That way obviously is to present the results in a way in which all scientists do. That is why they use math, and that is why eventually it will be used in linguistics. Finally, once again, you are arguing by pseudo-induction. Let me ask if this argument works: Once, I broke my leg going up an escalator. So don't go up escalators. This is not even induction which itself does not work for science. Here is a case of 'bad induction": The farmer feeds the goose 35 days in a row. So he will feed it on the 36th day. No. He might decide to have roast duck instead. Here is another bad case of induction: Every saturday for the past 10 weeks it rained. So it will rain on the 11th. Wrong. So why do you think that I would take your argument seriously? Surely, there are bad linguists as well as bad engineers. And there is bad linguistics as well as good. So what? How do you propose to fix this problem? Let me guess. You want to have a stamp by the International Linguistics Association and you want to stamp every book that even touches upon linguistics as bad and good. Is that a solution? Well, the arguments often made by linguists, (like a few already made here) are essentially informal versions of this. And they stink as badly as the formal version would stink. > > In 1985, V. Chirikba published a comparison of Basque with various North > Caucasian languages, mainly Abkhaz, arguing that his comparisons > constituted evidence for a genetic link. Chirikba's work was cited in > part by Shevoroshkin and Manaster Ramer in their article in the 1991 > Lamb and Mitchell book. S and M-R conclude that there is "some > plausibility to many of [C's] comparisons". By this I suppose they mean > that the comparisons look good on the page. Fine, but unfortunately the > Basque *data* on the page are mostly wrong. C is a specialist in > Caucasian, but he doesn't know Basque, and he has relied upon various > secondary sources for his Basque data. Bad move: these sources, > whatever they were, were clearly anything but reliable. Moreover, C has > introduced some further errors of his own, mainly in his erroneous > segmentations. > > The 34 of C's comparisons repeated by S and M-R are numbered 78-111. On > the Basque side, some of the forms are correct. However. > > (When I say that a Basque word is "arbitrarily segmented", I mean that C > has arbitrarily extracted a portion of it which does not match and > thrown it away without explanation.) > > [78] is cited wrongly, badly wrongly. > > [80] does not exist. > > [81] is hopelessly misglossed. > > [84] is slightly misglossed. > > [89] is a bimorphemic form containing a case-suffix but wrongly cited as > monomorphemic, even though the case-suffix itself constitutes item [83]. > > [90] consists of another bimorphemic (case-suffixed) form wrongly cited > as monomorphemic, plus a second truly monomorphemic item unrelated to > the first (and wrongly glossed). > > [91] does not exist, but obviously represents a garbled attempt at > citing a stem whose meaning is utterly different from what is suggested. > > [92] does not exist. > > [93] does not exist. > > [95] is wrongly glossed and assigned to the wrong part of speech. > > [96] is wrongly cited. > > [98] is cited only in a secondary sense. > > [99] does not exist, and is moreover arbitrarily segmented. > > [100] is wrongly glossed. > > [101] is cited only in a secondary regional form, and is arbitrarily > segmented. > > [102] is arbitrarily (and impossibly) segmented. > > [104] is a transparent compound wrongly presented as monomorphemic. > > [105] is cited only in a regional secondary variant, and it is both > wrongly glossed and assigned to the wrong part of speech. > > [108] does not exist. > > [109] is cited only in a regional variant lacking the initial /g/ found > elsewhere, and is glossed only in a secondary sense. > > [111] is cited as monomorphemic, even though the second half of it is > precisely the item cited separately as item [80]. > > So, of the 34 Basque items compared, 21 are erroneous or non-existent. > Of the remaining 13, three are single segments, seven are monosyllables > in which only a single segment matches anything in Caucasian, and one > more contains a sequence of three segments unmatched in the Abkhaz > comparandum. The case therefore rests almost entirely upon the > erroneous Basque data, while the real Basque data lend no support. > > It is very, very dangerous to try to do comparisons on languages you do > not have a specialist knowledge of. You have to assume that everybody > who has intervened between the native speakers and you is virtually > infallible. > > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From hubeyh at montclair.edu Tue Nov 10 23:21:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:21:25 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > That being so, certain sound changes are indeed reversible. For > example, PIE */t/ generally changed to theta in Germanic, but this theta > has changed back to /t/ in the continental Scandinavian languages. This > is `Rueckverwandlung', or `reversal'. It is not particularly common. So what does that mean? If there are only 20 consonants and 10 vowels there are (30)(30)-30 = 870 possible sound changes and out of this only 30 can be changing back to the original. > But some sound changes are quite irreversible. Consider loss. Why irreversible? don't languages sometimes add sounds? > Another type of irreversible change is merger. Once accomplished, a > merger cannot be reversed by purely linguistic means -- though it *can* What happened to adding of sounds? > But there are other mechanisms, such as unpacking. In some varieties of > Basque, the historical palatal nasal /n~/ has been unpacked into a > cluster /jn/, and in some varieties of French palatal /n~/ has been > unpacked into the cluster /nj/. So sounds can be added after all. > English has acquired some final clusters by excrescence: `vermin' to > regional `varmint', `no' to `nope', `amiddes' to `amidst', `betwix' to > `betwixt', and so on. And note also cases like `empty' and `thunder', > whose /p/ and /d/ were formerly absent but have been inserted by > epenthesis, presumably to ease the transitions between unlike sounds. Ditto. > So, all of the questions that Mark asks are interesting ones, but they > have answers which have been largely worked out in the only way > possible: by looking at the evidence. CAn you look at evidence that you don't have? This is only what we can see. So then what is wrong with "if it can happen once, it can happen twice" or "if it happened in the past, it can happen in the future". That means that unless there are many many examples collected over many many years of changes that "never" seem to occur, then it is still "all possibly go". -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:21:52 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:21:52 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Scott DeLancey wrote: > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > > > Historical linguistics is circular, especially as it is based mostly on > > IE. > > IT says; > > > > 1. The set of languages, {x,y,z...} constitutes a family because of > > 'regular > > sound correspondences'. > > 2. 'Regular sound correspondences' indicate a 'language family'. > > Could you give us a couple of examples of recognized language families > that have been established on the basis of this kind of reasoning? > There are a few, but it's not the standard methodology at all. > Not Indo-European, for sure, which was first proposed and established > on the basis of extensive correspondences in morphological paradigms. Aren't morphological paradigms also part of 'regular sound change"? > Scott DeLancey > Department of Linguistics > University of Oregon > Eugene, OR 97403, USA > > delancey at darkwing.uoregon.edu > http://www.uoregon.edu/~delancey/prohp.html -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:23:19 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:23:19 EST Subject: [Fwd: Turkic-Sumerian Connection Message-ID: Here are some of the words, some of them with additions by me to the original post by someone else. -------------------------------------------------- Here are some more additions. In a couple more iterations we can have more words for everyone to look at. On Mon, 3 Nov 1997, TIMUR KOCAOGLU wrote: >( Sm. = Sumerian, Tk. = Turkic, % = the loss of the consonant) > >A. Sound equations which CANNOT be recognized easily: > >a. Initial consonants: D, g, m, n, S, Sh, % > >I. Sm. .D ~ Tk. .y, % > >1. Sm. dar : 'spalten, zerschneiden, zerstoeren' (D.71; MSL, III, 100) > Tk. yar-: 'yarmak' [to break in the middle] (DLT, I, 399; KBI, 523) > >2. Sm. dib : 'Band' (D. 83) > Tk. yip : 'ip' [string, rope] (KBI, 546) Note that in some Kipchak dialects these words would be /car/ and /cip/. Furthermore, in d-Bolgharic they would be /dar/ and /dip/. Here are a few more: Sm. dir : zerspalten, zerstoren, umwerfen (D.86) Tk. yir : yirmek (DLT, III, 58) Sm. dirig : 'to be excessive, to be too much, too many (Grd.341) 'ubervoll sein, voll sein (D.87), diri 'ubergrosss'. Tk. irig : 'sert, kaba, hasin, gayretli' (KBI,199) iri 'kaba, sert', 'large, huge, voluminous, big' (Rd. 546) This is the word that showed up in the name Eridanus (i.e. iri) by Herodotus. And again, Doerfer has reconstructed proto-Turkic as *d > *d' > c > y > 0. In some cases the initial y >0 in Turkic languages such as cip/ip, cilan/yilan/ilan, cirt/yirt, cer/yer. These would have originally been dip, dilan, dirt, der, etc according to Doerfer, and Doerfer did this decades before Tuna's book came out, and Doerfer is a big name in Altaic/Turkic studies. Furthermore the old name for the Yayik River (Volga/Don?) shows up in Greek as daichs/daix exactly as it should have been. And it shows up circa 200 BC. Other d-Bolgharic words have been deciphered on Runic writings such as daga, dog, der, in addition to one that was already known, dilom (yilan=snake). Sm dirra : 'Hulfe' (D.87) Tk. yarI " (yardim) (help) Sm. tir : 'country' (MSL,III.87) Tk. yir : 'yer, toprak, yeryuzu' (KBI, 546)(land, earth) Sm. tu(5): 'waschen, baden, libieren' (D.206) Tk. yu(w): yikamak (to wash) > >II. Sm. .g ~ Tk. %, .y > >1. Sm. gamar : 'wuctig sein' (D.41) > Tk. agIr : 'aghIr' [heavy] (DLT, I, 52) > >2. Sm. garim : 'Fluss-Aue' (MSL, III, 109) > Tk. arIq : 'Irmak' [river, canal] Sm. gaz : 'to crush' (Grd.356), gaza (zerbrechen (MSL,III.143) Tk. ez : 'to crush, to pound', ezme 'crushed, pounded' Sm geme: 'Magd' (MSL,III.1250 Tk. eke : 'buyuk kiz kardes' (DLT,I,685) (older sister) Sm. gishig : 'Tur' (D.130), 'door' (EHG,436) Tk. eshik : 'kapi' (KBI,206) (door) >III. Sm. .m ~ Tk. .K > >1. Sm. mal : (Emesal) 'to stay' (for gal), (Grd. 384) > Tk. kal-: 'kalmak' [to stay] (DLT, I, 41; KNI, 215) > >2. Sm. marun : 'Ameise' (D. 160) > Tk. karIncha: 'karInca' [ant] (DLT, I, 501; III, 375) Sm. mir : 'anger' (MSL,IV,35) Tk. kiz : 'to be angry, cross..." SM. mu : 'name, fame' (Grd.388, MSL,V,220) Tk. ku : "un, shan' (DLT,III,212) (fame) >IV. Sm. .n ~ Tk. .y > >1. Sm. nad(9) : 'sich niederlagen, beschlafen; sich lagern' (D. 168; MSL, III,152) > Tk. yad- : 'yaymak, doshemek, sermek' [to spread, to lay down] (DLT, I, 15) The name "yayik", or "daichs" comes from this root. Sm. nigin : 'Summe' (D,171,MSL,III,111) Tk. yigin : 'yigin,kume, yigilmis' This word shows up closer to original form in Balkar as nigish. Sm. nunuz : 'bead' (EHG, 21; Falkenstein, 29) Tk. yinchu: 'inci,cariye (DLT,I,273) Sm. nurum : 'Licht' (D.170) Tk. yaruk : 'isik, aydinklik, parlak' (DLT,I,46)(light) >V. Sm. .S ~ Tk. .y, % > >1. Sm. sar : 'schreiben' (MSL, III, 113), 'to write' (Grd. 403) ~ shar id. > Tk. yaz-: 'shashmak, yanIlmak, chozmek, yazmak' [to write with > other meanings] (DLT, I, 192; II, 20, III, 59) Sm. sig : 'prime, good' (Salonen,22) Tk. yig : 'yeg,iyi, daha iyi' (good) Sm. silig : (II2) 'Hand' (D.182) Tk. elig : 'el' (DLT, I, 72; 4KBI,145) (hand) Sm. sheg : 'rain' (Grd. 412) TK. yag : 'yagmak' (to rain, to precipitate) Sm. shir : 'singen und spielen' (MSL,III,150) Tk. yir : 'kosma, turku, hava,.." This word is still 'jir' in Balkar, Tatar, and 'ir' in kyrgyz. 'Shiir' is poem Turkish and related words mean poem in lots of Turk* languages. Sm. zag : 'Grenze' (MSL,III,85) 'border' (MSL,V,70) 'shoulder, outer edge, boundary, border' (Grd,432) Tk. yaka : 'taraf,yan,civar' (EUSz,280) Tuna missed "chek" which still means 'border' today in KB. Szekely (Chek-eli) in Hungary probably means "borderlands" and Czechs were probably named that way the same way Ukrain, and Krajina got their names. BTW, in Kipchak languages, 'yaka' would be more like 'jaga'. >VI. Sm. .Sh ~ Tk. .ch > >1. Sm. sag : 'small child' (MSL, III, 78) > Tk. chagha: 'yeni dogmush' [new born infant] (YTsz. 48) Sm. sag : 'schlagen' (D.175) Tk. chak : 'chakmak, vurmak' (EUSz,58) The word for nail, "chivi" in Turkish, and "chuy" in KB seem related. The word was probably more like "chuk" for nail (although it now means something else in slang). Sm. zibin : 'Insect' (D.120) Tk. chibin : 'sinek' ... (i.e. fly) >VIII. Sm. .u ~ Tk. .kV/a > >1. Sm. ubur : 'weibliche Brust' (MSL, III, 145; D. 102), 'teats' (Grd. 426; Falkenstein, 26) > Tk. koguz: 'goeghuez' [breast] (KBI, 274; EUSz, 114: DLT, I, 366) Sm. ud: 'day, time (in general' (Grd.425) ud 'day' (MLSb) the consonants after the initial vowel of the word: d, d, m, r, sh > >I. Sm. d/ ~ Tk. d/ > >1. Sm. adakur : 'ein opfergefass fuer Getraenke' (Sm. Lw.)(Akk. Hwb. 9) > Tk. adak : 'ichki kadehi' [wine glass] )Hs. S. 559; Nh. F. 370-8; Muh. 7) Sm. gid : 'l. entfernen' (D.60) Tk. id : 'salmak, gondermek.." Although in most Turk* languages this seems to be something like "jiber" in KB we still find this in the form "iy". Sm. ud : 'Zeit' (D.104) Tk. o"d : 'zaman, vakit..'(DLT,I,44...) (time) >II. Sm. d/,. ~ Tk. n/, . > >1. Sm. dugud : 'schwer' (MSL, III, 141) > Tk. yogun : 'kalin, yoghun' [thick, heavy] (KBI, 549; DLT, III, 29) Sm. kid(2): 'Sonne' (D.149) (il-SHAMASH..) TK. kun : 'gun, gunes,..' (sun, day) SM. mud : 'blood' (Grd. 389) Tk. kan : 'kan..' (blood) Sm. udu : 'sheep' (Grd. 427), udu, 'Schaf' (MSL,III,111) Tk. kong : 'koyun, (sheep) Sm. shid: 'number, voting board' (MSL,V,15); shiti 'Rechnung, 'Zahl' (MSL,III,144) Tk. san : 'sayi, sayma...' (to count) >III. Sm. VmV ~ Tk. VKV > >1. Sm. amash : 'Schafhuerde' (MSL, III, 145), 'Umfriedung, Stall' (D. 13) > Tk. agIl : 'aghIl, koyun yataghi' [sheepfold] (DLT, I, 65, 73) Sm. imma : 'two' (Emesal, (Falkenstein) Tk. ikki : SM. umush: 'discernment' (Grd, 428), 'Verstand' (D.108) Tk. ukush : 'anlayish; (understanding) >IV. Sm. r/ ~ Tk. z/ > >1. Sm. bur : 'to spread abroad, to disperse of (a thing)' (Grd. 336; MSL, III,140, 170) > 'undo' (especially a spell),' 'to make a hole' (MSL, III, 67), buru 'harvest (moun)" (Grd. 336) > Tk. buz-: 'bozmak, yIkmak' [to demolish, to ruin] (DLT, III, 8) Sm. gur(5) : 'zerbrechen, zerschneiden,abtrennen' (D.55) Tk. u"z: 'to break'... Sm. har :'dig,dig quickly' (Prince,176) Tk. kaz : dig Sm. mir: 'anger' (Emesal) (MSL,IV<35) Tk. kiz: (to anger) Sm. sur : 'to squeeze, to press out (oil, juice)' (Grd. 408) Tk. suz : 'suzmek: (to filter, to squeeze out) >V. Sm. sh/ ~ Tk. l/ > >1. Sm. ashsha : 'six' (Falkenstein, 41) > Tk. altI : 'altI' [six] (EUSz. 13) Sm. tush: 'seat, to sit' (MSL,III,58), 'to dwell (in a place) Tk. ol : 'bulunmak, kalmak..." (to be) Sm ush : 'dead, to die' (Grd.431) Tk. o"l : (die) DLT,IT,38) This word shows up in both versions in Karachay i.e. aush (to die). >VI. c. The word endings in ae, g, m, Vr/z > >I. Sm. ae. ~ Tk. An. > >1. Sm. mae : 'I' (Grd. 386) > Tk. men : 'ben' [I] (KBI, 309, DLT, I, 20) SM. zae : 'you' (sing) Tk. sen : >II. Sm. g. ~ Tk. ng. > >1. Sm. asha(g) : 'field' (Grd. 326) > Tk. alang : 'alan, duz ve achIk yer' [field, open space] (DLT, I, 135) Sm. dag : 'daybreak, morning, dawn' (D.43) Tk. tang : (daybreak, morning) Sm. kalag : 'to be strong, vigorous, have power' (Grd.349) Tk. kalIng: 'kalabalik, cok suru, kalin, kesif' >III. Sm. m. ~ Tk. K. > >1. Sm. alim : 'Steppentier, Widder' (D. 13) > Tk. elik : 'geyik' [deer] (ETY, Ir. 97, II, 90; ....) Sm. shurum : 'a cattle stable' (D.201) Tk. surug : 'suru' (herd) >IV. Sm. CVr/z. ~ Tk. Cr/chV. > >1. Sm. dingir : 'Gott' (D. 84), 'god' (Grd. 341) > Tk. tengri : 'TanrI' [god] (DLT, I, 53, 68) > 'goek, sema' [sky] (DLT, III, 377) Sm. dubur : 'Hode' (D.78) Tk. yumru : 'top gibi yuvarlak' (Mn.Gz.78, v5)..(round) >B. Sound equations which CAN be recognized easily: > >1. Sm. agar : 'lead (metal)' (EHG. 34, 58) > Tk. agIr : 'aghIr' [heavy] (DLT, I, 52) > >2. Sm. azgu : 'neck-stock (for use with animals)' (Grd. 331) > Tk. asgu : from as- 'asmak' [to hang] 'asgu' [hanger] (Dsz. 342) > 'asgI', 'asku', 'askI' > >3. Sm. bulug : 'Grenze, Grenzegebiet' (D. 31) > Tk. bulung: 'koshe, bucak, zaviye' [corner] (DLT, II, 371) > >4. Sm. di : 'to speak' (Grd. 342) > Tk. ti-: 'demek' [to say, to speak] (DLT, III, 231) Sm Tk Meaning esh es blow gim kipi/kimi like,as hum kom Lager, stall iduga yidig perfume kad kada knupfen kash kach speed, run, ki kI-L make, do ku ko werfen,lagen,grunden, koymak kur koru guard, watch,protect men men ich, I nammu neme how much, what sum sun geben, give te(ga) teg touch, attain tin tIn Leben, life, breath u u schlaf, uyku (sleep) umush yumush hizmet, vazife,werk ush us verstand,akil zag sag right side >Prof. Tuna compares about 165 Sumerian words with 149 Turkic, 7 >Mongolian, and 1 Akkadian words. > >Prof. Tuna discussed and argued with his thesis and examples with the >following scholars: > >1. 1971 in Harvard University (Boston): Erica Reiner, Guterbock, >Herbert Paper, Omeljan Pritsak, Denis Sinor. > >3. In 1974: with George Cardona & Henry Hoenigswald [bothIndo-Europeanists], >John Faught [American Native Languages], Ake Sjoberg [Sumerologist], >Earl Leichty [Akkadian], Omeljan Pritsak [Turcologist, Slavist]. > >In conclussion, Prof. Tuna argues that there is an historical link >between the Sumerian and the Turkic languages in the past, and whether >they are related or not is not relevant at this point. But, with >these 168 words, he thinks he has established a sound ground for this >"historical link." > >Any comments? > >Timur Now, the serious question of historical linguistics cannot be avoided. Everyone talks about 'regular sound change'. But not too many can answer the question of 'how many'. This data of Dr. Tuna (who is a linguist with a PhD from University of Pennsylvania) cannot be evaluated correctly until and unless we have an idea of how many cognates are required to establish "genetic relationship". Something this important, something that strikes deep into the very core of historical linguistics methodology should have shaken any normal scholar into action. PS. The Sumerian Dictionary from the University of Pennsylvania is by Ake Sjoberg. PPS. I can easily find 30-50% cognates between Turkic and the Sumerian list published by Halloran on the WWW, without even trying too hard. PPPS. If anyone wants more, I have a 20 page paper titled "Story of Life & Death, and Love & War" that treats related topics in historical linguistics via examples of sound changes in Turkic languages. Finally, there have been a few people who have done some mathematical calculations of how many words can be found to be cognates accidentally or how many pairs of cognates are needed to establish geneticity. The numbers are quite small, say 3-7 pairs (by people like Cowan, Bender, Greenberg). I have also made calculations of this type, much more thorough than these people and even tested it via a computer simulation. They can be found in my book, "Mathematical and Computational Linguistics", Mir Domu Tvoemu, Moscow, Russia or in electronic form from http://www.1stbooks.com. The chapter that deals with these computations can be found on my website for free. Enjoy your readings! Regards, Mark Hubey http://www.csam.montclair.edu/~hubey -------------- next part -------------- An embedded message was scrubbed... From: unknown sender Subject: no subject Date: no date Size: 11331 URL: From larryt at cogs.susx.ac.uk Tue Nov 10 23:23:43 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:23:43 EST Subject: The good Dr. Tuna In-Reply-To: <364381BA.3D0EC98D@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 6 Nov 1998, H. M. Hubey wrote: [on Tuna's knowledge of Sumerian] > 1. You mean you would not believe me if I said 2+2=4 because I am > not a mathematician? Nope. I mean only that I am unwilling to believe that something is true of Sumerian merely because Tuna has asserted it -- especially if he is not a specialist in Sumerian. > 2. There's a thing called "Argument from Authority". It is one of > the classic fallacies of logic. But I am not arguing from authority. I am merely pointing out that facts are facts, independently of what anybody, authoritative or not, thinks about them. But an authority, by definition, is far more likely to know what the facts are. In Dr. Tuna's case, his cited Sumerian word for `two' is either right or wrong. If it is not in accord with the facts, then it is wrong, no matter *who* he is quoting. > 3. What exactly is there to not being a Sumerian except that you > spent N years more on it than someone else? `Sumerianist', I presume. A Sumerianist is somebody with a detailed knowledge of the Sumerian language. Reading somebody else's book on Sumerian does not make you a Sumerianist, just as reading a book on how to play baseball does not make you a baseball player. > 5. The simple facts are that there are simple rules in linguistics and > it does not take longer than a few minutes to get the hang of the basic > idea, Really? You mean all of us have wasted our time in spending years learning our trade, when all we needed to do was to sign up for Doctor Hubey's Patented Ten-Minute Education in Linguistics? Gosh. OK. Here's a problem from my field. The four major regional variants of the Basque word for `ear' are as follows: So: what's the proto-form? And what "simple rule" should be invoked to discover it? > So far I have found no book on historical linguistics in which there > is a clear algorithm for reconstructing protoforms. That's because there isn't one. Reconstructing proto-forms is not the sort of thing can be reduced to an algorithm. Reconstruction requires a profound knowledge of the languages involved and a good understanding of language change -- at least. The process is not determinate, and it cannot be mechanized. Historical linguists can no more use algorithms to reconstruct languages than historians, archeologists and paleoanthropologists can invoke algorithms to reconstruct the bits of the past they're interested in. We all just have to dig out the available evidence, sweat over it, and then try to figure out what happened. > Some of the worst linguists are those who merely repeat what they > have managed to memorize. It's better to become a gardener than to > get a PhD in that manner. Nobody in the English-speaking world gets a PhD in linguistics by memorizing things. But it *is* a good idea to learn the facts, since external examiners have a tiresome habit of expecting the candidate to know the facts. > Let us also recall some of the things done by amateurs. Gauss, one of the > greatest mathematicians of all time, had to publish his own book on > Arithmetic and Group theory. You can find similar stories in every field. So you can. But you appear to be arguing as follows: *some* amateurs have been successful, therefore *all* amateurs are successful. But this is precisely the deeply flawed inductive reasoning which you have (wrongly) imputed to me. In my experience, most amateur linguists are incompetent, and some of them are downright crazy. If you thing I'm being unreasonable, then answer this question: would you be willing to fly on the maiden flight of an airliner designed by an untrained amateur designer and piloted by an untrained amateur pilot? Even if they've both spent "a few minutes" learning "a few simple rules"? [on the danger of relying on secondary sources] > That is pretty useless. Nobody can learn every language. True, of course, but this sad observation cannot change the fact that working on languages you don't know is dangerous. > and nobody can be a mathematician, physicist, engineer, computer > scientists, chemist, economist, etc. That is not how it works at > all. In fact, there is a thing called "specialization of labor" and > has been around for a long time. In fact, it is used by linguists > all the time, since they defer to specialists all the time, even to > the point of committing the "argument from authority" fallacy. Not guilty. Look. One of the non-existent "Basque" words cited as a comparandum by Chirikba in that article I mentioned is the alleged * `this same'. Why do I declare it to be non-existent? Because some authority has said so? No, not at all. I have never heard this word from any Basque-speaker. I have never read it in any Basque text. I have never encountered it in any specialist work on Basque. I cannot find it in any of my numerous Basque dictionaries, some scholarly, others popular. My Basque-speaking friends do not know it, and my specialist colleagues have never heard of it. I am therefore arguing, not from authority, but from evidence. [on learning Sumerian] > What exactly is there that is going to take 12 years to learn? Some symbols > arranged in some order with some presumed meanings which can be gleaned > from multilingual transcriptions. What is there to gain? Mr. Hubey, are you suggesting that one need not spend years studying Sumerian in order to know Sumerian? Time for a reality check, I think. ;-) > Surely, there are bad linguists as well as bad engineers. And there is > bad linguistics as well as good. So what? How do you propose to fix > this problem? It is beyond my power to fix it. All I can do is to point out errors when I encounter them -- which, in my case, is pretty damn often. > Let me guess. You want to have a stamp by the International > Linguistics Association and you want to stamp every book that even > touches upon linguistics as bad and good. Is that a solution? Well, > the arguments often made by linguists, (like a few already made > here) are essentially informal versions of this. And they stink as > badly as the formal version would stink. Mr. Hubey, I fear that you do not know very much about linguistics. Perhaps that ten minutes wasn't quite long enough after all. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Tue Nov 10 23:24:16 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:24:16 EST Subject: The good Dr. Tuna Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > > Nope. I mean only that I am unwilling to believe that something is true > of Sumerian merely because Tuna has asserted it -- especially if he is > not a specialist in Sumerian. > But I am not arguing from authority. I am merely pointing out that > facts are facts, independently of what anybody, authoritative or not, > thinks about them. But an authority, by definition, is far more likely > to know what the facts are. How are both true? Tuna only points out Sumerian words as found in scholarly journals, which presumably are written by Sumerologists or at least read and reviewed by Sumerologists. Tuna merely lists the Sumerian words with their meanings as told by Sumerologists. But now you are claiming that you will not believe Tuna if he is not a Sumer specialist which means that you will only believe authorities in Sumerian. To me it sounds like you believe that Tuna has falsified evidence, and also believe that only Sumerologists should make use of Sumerian words in any journal or book. > > 5. The simple facts are that there are simple rules in linguistics and > > it does not take longer than a few minutes to get the hang of the basic > > idea, > > Really? You mean all of us have wasted our time in spending years > learning our trade, when all we needed to do was to sign up for Doctor > Hubey's Patented Ten-Minute Education in Linguistics? Gosh. Why don't you post some complex rules of historical linguistics and let us see how long it takes to comprehend. > > OK. Here's a problem from my field. The four major regional variants > of the Basque word for `ear' are as follows: > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > discover it? I think you should try 1) reading what is written instead of what you think is written 2) explaining what I asked you last year on another list On the other list and other lists when I ask "experts" to explain what rules are used to construct protoforms and why they can't be found in textbooks, I notice that there is a lot of hemming and hawing. It sounds like Truman's refrain about economist; "I wish I had some one-handed economists". When I ask for rules on constructing protoforms (i.e. algorithms) there is no answer. When I ask why the field is soft and fuzzy, people like you get insulted and shout that it is a real science. Either there are rules for constructing protoforms or there aren't. My original question is/was why there is no algorithm for producing protoforms. Either what you practice is a science or it is not. If it is magic you don't have to explain it. If it is science it should be possible to see it in writing in some book. As for the "regular sound change rule" as applied to Basque ear, it is plainly possible to see that the algorithm is a "partial Caesar cipher". A Caesar cipher (the easiest cipher to crack) consists of changing every letter (sound) to another one by shifting them by an integer n mod M. In the case of linguistics it is "partial" because only some sounds get changed. In the case above if we make the equivalence l=h=g=* we can write the set as {be*arri,biarri}. Now we equate *=0=# and e=i=@ and obtain b@#rri, where @=ei and #=l=h=g=0. Now we go back to the original problem of language universals, and whether h>g is more common than g > h etc. This is why I asked the question of how protoforms are constructed in the first place. How, if it is not clear which is more common (e.g. h>g or g>h) then do linguists (i) choose one, AND, (ii) at the same time claim that it is a science. > In my experience, most amateur linguists are incompetent, and some of > them are downright crazy. I will refrain from telling you what my experience with (some) linguists is :-) However I could easily tell everyone what I know of your competencies as displayed already on other lists, but I will also let that pass and let you demonstrate it on your own the same way you demonstrated it on other lists (re: context-free grammars vs contex-sensitive grammars, definition of language, etc). All you have to do is answer the question which you have dodged for over a year. I will repeat it: How, if it is not clear which is more common (e.g. h>g or g>h) then do linguists (i) choose one, AND, (ii) at the same time claim that it is a science. Of course, I already know the answer; just pleases me to watch you trip over yourself :-) or try to evade it, (again, and again). > > True, of course, but this sad observation cannot change the fact that > working on languages you don't know is dangerous. It's even more dangerous to take guesses on topics which should be answered and answerable by mathematics. > Mr. Hubey, are you suggesting that one need not spend years studying > Sumerian in order to know Sumerian? In order to produce a list of cognates all you need is a dictionary. There are other things that will help, but since no linguist seems to do this, it does not matter. > It is beyond my power to fix it. All I can do is to point out errors > when I encounter them -- which, in my case, is pretty damn often. Pointing out an error is not the same thing as trying induction from one single example and pretending that you seem to have discovered a law. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mghiselin at casmail.calacademy.org Tue Nov 10 23:24:47 1998 From: mghiselin at casmail.calacademy.org (Michael Ghiselin) Date: Tue, 10 Nov 1998 18:24:47 EST Subject: Cladistic language concepts Message-ID: ----------------------------Original message---------------------------- Dear Dr. Dyen: Yes, we can separate different parts of the genome according to how rapidly they change. The DNA that codes for RNA changes relatively slowly though it depends on the part of the RNA molecule. It is thought that the most conservative parts of the RNA molecule are areas where a change would abolish function: and since the function is making proteins that would be a lethal. Mitochondrial RNA evolves faster than nuclear RNA and the reason seems to be that there is less of that kind of constraint. I am not sure what you mean by rates of branching. If there is rapid speciation the probability that there will be evidence of this in the data is greater if the molecule in question is evolving rapidly. But remember that branching rate is a different thing from the rate of substitution. It would be wonderful if all substitutions took place at a constant rate, for then we would have a "molecular clock" but we don't Sincerely, Michael Ghiselin From hubeyh at montclair.edu Tue Nov 10 23:25:02 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:25:02 EST Subject: [Fwd: [Fwd: Wake up call for the semester :-)]] Message-ID: ----------------------------Original message---------------------------- How do historical linguists evaluate these? > H.M.Hubey wrote: > > > > Other lists are doing it, why not us :-) > > > > How about these: > > > > 1. Meroitic -k, Barea -ge: Fenno-Urgic -k 'to' (e.g. Ingrelian ala-k > > 'down'. > > > > 2. Meroitic -te, Nubian -do locative suffix 'in': Old Turkish -ta, -da > > 'in' Finnish -ta 'in' > > > > 3. Meroitic -k feminine suffix: Mongolian -k-chin feminine of > > adjectives; Meroitic kdi 'woman': Turkish kari 'woman' (correspondence > > d:r looks better than d:ss but to make the matter even more surprising, > > there is one Eastern Turkish language, where the word for woman is > > kissi!) > > > > 4. Meroitic t demonstrative, Nubian ter 'he' etc. Mongolian tere 'he' > > 'that', Finnish 'te' 'this one' (I used te instead of ta-unlaut) > > > > 5. Old Nubian -ka accusative suffix: Old Turkish -g, -ig, Mongolian -g, > > -gi accusative suffix. > > > > 6. Old Nubian -ka dative suffix: Old Turkish -qa, -ke dative suffix > > > > 7. Old Nubian -n(a) genitive suffix: Mongolian -in, -n, Fenno-Ugric -n > > genetive suffix. > > > > 8. Old Nubian -r 'intentive' verbal suffix; Old Turkish -r, Finno-Ugric > > -r factitive verbal suffix. > > > > 9. Meroitic tar 'give' causative verbal affix (according to Dr. Priese) > > Old Nubian tir 'give' causative verbal affix: Old Turkish -tur 'give' > > causative verbal affix > > > > 10. Old Nubian -a participle, conjunctive converb: Old Turkish -a > > conjunctive converb > > > > 11. Old Nubian -ra predicative converb: Mongolian -ra final converb > > > > 12. Old Nubian -sa verbal participle praeteriti: Mongolian -san > > participle praeteriti > > > > 13. Old Nubian -s verbal suffix, praeteritum: Fenno-Ugric -s verbal > > suffix, praeteritum (cf. Old Nubian ki-s-in 'you came' with Wogulian > > min-s-en 'you came') > > > > 14. Old Nubian -men (-m-en) negation of verbs: Old Turkish -ma negation > > of verbs > > > > 15. Old Nubian -in,-en verbal suffix, 'you' 2 sg: Wogulian -en verbal > > suffix 'you' 2. sg. > > > > 16. Old Nubian possessive pronoun=genetive of personal pronoun (ir > > 'you', in-na 'your'): Old Turkish the same ('sen' 'you', san-ing > > 'your'), Mongolian the same (chi 'you', chinu 'your') > > > > 17. Old Nubian -t, -it deverbal nouns: Old Turkish -t, -it,-id deverbal > > nouns > > > > 18. Old Nubian -ki deverbal nouns: Turkish -ki abstract nouns, > > Finno-Ugric -k deverbal nouns > > > > 19. Old Nubian min 'what', Mongolian men 'what': Wogulian men 'what', > > Hungarian mi 'what'; > > > > 20. Old Nubian -guria 'because of': Turkish -gore 'because of' > > > > This connects Eastern-Sudanic (Old Nubian) with Uralo-Altaic. > > > > OK. I spill the beans: these are from Fritz Hintze's article. That is > > the reason for the strange spellings. Anyway, I think family > > relationships based on less than this have already been proposed and > > accepted in some cases in the Americas and in Africa. Is that right? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From tonybreed at juno.com Tue Nov 10 23:25:59 1998 From: tonybreed at juno.com (D. Anthony Tschetter-Breed) Date: Tue, 10 Nov 1998 18:25:59 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- Forgive me for being anectdotal, but when topic comes up of sound changes from X to Y and back to X, I always think of the English: askan > ask > ax (in some dialects, for example Black American English) One could propose that this is not an example of X>Y>X either on the grounds that the "ax" form is not an innovation but rather a retention of an earlier form, or perhaps that it's not systematic. I don't know. Any thoughts? -Tony ___________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866] From larryt at cogs.susx.ac.uk Tue Nov 10 23:26:22 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:26:22 EST Subject: Doing historical linguistics (part 2) In-Reply-To: <3644B37B.BC3CE929@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Sat, 7 Nov 1998, H.M.Hubey wrote: [LT] > > OK. Here's a problem from my field. The four major regional variants > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > discover it? > I think you should try > 1) reading what is written instead of what you think is written > 2) explaining what I asked you last year on another list > On the other list and other lists when I ask "experts" to explain > what rules are used to construct protoforms and why they can't be > found in textbooks, I notice that there is a lot of hemming and > hawing. That's because there are no "rules", in the mechanical or algorithmic sense of the term. Performing good reconstructions requires both knowledge of the languages and experience of the craft. Asking a linguist how to do reconstruction is not like asking a mathematician how to solve a differential equation. It's more like asking a professional cyclist how to ride a bike. (Not a good analogy, but the best I can do off the top of my head.) Your experts are hemming and hawing because they can't find any simple way of explaining the procedure to a novice, not because they don't know how to do it. > It sounds like Truman's refrain about economist; "I wish I had > some one-handed economists". When I ask for rules on constructing > protoforms (i.e. algorithms) there is no answer. When I ask why > the field is soft and fuzzy, people like you get insulted and shout > that it is a real science. No, not insulted, just exasperated. And anyway you've already had your answer: there are *no* algorithms for constructing proto-forms, and there cannot be. > Either there are rules for constructing protoforms or there aren't. Oh, there are certainly rules, but there are no algorithms. For example, given the observed variant forms of the Basque word for `ear', there is only one reconstruction that obeys all the rules. > My original question is/was why there is no algorithm for producing > protoforms. Either what you practice is a science or it is not. If > it is magic you don't have to explain it. If it is science it should > be possible to see it in writing in some book. No; this doesn't follow. In chemistry, for example, there are no rules for discovering new classes of compounds, but chemists frequently discover new classes of compounds nonetheless, and they are pleased to call this activity "science". Scientific activity is not, in general, algorithmic in nature. Algorithms, in fact, are in most cases only useful for obtaining answers to practical problems, and not in obtaining new knowledge. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Tue Nov 10 23:26:41 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Tue, 10 Nov 1998 18:26:41 EST Subject: Doing historical linguistics (part 3) In-Reply-To: <3644B37B.BC3CE929@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Sat, 7 Nov 1998, H.M.Hubey wrote: [LT] > > OK. Here's a problem from my field. The four major regional variants > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > discover it? > As for the "regular sound change rule" as applied to Basque ear, > it is plainly possible to see that the algorithm is a "partial > Caesar cipher". A Caesar cipher (the easiest cipher to crack) consists > of changing every letter (sound) to another one by shifting them > by an integer n mod M. In the case of linguistics it is "partial" > because only some sounds get changed. > In the case above if we make the equivalence l=h=g=* we can write the > set > as {be*arri,biarri}. Now we equate *=0=# and e=i=@ and obtain b@#rri, > where @=ei and #=l=h=g=0. > Now we go back to the original problem of language universals, and > whether h>g is more common than g > h etc. This is why I asked the > question of how protoforms are constructed in the first place. > How, if it is not clear which is more common (e.g. h>g or g>h) then > do linguists (i) choose one, AND, (ii) at the same time claim that > it is a science. Neither h > g nor g > h strikes me as a very plausible change, at least not as a single step. If we allow these to be multi-step changes, then I would expect g > h to be more frequent than the reverse. However, none of this is relevant, because your partial Caesar cipher is of no obvious relevance to my question. Not one of the four attested forms is a plausible ancestor for the other three, and the original form must have been different from all of them. In fact, our preferred reconstruction is *, which accounts perfectly for all four of the recorded variants. But, of course, the probable correctness of this reconstruction is only obvious if you are intimately acquainted with the facts of Basque. Without such knowledge, * looks no more plausible than, say, * or * -- both of which are quite impossible. > I will refrain from telling you what my experience with (some) > linguists is :-) However I could easily tell everyone what I > know of your competencies as displayed already on other lists, but > I will also let that pass and let you demonstrate it on your own > the same way you demonstrated it on other lists (re: context-free > grammars vs contex-sensitive grammars, definition of language, etc). Oooooh, Mr. Hubey. [Note of explanation: Some time ago, on another list, Mark asked me to explain what a context-free grammar was. I did my best, but apparently I was unsuccessful.] > All you have to do is answer the question which you have dodged for > over a year. I will repeat it: > > How, if it is not clear which is more common (e.g. h>g or g>h) then > do linguists (i) choose one, AND, (ii) at the same time claim that > it is a science. Mr. Hubey, I'm afraid this question is unintelligible. It's rather like asking "What is the commonest word for `dog' on the planet?" and expecting the answer to be of some relevance to doing historical linguistics. Anyway, this business of "choosing" between h > g and g > h is something you have invented for yourself. We don't do historical linguistics by choosing from a Chinese-restaurant menu. [LT] > > Mr. Hubey, are you suggesting that one need not spend years studying > > Sumerian in order to know Sumerian? > In order to produce a list of cognates all you need is a dictionary. No, Mr. Hubey, a thousand times no. This is a common misconception among non-linguists, but it is grossly false. To produce a list of cognates you need far more than a dictionary. To start with, you need to know what you're doing. Trying to compile lists of cognates with nothing but dictionaries -- a common but forlorn activity -- is about as sensible as trying to construct a theory of music by counting the occurrences of all the notes in the entire repertoire of operas. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Tue Nov 10 23:27:04 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:27:04 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Sat, 7 Nov 1998, H.M.Hubey wrote: > My Sumerianist colleagues assure me that there exist a number of > (especially older) works on Sumerian which are outdated and which > contain errors of fact and interpretation, and they complain that > various attempts at seeking relatives for Sumerian have relied upon > these defective and unreliable sources. Yes, on the one hand, no science is immune to error, and on the other hand it is unlikely that pages and pages of materials can be "mostly wrong" and still convey more or less the correct message. If there are going to be changes it will probably be with finer shades of meaning. If I now write a book of cognates from Halloran's list, then some of your "Sumerianist" colleagues will again change them. The chances are, reasoning probabilistically, that small changes will create small effects and large scale changes will void past Sumerian works. Since when does a field get a thorough overhaul? Do you think Hittite needs a thorough overhaul? > > So, the question is not whether Tuna has taken his forms from published > sources, but rather how reliable those sources are. And the only people > who are in a position to judge the reliability of Tuna's sources are > those who have a good specialist knowledge of the language. > > Moreover, even when a Sumerian word is cited correctly, it does not > follow that it is a valid comparandum in comparative enterprises. Many > Sumerian words are attested only in late texts, and not in early ones, > and hence cannot be used as comparanda. (This I know is true.) IT does not matter. Because the Sumer-Turk cognates come from a "proto-Euphratic" substrate and these people happened to be around that neighborhood a couple of thousand years BC. That means that "Turk" (whatever it is) peoples were not living in the forests of the Altai-Siberia region. There is too much evidence against it and Turko-Sumer cognates are just a small part of the story. In other cases, a word may have a transparent etymology within Sumerian, > say as a compound or a derivative, and hence it too is unavailable as a > comparandum. (I don't know if this is the case in Sumerian, but it is > very commonly the case with other languages invoked in remote > comparisons.) In still other instances, a word may have changed its > form or meaning substantially during the historical period of the > language, in which case it is only the earliest form and meaning which > is available for comparison, and not the later ones. So what? In NJ among the hispanics "mines" has replaced "mine" because of hypercorrection. That is one mechanism of regularization and things like that happen all the time. What difference does that make except give ammunition to those who want to make Sumerian fall from the sky? > In short, merely finding a Sumerian word cited in the specialist > literature, even if that literature is reliable (which it may not be), > does not automatically permit you to cite that word as a comparandum in > seeking genetic links. Sure it does. It is roots that count, and the root is presumably in many words that have that root and are derived from it. Where did you find these rules? > A few examples from Basque, since I don't know Sumerian. Among the > Basque words frequently cited as comparanda in seeking genetic links are > the following. > > `dry'. This looks very similar to something in Caucasian and > even more similar to something in Berber. But the earliest sense of the > word is `barren, sterile', and it has developed semantically, via > `dried-up' (of a well or a spring) to `dry'. So you can't have the > modern sense. So what? Many words shift meaning over time. Do you think humans 30,000 years ago had 100,000 words in their lexicons? How about 100,000 years ago? Was there ever a time when humans had maybe 1000 words? How can you be sure that none of these roots has survived (albeit in semantically shifted form)? You can't. That is so simple, it makes me wonder about the motives of people who become so irrational. What are you afraid of losing? > `woman'. The second half looks like words for `woman' or > `wife' in several other languages, and eager comparativists have > brandished this delightedly while throwing away the first half as > meaningless junk. But the earlier sense of this word was `girl', and it > is transparently bimorphemic, from , the regular combining form of Wow. YOu mean there is no semantic connection between girl, woman and female? Do the hipsters and hippies who used the word "bitch" for women make a gigantic error, and there is no connection between female dog and female person? > `female' (which itself is borrowed from Occitan), plus <-kume> And where is that borrowed from? em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), emesal (female speech in Sumerian), emcek (breasts, udder), meme (breast), emzirik, etc etc. Do you understand the probabilistic implication of such patterns? > `offspring, child'. So you can't have the word at all, and you > certainly can't assign the `female' element to the second half. Are you trying to convince me that historical linguistics works only with "exact" meanings and "exact" phonological/phonetic/sound shapes of words? Are you kidding? > `burdock', `limpet'. This looks strikingly like something in > Caucasian. But it is a transparent (and fully regular) borrowing of > Latin `burdock' and its Spanish descendant `limpet'. So > you can't have this one either. Accidents happen all the time. That is the whole purpose of learning and understanding probability theory and statistics. AT least then you don't have to repeat yourself over and over trying to prove general statements by giving 2-3 examples. Did you read what I wrote about the farmer who fed his duck for 33 days? How much value do these examples have? None for me, because I got the general idea years ago after 2-3 examples. I already know that generalities cannot be proven or assumed because of 2-3 examples. I also know how probability theory works, and I can recognize those who are ignorant of it and hence constantly make the same mistake over and over and over again. I also have to tell you that there is no such thing as "proof by repetetion". If you want to pile up your examples and then try statistics please be my guest, go right ahead. > All genuine Basque words, all found in any halfway decent dictionary, > but all totally unavailable as comparanda in seeking relatives for > Basque. People who cite such data as comparanda are wasting their time. Hardly. It is more likely that you are wasting your time trying to convince me that I should take your 2-3 examples as proof and substitute "proof by repetetion" and "proof by example" for logic, probability theory and fuzzy set theory, and the rest of math. > So, I'm not accusing Tuna of falsifying data. But, if he doesn't have a > profound knowledge of Sumerian, and is merely extracting words > incomprehendingly from other people's work, then he is running a very > grave risk of obtaining meaningless results -- just as many others have > done in extracting Basque words from dictionaries without having any > knowledge of the language. Fortunately, most people have common sense and that eventually will drive them to accept conclusions derived from probability theory and logic, and the rest of math because they know that the world around them was built by people who use all the tools in this bag of cheap tricks called math. What is funny is why when you extoll the virtues of correlation-regression analysis (Labov?) to high heaven in your book, you seem to have turned against math on this list? Is that because other linguists extolled Labov and you decided you had to do it? OR is it because you can understand CR analysis but don't understand probability theory? Or is it because you think Labov was a linguist and I am not? OR is it because you think Sumerian is too valuable to allow to be related even in the remotest sense to barbarians when truly civilized Aryans are hanging around? For what reason do you find it necessary to stick to this irrational manner of discourse? > > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Tue Nov 10 23:27:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Tue, 10 Nov 1998 18:27:25 EST Subject: Doing historical linguistics (part 2) Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Sat, 7 Nov 1998, H.M.Hubey wrote: > > [LT] > > > > OK. Here's a problem from my field. The four major regional variants > > > of the Basque word for `ear' are as follows: > > > > > > > > > > > > > > > > > > > > > So: what's the proto-form? And what "simple rule" should be invoked to > > > discover it? > > > I think you should try > > > 1) reading what is written instead of what you think is written > > 2) explaining what I asked you last year on another list > > > On the other list and other lists when I ask "experts" to explain > > what rules are used to construct protoforms and why they can't be > > found in textbooks, I notice that there is a lot of hemming and > > hawing. > > That's because there are no "rules", in the mechanical or algorithmic > sense of the term. Performing good reconstructions requires both > knowledge of the languages and experience of the craft. Ehhh. "craft"?????????????? You mean it is not a science? >Asking a > linguist how to do reconstruction is not like asking a mathematician how > to solve a differential equation. It's more like asking a professional No it isn't. A mathematician can tell you how to solve it. > cyclist how to ride a bike. (Not a good analogy, but the best I can do > off the top of my head.) Your experts are hemming and hawing because I know how to ride bikes and I can tell you how to do it. > they can't find any simple way of explaining the procedure to a novice, > not because they don't know how to do it. Then there must at least be some "fuzzy rules" but I forgot you don't like that either. Maybe there are tendencies and propensities and they can be analyzed using Markov process concepts and you don't like that either. Nevertheless you are sure they can't be used despite not knowing what they are and how they can be used. Is this some special form of logic available only to historical linguists or only to you? Despite all of these problems, nevertheless, you are sure it is a real science deserving of the Nobel prize and does not need any improvement whatsoever because it has already reached the pinnacle of what science can be? Is this the logic and conclusion you expect me to accept? > > It sounds like Truman's refrain about economist; "I wish I had > > some one-handed economists". When I ask for rules on constructing > > protoforms (i.e. algorithms) there is no answer. When I ask why > > the field is soft and fuzzy, people like you get insulted and shout > > that it is a real science. > > No, not insulted, just exasperated. And anyway you've already had your > answer: there are *no* algorithms for constructing proto-forms, and > there cannot be. That again, is the wrong conclusion. What you want to say is this: "I do not know how to construct protoforms and I do not know anyone else who does." Certainly you cannot know whether or not there can ever be an algorithm and you cannot know in what manner they might be constructed. > > Either there are rules for constructing protoforms or there aren't. > > Oh, there are certainly rules, but there are no algorithms. For > example, given the observed variant forms of the Basque word for `ear', > there is only one reconstruction that obeys all the rules. OK. PRoduce the rules, and I will create the algorithms. > > My original question is/was why there is no algorithm for producing > > protoforms. Either what you practice is a science or it is not. If > > it is magic you don't have to explain it. If it is science it should > > be possible to see it in writing in some book. > > No; this doesn't follow. In chemistry, for example, there are no rules > for discovering new classes of compounds, but chemists frequently > discover new classes of compounds nonetheless, and they are pleased to > call this activity "science". I don't believe that. Once again, you are confusing yourself for the scientific community. When someone says "I don't know" it could mean "I don't know but others might know." "I don't know but probably nobody else knows either." "I know that nobody knows." .... > Scientific activity is not, in general, algorithmic in nature. A lot of people would disagree. For example, the great philosopher Popper, or Lakatos. > Algorithms, in fact, are in most cases only useful for obtaining answers > to practical problems, and not in obtaining new knowledge. YOu might be partially right here, but only that; "might" and "partial". But don't give up hope yet. There is a thing called "data mining" and "knowledge discovery science". We will lick that problem too :-) Why do you think there are conferences in which people try to create "artificial scientists" and "artificial intelligence". > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From bwald at HUMnet.UCLA.EDU Tue Nov 10 23:28:09 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Tue, 10 Nov 1998 18:28:09 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- In passing, I noticed Alan King acknowledge a correction by Miguel Carrasquer. Alan wrote: >Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact >that besides Castilian, both Galician and Romanian are also exempt from >this general Romance development. In Castilian and Galician (but NOT in >Romanian, as you point out), there is no /z/ phoneme. I was confusing >these two things. I am curious about this situation because I have long had the impression that what is historically distinctive of Castillian and the Spanish-area around it, but not extending to Portuguese, Catalan or other adjacent Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among them. That is, I never thought that the ancestor of Castiliian, Spanish, whatever, was exempt from the EARLIER Romance process of intervocalic voicing of Latin -s- (among other sounds), but that by LATER developments it devoiced the resulting -z- (in most environments -- in general,in effect). Similarly, /zh/, as in ancestral "*g*ente" 'people' (current "*h*ente"), was devoiced, something like /zh/ > /sh=c,/ (ichlaut) (> h), cf. ojala (o*h=x*ala:) < Arabic in*SH*alla:h 'God willing'. As far as I know, such fricative devoicing is a distinctive feature of the divergence of Spanish from other Romance languages, and Spanish should be celebrated as a language which exemplifies historical devoicing of (certain) fricatives (the ones made with the tongue against the palate). P.S. Interesting in Spanish is the tenacity of the devoicing, so that rather than revoice intervocalically, /s/ has a tendency (found elsewhere) to reduce to "pure" devoicing, i.e., /h/ (and ultimately loss of the segment). Though /s/ > /h/ also occurs before consonants, the tenacity is reminiscent of the previous movement of /sh/ to the back, to /x/ or /h/, rather than toward "revoicing" in any environment. (Unlike /s/, /sh/ only occurred before vowels. /s/ could easily be neutral or even associated with the phoneme /sh/ before consonants, cf. Portuguese, and, indeed, the numerous studies of s > h variable dialects of Spanish find that s > h is more favored to occur before a consonant than a vowel, cf. the history of French /s/+consonant.) In a separate message, Robert Ratcliffe wrote: >But in general I do not see any reason to assume >that no sound change is reversible given enough time. It should be made clear that the issue is directionality of sound change. The issue of "reversibility" usually implies that the set of words affected by the first sound change remain distinct at the time of the second sound change. However, the "given enough time" qualification suggests that preservation of the identity of the class of words is not at issue (since other changes will most often obscure the original word class affected). The issue is only that a second sound change HAPPENS to be the reverse of the first one -- but historically has nothing to do with the first change. According to my understanding, Latin s > z and then Spanish z > s illustrates this point. From R.J.Penny at qmw.ac.uk Tue Nov 10 23:28:26 1998 From: R.J.Penny at qmw.ac.uk (Ralph Penny) Date: Tue, 10 Nov 1998 18:28:26 EST Subject: r and s: Galician Message-ID: ----------------------------Original message---------------------------- Rhotacization of /s/ to [r] before /d/ commonly occurs in Cantabria, at least in the rural varieties I investigated in _El habla pasiega_ and _Estudio estructural del habla de Tudanca_. Ralph Penny > ----------------------------Original message---------------------------- > > > > > > The change of s to r before another consonanant is widespread in southern > > Spain and also in Madrid, so that los dedos may be articulated as lor > > dedoh and buenos dias as buenor diah. > > it is nice to read it, because I never heard a change like "buenor > diah" in Madrid, nor in Toledo, Galicia nor in Southern Spain. You say > "widespread", but I never, *really never*, met it. > > You can hear following (sub)standard variants: > > in Madrid > 1) buenos dias > 2) bueno dia > in the South (from Toledo toward andalucia) > 3) buenoeh diah > 4) buenoh diah > > a Galician should use the standard variation, also "buenos dias", but > "bos dias" (in Galician). > > > Have you ever heard "buenor diah" ? Where do you get such informations? > I would like to read about it. > > j.m. > Ralph Penny School of Modern Languages Department of Hispanic Studies Queen Mary and Westfield College Mile End Road London E1 4NS. Tel: +44 171 775 3139 Fax: +44 181 980 5400 From fcosw5 at mail.scu.edu.tw Tue Nov 10 23:33:13 1998 From: fcosw5 at mail.scu.edu.tw (Steven Schaufele) Date: Tue, 10 Nov 1998 18:33:13 EST Subject: Ket-Na-Dene affiliation? Message-ID: ----------------------------Original message---------------------------- Dear Colleagues, My wife has just forwarded to me a Reuters story (posted on www.cnn.com) about an article by Merritt Ruhlen appearing this week in the Proceedings of the National Academy of Sciences, claiming an affiliation between Ket, the sole surviving Yeniseian language, and the Na-Dene family of North America. The Reuters story begins with the (to a historical/comparative linguist) ominous statement, `A few words in one of the planet's most obscure languages support the theory that Native Americans left Asia in several separate migrations', and goes on to note that Ruhlen's paper `gives examples of 36 words that are similar in the two language families, including the words for birch bark, children and rabbit'. (However, the only actual lexemes mentioned in the Reuters story are the Ket word for `birch bark' -- it's simply asserted that `in several existing Na-Dene languages it is pronounced similarly' -- and the Ket and Koyukon words for `breast'.) The Reuters story goes on to reassure us that the whole hypothesis isn't really based only on 36 lexemes: `Ruhlen found enough other similarities to convince him of the link. "I just picked ouit 36 for this article that looked like the best and most obvious and strongest," he said.' I'm wanting to know, does anybody subscribing to this EBB know anything about this? Is there anything to this proposed affiliation? And if not, is anybody doing anything about clarifying the issue for the general public? I note with great trepidation that the Reuters article trots out mouth-watering statements like `Related words are often easy to spot -- for instance the German word "mutter" is similar to its English counterpart "mother", while the Russian word "brat" looks very much like "brother" and is similar to the Latin root for words like "fraternal".', but completely ignores any mention of such nasty little things as false friends, etc. Considering the trouble i go to in my introductory courses explaining how unreliable this kind of argument is, i'm worried about a story like this in the general press. Best, Steven -- Steven Schaufele, Ph.D., Asst. Prof. of Linguistics, English Department Soochow University, Waishuanghsi Campus, Taipei 11102, Taiwan, ROC (886)(02)2881-9471 ext. 6504 fcosw5 at mail.scu.edu.tw http://www.prairienet.org/~fcosws/homepage.html ***O syntagmata linguarum liberemini humanarum!*** ***Nihil vestris privari nisi obicibus potestis!*** From DISTERH at UNIVSCVM.SC.EDU Tue Nov 10 23:40:23 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Tue, 10 Nov 1998 18:40:23 EST Subject: HISTLING is back Message-ID: Dear Colleagues, As you have doubtless noticed, HISTLING is back from her weekend away from the computer. I apologize for having inconvenienced your mailboxes with the 20 postings I've just sent out. Dorothy Disterheft From delancey at darkwing.uoregon.edu Wed Nov 11 19:37:15 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Wed, 11 Nov 1998 14:37:15 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- > My wife has just forwarded to me a Reuters story (posted on www.cnn.com) > about an article by Merritt Ruhlen appearing this week in the > Proceedings of the National Academy of Sciences, claiming an affiliation > between Ket, the sole surviving Yeniseian language, and the Na-Dene > family of North America. The Reuters story begins with the (to a This idea has been floating around for a while, among the usual suspects. The proposal of a Yeniseian-Sino-Tibetan link was first made quite a few years ago, I believe, though I don't know remember (if I ever knew) who was the original proponent. (I think this is something which Starostin has endorsed, but I'm sure the original proposal is much older than that). And Na-Dene-Sino-Tibetan goes back to Sapir, so it must be true. So, by transitivity, Yeniseian-Na-Dene must be true, too. Simple logic. Actually, though seeing Ruhlen's name attached to any proposal for genetic grouping always makes me suspicious, I'd be very interested to see these data. I spent a couple of days several years ago straining my almost nonexistent Russian and groaning through some Ket verb paradigms, and one of the striking things about the language is the existence of semantically obscure elements in the prefix string, apparently lexically empty preverbs of some sort, which indeed reminded me a lot of Athabaskan "thematic" prefixes. Not the sort of resemblance I'd be ready to hang a major proposal for trans-Pacific relationship on, but it's a rather unusual typological feature. > I'm wanting to know, does anybody subscribing to this EBB know anything > about this? Is there anything to this proposed affiliation? And if > not, is anybody doing anything about clarifying the issue for the > general public? Good question, but--does the general public really want to know better? Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Wed Nov 11 18:20:34 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Wed, 11 Nov 1998 13:20:34 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- On Tue, 10 Nov 1998, Steven Schaufele wrote: > My wife has just forwarded to me a Reuters story (posted on www.cnn.com) > about an article by Merritt Ruhlen appearing this week in the > Proceedings of the National Academy of Sciences, claiming an affiliation > between Ket, the sole surviving Yeniseian language, and the Na-Dene > family of North America. The Reuters story begins with the (to a > historical/comparative linguist) ominous statement, `A few words in one > of the planet's most obscure languages support the theory that Native > Americans left Asia in several separate migrations', and goes on to note > that Ruhlen's paper `gives examples of 36 words that are similar in the > two language families, including the words for birch bark, children and > rabbit'. (However, the only actual lexemes mentioned in the Reuters > story are the Ket word for `birch bark' -- it's simply asserted that `in > several existing Na-Dene languages it is pronounced similarly' -- and > the Ket and Koyukon words for `breast'.) The Reuters story goes on to > reassure us that the whole hypothesis isn't really based only on 36 > lexemes: `Ruhlen found enough other similarities to convince him of the > link. "I just picked ouit 36 for this article that looked like the best > and most obvious and strongest," he said.' > I'm wanting to know, does anybody subscribing to this EBB know anything > about this? Is there anything to this proposed affiliation? I haven't seen this particular article, but Ruhlen has in fact been pushing a link between Yeniseian (the family to which Ket belongs) and Na-Dene for some time. Take a look at chapter 4 of the following book: Merritt Ruhlen (1994), On the Origin of Languages, Stanford: Stanford University Press. (Do not confuse this with Ruhlen's 1994 book of nearly identical title published by Wiley.) This chapter, co-written with Sergei Starostin, proposes 300+ Proto-Yeniseian reconstructions, plus external comparisons with Na-Dene, Basque, Abkhaz-Adyghe, Nakh-Dagestan, PIE, Basque, Burushaski, and Nahali. All the items mentioned by Steven are there, though the details are perhaps different: for example, no Na-Dene comparandum is proposed for the Proto-Yeniseian word for `birch bark'. My own view is that the comparisons on offer are devoid of value. They consist of nothing more than the usual miscellaneous lookalikes. The Basque comparanda, I can testify, are characterized by the most awful collection of errors. Here are the particular items mentioned by Steven, with only Proto-Yeniseian (PY) and Na-Dene (ND) comparanda; diacritics are omitted: `birch back': PY *. No ND comparandum. `breast': PY *; Haida `heart', Tlingit `heart', Kutchin `breast', Tahltan `breast', Hare `breast', Mattole `breast'. `children': PY *; Haida `child', `son', Tlingit `child', `son', Eyak `child', `son', Navajo <[gamma]e?> `son'. `rabbit': PY * ~ *: Eyak `rabbit', Slave `rabbit', Tsetsaut `rabbit', Navajo `rabbit'. > And if not, is anybody doing anything about clarifying the issue for > the general public? This is not easy. Last year the Times of London published a solemn article about a particularly imbecilic book "proving" that Etruscan was Basque, and even added an editorial praising this "achievement". I wrote a letter to the Times drawing attention to the absurdity which they had perpetrated, but I never got so much as a reply, let alone anything in print. It appears that even seemingly serious publications prefer eye-catching drivel to sober assessment. > i'm worried about a story like this in the general press. Me too, but what can we do? Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Wed Nov 11 18:19:46 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 11 Nov 1998 13:19:46 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > > Aren't morphological paradigms also part of 'regular sound change"? > > No, they're not. > > But more to the point, contrary to what Mr Hubey seems to be > suggesting, the little couplet above is not an instance of circular > reasoning. It's merely the same statement phrased two different ways. > Or rather, 2. is a statement of a general principle, of which 1. is a specific application. > > What Mr Hubey may be trying to articulate is the superficially > circular-looking: > > 1. A,B,C are a language family ===> A,B,C have regular sound > correspondences. > > 2. A,B,C have regular sound correspondences ===> A,B,C are a language > family. > Perhaps only some people use it circularly, or many linguists use it circularly, but it is in use. > But the attribution of circularity rests on a misreading of the > relations between the propositions in 1. and 2. as the same. 1. is a > causal relation -- regular sound correspondences result from the > definition of a language family, plus the fact that sound change is > regular. But borrowings also create regular sound correspondences. 2. is a progression from evidence to inference -- we observe > regular sound correspondences, from which we conclude these languages > are a family. (Whether this is an accurate account of what we > actually do is not the question here.) It's no more circular than: > > 1. Patient has measles ===> patient has spots on face. > 2. Patient has spots on face ===> patient has measles. This rests on something different. 1. Patient has measles (definition comes from some other place, but may include spots on the face). These days the defn would come from being able to culture the bacteria. Then the spots on the face and measles correlate. The time depth is short and one can see a non-measles person get it, get sick, etc. Because of the correlation of measles and spots, 2 then becomes an implication. But that does not work so in historical linguistics because we never had a record of any language family (knowing its relatives, etc) but all of it rests on a larger theory of which regular sound correspondence must be a part. So it is a whole mess of correlations which lead towards that conclusion. More to the point it is based on this reasoning. 1. These languages have too many things in common. IOW, there are many words in these languages which can be made to look like each other with similar meanings and which could not be due to chance. 2. If that is not due to chance then either they got these words from each other or the words are all descended from a common language. 3. We have plenty of evidence (what?) that these languages did not get these words from each other. 4. Therefore these words in these languages must all come from an earlier common source. This is how it is supposed to work, but you can see rather easily how and where problems crop up, and where arguments occur. Isn't this basically right? > This may be rough and ready diagnostic practice, but it isn't a > logical fallacy. > > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From martinez at eucmos.sim.ucm.es Wed Nov 11 16:41:09 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 11 Nov 1998 11:41:09 EST Subject: Gr. #s- > #h- Message-ID: ----------------------------Original message---------------------------- > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting and of course a problem like Gr. hu^s / s^us "pig" From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:40:56 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:40:56 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- At the risk of getting caught in the crossfire, I'd like to interject myself in the debate between Larry Trask and Mark Hubey. I'm in agreement with about 90% of what LT has to say, and have to say (with no malice intended) that MH's postings reveal a great ignorance of the field of linguistics. Nonetheless I do think that MH is asking some legitimate questions and deserves a better response from the profession than the old refrain that the esoteric knowledge is only revealed to those who have joined the secret brotherhood. re the following sequence in particular: H. M. Hubey wrote: > So far I have found no book on historical linguistics > in which there is a clear algorithm for reconstructing protoforms. And > > i have asked many people about this. They don't know either. Sometimes > > they can justify it easily sometimes it is a guess and sometimes it is > no more than a concensus. MH again: > > On the other list and other lists when I ask "experts" to explain > > what rules are used to construct protoforms and why they can't be > > found in textbooks, I notice that there is a lot of hemming and > > hawing. LT's response: > That's because there are no "rules", in the mechanical or algorithmic > sense of the term. Performing good reconstructions requires both > knowledge of the languages and experience of the craft. Asking a > linguist how to do reconstruction is not like asking a mathematician > how > to solve a differential equation. It's more like asking a > professional > cyclist how to ride a bike. (Not a good analogy, but the best I can > do > off the top of my head.) Your experts are hemming and hawing because > they can't find any simple way of explaining the procedure to a > novice, > not because they don't know how to do it. My view: That "reconstruction" is (or even could be) a matter of algorithm, or rule is a widely held misunderstanding outside the field, no doubt supported by a misinterpretation of the technical term "reconstruction" in its ordinary sense. We cannot literally "reconstruct" anything. A "reconstruction" is a hypothesis, neither more nor less. Like any hypothesis in any field of science we arrive at it by guesswork, by intuition, by imagination, by accident. There is no path of deductive reasoning, no "discovery procedure", no algorithm which leads from the data to the hypothesis. The rigor in historical linguistics as in the natural sciences is not in the way in which hypotheses are reached but in the way in which the hypotheses are submitted to the test of the data. A good reconstruction should be testable in principle-- it should be a specific hypothetical prediction about the way a particular language was spoken by a particular group of people at a particular point in time. A reconstruction is virtually never (directly) testable in practice-- it is so only on those extremely rare occasions when new texts of ancient languages are unearthed. How do we test it then?-- By implication. Each reconstruction (of a proto-phoneme for example) has implications for the whole system of the proto language (the whole phonological system, eg), for the development path leading from the proto-language to the attested languages (the sequence of sound changes, eg), and for the forms of the reflexes in the descendant languages. Only the last is directly observable, of course, and only this real data can be used to rule out a proposed reconstruction absolutely. The proposed development path implied by the reconstruction cannot be tested as right or wrong but only as plausible or implausible based on the statistical frequency of observed patterns of change. Simlarly the implications which a reconstruction has for the whole system can only be tested under the assumption of the "uniformitarian principle" - the assumption that prehistoric languages were not fundamentally different in kind from attested languages and hence should not show structural anomalies of a type not found in attested languages (languages shouldn't be reconstructed with no vowels, or all verbs, for example). Probability and statistics is involved in all of this, and much could be gained from making it more explicit. Much could also gained if historical linguists made an effort to base assumptions about plausible directions of changes on an explicit, accesible, statistically analyzable body of evidence of changes, rather than on implicit personal knowledge. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From mcv at wxs.nl Wed Nov 11 16:39:30 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:39:30 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36463306.60B6E2F@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: >Larry Trask wrote: >> `female' (which itself is borrowed from Occitan), plus <-kume> > >And where is that borrowed from? > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), >emesal (female speech in Sumerian), emcek (breasts, udder), meme >(breast), emzirik, etc etc. This illustrates the fact that a little knowledge of the languages involved is never a bad thing. Larry should have been slightly more precise by indicating that Basque was not borrowed from vanilla Occitan, but from Gascon/Bearnais. Then it immediately becomes clear that the word is simply Latin FEMINA, which regularly becomes hemna (hemm@ ~ hemno) in Gascon (f > h). The ultimate root is IE *dheH1- "to suck" (*dheH1-mHn-oH2 > femina). The coincidence with PTurkic *eme "woman" is entirely coincidental. And so is, with even more reason, Sumerian where the element doesn't mean "woman" at all, but "language, tongue". The word means "thin, refined" (it used to be thought that was "woman", and one can still find that in older books, but in fact the Sumerian word was ). >Do you understand the probabilistic implication of such patterns? Yes. In short: "Coincidences happen" and "Garbage in, garbage out". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From Georg at home.ivm.de Wed Nov 11 16:39:11 1998 From: Georg at home.ivm.de (Ralf-Stefan Georg) Date: Wed, 11 Nov 1998 11:39:11 EST Subject: The good Dr. Tuna In-Reply-To: <364381BA.3D0EC98D@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- >Where are the books on Sumerian? > >Where are all the books on Hittite? I have a batch of them. It took >be 6 months to collect some and xerox the others. Does this mean that >I must not trust that the OI has done poor work and that I should not >rely "on others". What if I start to look at the original text? > >What exactly is there that is going to take 12 years to learn? A textbook, or well a handbook of, say, Hittite is but a snapshot of common knowledge about such a language at a very precise given point in time. Just as an example: I consider myself well-trained in Indo-European and did study Hittite (including cuneiform) for a while. Yet I'd not even think about citing a Hittite form which may be crucial for some argument of mine without consulting my Hittitologist friends before, who work night and day with the actual texts. Hittite is language which is amply (re: the number of texts), but still quite fragmentarily (re: its vocabulary and morphology) attested. *Every day* during the summer campaigns new texts are unearthed, transcribed, translated and eventually published. Every year at least, some of them lead to sometimes considerable changes in our knowledge of the language. Words are re-assigned to different inflectional classes, meanings are fixed, even the *readings* of cuneiform signs are fixed sometimes even today, previously well-known words, which already made it into the dictionaries are unveiled as errors of previous investigators and so on, and so on. It is quite unlikely that things like that happen too often for languages like Danish, Mandarin Chinese or Guugu-Yimidhirr (one modern language, for which a constant reanalysis of synchronic data is however quite common is Ket), but for a language whose records are three millennia old this is quite common and should surprise nobody. The same holds for the other cuneiform languages, like Sumerian, Hurrian, Urartaean aso. In the case of Hattic, one monograph published last year manages to render almost the entirety of previous publications on this language obsolete (yet Hattic had been assigned to various language families before, too bad). This has nothing to do with the languages being so utterly difficult and unmanagable, the writing system is, the cultural setting of the texts is, and the fragmentary nature of the stone tablets after so many centuries is (there is actually *one* (1) entirely undamaged text of any interesting length in Hittite, the famous bronze-tablet contract, found in the eighties). Learning Hittite does *not* take 12 years, it takes your whole life. Making use of the present state-of-the-art (which may be different after two years) for linguistic purposes takes a phone-call to the experts (but having spent 12 years learning Hittite before that surely helps to understand what they are talking about ;-). St.G. Stefan Georg Heerstrasse 7 D-53111 Bonn FRG +49-228-69-13-32 From Georg at home.ivm.de Wed Nov 11 16:38:47 1998 From: Georg at home.ivm.de (Ralf-Stefan Georg) Date: Wed, 11 Nov 1998 11:38:47 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- >I'm wanting to know, does anybody subscribing to this EBB know anything >about this? The Yenisseyan-Na-Dene affiliation is subscribed to mainly by Sergei Starostin and his Moscow-based school of Nostraticists. It forms part of a much larger grouping, called Dene-Caucasian, comprising no less than the following languages and families: Basque, Iberian (yes, Iberian, though nobody knows anything about Iberian !), Burushaski, Yenisseyan, "North-Caucasian" (comprising NW- and NE-Caucasian which are not generally grouped together by Caucasianists), Hurro-Urartean (lumped with N-Cauc.), Kusunda (an extinct language of Nepal, on which little is known; Bengtson, another supporter of this grouping, recently informed me that Kusunda has been taken out of the grouping; I'm inclined to call this a step in the right direction ;-) and Na-Dene (I'm currently unsure whether further North American lgs. have made it into the family yet), and, sorry I forgot, the whole of Sino-Tibetan, and of course Sumerian. It is clear that one of the goals of this grouping is to hoover up most languages of the Old World, which are currently thought to be isolates. Larry Trask has shown on numerous occasions that (the supporters of) this theory treat(s) Basque data in a less than competent manner; the same can be said about Yenisseyan and much of the Tibeto-Burman (part of Sino-Tibetan) data I've seen in connection with this theory. Parts of this giganto-macro-grouping have some history, though: a Yenisseyan-Sino-Tibetan connection has been en vogue in the earlier days of Yenisseyology with investigators like Donner, Bouda and the outsider Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as I remember this was based on typological resemblances only), Sumerian has been compared to pretty much everything, so has Basque, but I think it is safe to say that a Basque - (unspecified) Caucasian connection has probably lured more early researchers than anything else (probably originally instigated by ergativity, which once made up for a quite exotic look-and-feel of a language - which is hardly the case today). Readers may already have inferred from my slightly ironic tone that I'm personally disinclined to buy much of this (I have working experience with Tibeto-Burman, Yenisseyan, NW-Caucasian, NE-Caucasian and a bit of Burushaski). St.G. Stefan Georg Heerstrasse 7 D-53111 Bonn FRG +49-228-69-13-32 From mcv at wxs.nl Wed Nov 11 16:38:31 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:38:31 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: <36490FB0.2995@mbm1.scu.edu.tw> Message-ID: ----------------------------Original message---------------------------- Steven Schaufele wrote: >My wife has just forwarded to me a Reuters story (posted on www.cnn.com) >about an article by Merritt Ruhlen appearing this week in the >Proceedings of the National Academy of Sciences, claiming an affiliation >between Ket, the sole surviving Yeniseian language, and the Na-Dene >family of North America. Old news. You can find the Yeniseian ~ Na-Dene comparisons (as well as Yen ~ NCauc./SinTib./Basque/Burush./Nahali) in Ruhlen's 1994 book "On the Origin of Languages" (make sure it's the Stanford UP one, not Ruhlen's "The Origin of Language", Wiley & Sons, 1994). There's also a chapter on Na-Dene itself. For an updated list of Proto-Yeniseian reconstructions (with some comparisons to North Caucasian and Sino-Tibetan only -- the Na-Dene, Basque, Burushaski, Nahali ones were added by Ruhlen [and very poorly so for Basque, I must say]), you can consult or download Sergej Starostin's online database at http://starling.rinet.ru/ (also contains online etymological materials on North Caucasian [Nikolaev & Starostin's NCED], Dravidian, Altaic, Sino-Tibetan, Chuckchi-Kamchatkan, as well as Ozhegov, Zalizniak and Mueller's Russian(-English) dictionaries). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Wed Nov 11 16:38:19 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Wed, 11 Nov 1998 11:38:19 EST Subject: s > r (Iberian) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- bwald wrote: >I am curious about this situation because I have long had the impression >that what is historically distinctive of Castillian and the Spanish-area >around it, but not extending to Portuguese, Catalan or other adjacent >Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among >them. The phenomenon does extend to "apitxat" Valencian Catalan, i.e. the central Valencian dialects of Valencia city and province (not N and S Valencian of Castello' and Alacant provinces). The rules are: -v- > -b- venia = /be'nia/ (also in std. Cat. /b@'ni@/) -z- > -s- casa = /'kasa/ (std. Cat. /'kaz@/) -Z- > -S- pujar = /pu'Sar/ (std. Cat. /pu'Za/) -dZ- > -tS- metge = /'metSe/ (std. Cat. /'medZ@/) I think "apitxat" also shares with Lleidata` Z > dZ initially and after /n/, so that the tongue twister ("16 judges of a courthouse eat liver from a hanged man") becomes: std. lg. /sEdz@ ZudZ at s dun ZudZat menZ at n fedZ@ dun p at nZat/ Lleidata` /sedze dZudZes dun dZudZat mendZen fedZe dun pendZat/ Valencia` /setse tSutSes dun tSutSat mentSen fetSe dun pentSat/ The big difference with Castilian is that there is no Z/S > x development (nor dz/ts > /T/). ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From martinez at eucmos.sim.ucm.es Wed Nov 11 16:38:07 1998 From: martinez at eucmos.sim.ucm.es (Javier Martinez) Date: Wed, 11 Nov 1998 11:38:07 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > But some sound changes are quite irreversible. Consider loss. In the > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting > /h/ was later lost. I predict confidently that the Greeks will never > reverse this change by re-introducing those long-gone /s/s, yes. as a product of analogy, see the -s- futures etc. From mccay at redestb.es Wed Nov 11 16:37:36 1998 From: mccay at redestb.es (Alan R. King) Date: Wed, 11 Nov 1998 11:37:36 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Benji Wald wrote: >In passing, I noticed Alan King acknowledge a correction by Miguel >Carrasquer. Alan wrote: > >>Yes, a slip. I guess I was thinking of intervocalic s > z, and the fact >>that besides Castilian, both Galician and Romanian are also exempt from >>this general Romance development. In Castilian and Galician (but NOT in >>Romanian, as you point out), there is no /z/ phoneme. I was confusing >>these two things. > >I am curious about this situation because I have long had the impression >that what is historically distinctive of Castillian and the Spanish-area >around it, but not extending to Portuguese, Catalan or other adjacent >Romance languages is the *devoicing* of voiced fricatives, /z/ > /s/ among >them. That is, I never thought that the ancestor of Castiliian, Spanish, >whatever, was exempt from the EARLIER Romance process of intervocalic >voicing of Latin -s- (among other sounds), but that by LATER developments >it devoiced the resulting -z- (in most environments -- in general,in >effect). Oh dear, I put my foot in it again. Apologies. I didn't really mean to say that, I can't even imagine how it happened. Must have been in a hurry. Of course I am wrong again, and you're right to pull me up on it. What I should have said (see if I get it right this time) is that the intervocalic s > z development, general in Romance (all? most?), was later reversed in Castilian and Galician, as opposed to most other Romance languages. In modern Castilian and Galician there is no /z/ phoneme. In most other Romance languages there IS such a phoneme, AND it comes from intervocalic /s/ in Latin. I know of one other modern Romance language which, synchronically, does not have /z/ deriving, historically, from Latin intervocalic /s/. As pointed out, Romanian (unlike Castilian and Galician) DOES have a /z/ phoneme, but from different historical sources than the /z/'s in the other Romance languages we're talking about. Whew! I hope that's better. So thanks for the correction. But what I was (we were) talking about, with regard to which this was a mere tangential point, was this (this is a reiteration of what I've already said, or tried to): (1) I believe that all languages in the Iberian peninsula, in their present (synchronic) forms, voice /s/ before voiced consonants in those cases (varieties) where the sibilant is not subject to some more radical transformation (rhotacism, aspiration, loss...). Even peninsular Basque, for what it's worth. (2) As regards the (synchronic) structural consequences of this voicing, we must differentiate between: a. languages which possess a /z/ PHONEME (and also a /s/ phoneme, of course); here we can talk about neutralization in the context in question. b. languages which LACK (synchronically, of course) a /z/ PHONEME; here we can only talk about conditioned allophones of /s/. Within the Iberian Peninsula, the languages in the a. group are Catalan and Portuguese. Those in the b. group are Castilian, Galician and peninsular Basque. Some varieties of Castilian (and also of Galician) are exempt from this classification because they do other things with their sibilants, such as aspirate them or drop them. And some dialects of Galician, and of Castilian (I thought and still think), rhotacize in the pre-voiced-consonant position. (I can provide documentation for this for Galician, and actually already have done so; for Castilian, which I have not investigated systematically, it is merely an impression, and I stand to be corrected if need be!) I say nothing about other Iberian languages or dialects, such as Asturian, owing merely to ignorance, but the corresponding information is most welcome. I wonder what I mangled this time?! In humility, Alan From drc at antnov1.auckland.ac.nz Wed Nov 11 16:37:17 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Wed, 11 Nov 1998 11:37:17 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Tue, 10 Nov 1998 18:21:52 EST > Reply-to: hubeyh at montclair.edu > From: "H.M.Hubey" > Organization: Montclair State University > Subject: Re: rhotacism from Ray Hickey > To: HISTLING at VM.SC.EDU > ----------------------------Original message---------------------------- > Scott DeLancey wrote: > > > > On Fri, 6 Nov 1998, H.M.Hubey wrote: > > > > > Historical linguistics is circular, especially as it is based mostly on > > > IE. > > > IT says; > > > > > > 1. The set of languages, {x,y,z...} constitutes a family because of > > > 'regular > > > sound correspondences'. > > > 2. 'Regular sound correspondences' indicate a 'language family'. > > > > Could you give us a couple of examples of recognized language families > > that have been established on the basis of this kind of reasoning? > > There are a few, but it's not the standard methodology at all. > > Not Indo-European, for sure, which was first proposed and established > > on the basis of extensive correspondences in morphological paradigms. > > Aren't morphological paradigms also part of 'regular sound change"? No, they're not. But more to the point, contrary to what Mr Hubey seems to be suggesting, the little couplet above is not an instance of circular reasoning. It's merely the same statement phrased two different ways. Or rather, 2. is a statement of a general principle, of which 1. is a specific application. What Mr Hubey may be trying to articulate is the superficially circular-looking: 1. A,B,C are a language family ===> A,B,C have regular sound correspondences. 2. A,B,C have regular sound correspondences ===> A,B,C are a language family. But the attribution of circularity rests on a misreading of the relations between the propositions in 1. and 2. as the same. 1. is a causal relation -- regular sound correspondences result from the definition of a language family, plus the fact that sound change is regular. 2. is a progression from evidence to inference -- we observe regular sound correspondences, from which we conclude these languages are a family. (Whether this is an accurate account of what we actually do is not the question here.) It's no more circular than: 1. Patient has measles ===> patient has spots on face. 2. Patient has spots on face ===> patient has measles. This may be rough and ready diagnostic practice, but it isn't a logical fallacy. Ross Clark From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:36:15 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:36:15 EST Subject: probability & sound change (nee rhotacism from R.H.) Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > Robert R. Ratcliffe wrote: > > > >The proposal that given enough time X>Y>X is > > plausible is based on a hypothesis that the directionality of sound > > change is random. It isn't. There are clearly preferred (frequently > > This is something that used to bother me a lot and I used to think > that > there must be some universality. I am still not sure that there is not > > but I am more sure now that some of what we see is not really > universal but rather due to some specific local phenomena. This is not a matter of opinion, or something that can be decided simply by thinking about it. It is a matter of evidence, and that is what I meant by 'empirical'. Labov's conclusion, based on case studies of some twenty languages, is: ".. there are no directions of vowel shifting that are forbidden to speakers of human languages, but ... some directions are taken far more than others" (Principles of Linguistic Change 1994, p.116). Specifically he proposes that in chain shifts, long vowels rise, short vowels fall, and back vowels move to the front. > > attested) and disprefered (rarely or not attested) directions for > sound > > change. The only systematic work on this I know of is in Ch. 5 of > > Labov's 1994 "Principles of Language", which only deals with vowels. > This has to be from a specific time period and a specific set of > languages. Labov's sample (mistake in title, should be Prinicples of Linguistic Change, sorry) is small by the standards of typology, and most of the languages are European. Further work in the diachronic typology of phonological systems is certainly desirable. Nonetheless, as empirical scientists we have to prefer a theory consistent with some evidence (directional sound change) over a theory (random sound change) which has not been shown to apply to any evidence. It is certainly possible that directionality effects are part of a short term cycle which evens out to zero in the long run, but I haven't seen any evidence which would support this view. > There is a similar problem in probability theory, that of stationarity > of a signal. It's impossible to prove. No matter how many languages > you > look into you can only look at a specific time interval. It's hard to > say if the sound changes are due to universals or to a particular > combination of sounds and phonotactics. This passage is a bit cryptic. I haven't talked about universals, but only about emprically observed statistical tendencies, and haven't discussed explanatory factors at all. It would certainly be meaningless to say that something is "due to" a statistical tendency. The length of the time interval is only limited by the historical record, but even so, it isn't necessarily a problem. If directional effects are only part of a long term cycle X>Y>X (a cycle longer than the historical record), this should still be demonstrable, based on the fact that not all languages should be at the same point in the cycle. In some langues X>Y should be the trend in the attested period, in others Y>X. Breadth of the sample can substitute for depth. > However, what I wrote above referred, in general, to any two sounds. > It > may be that for some specific X and some specific Y, the sound change > X > Y for some specific language (i.e. specific set of phonemes and > phonotactics) may be irreversible. But in general I do not see any > reason to assume that no sound change is reversible given enough time. You're right, of course. Directional trends relate only to what is probable in sound change, not to what is possible or impossible. On the other hand it occurs to me that there is one type of sound change which IS irreversible, that is X>0, as found in mergers, phoneme loss, or assimilation in clusters. > Right. However, it is strange on the other hand to see those consonant > > clusters and lack of vowels in languages like Abaza, Georgian, or > Khoisan and its clicks. The foremost question is this: if there is > such a > universal trend (say toward lack of cases, or toward voicing, or from > stops to fricatives, or approximants) how then did the language (any > language) get those stops in the first place? Or how did some language > get consonant clusters at all? Directionality does not mean teleology. To say that sound change moves in a non-random direction is not to say that it is going anywhere in particular. One might propose (although I didn't) that over time phoneme inventories get smaller, or that 'marked' phonemes tend to be lost in favor of unmarked. I have yet to see a teleological proposal of this type which can survive the test of the evidence. The directionality trends discovered by Labov appear ultimately due to physiological constraints both articulatory and auditory. In other words the direction of sound change is constrained by the present not by the future. I suppose this is what you mean by talking about local versus universal (i.e. short term vs. long term). There are no long term directional trends (as far as the evidence now goes). But we cannot conclude from this that the sum of short term directional trends adds up to zero or no directionality or random directionality. The prequisite to a realistic probalistic model of language change is a systematic research program in diachronic typology. We have to establish the probable direction of change on the basis of statistical data from observed changes-- not on the basis of a priori reasoning. I suspect the ultimate model will look something like those typhoon maps we see here (don't know if you have them in other countries): There is a circle showing where the storm is and from the outer edges of the circle there is a wedge shaped projection showing the area into which the storm is likely to move. At the big end of the wedge there is another bigger circle showing where the storm is likely to be at the end of a certain period (24 hrs, e.g.). From this big circle another wedge is projected, with a bigger mouth showing where the storm should go in the next period, and so on. > The same applies to linguistics changes.There are many scales at which > changes occur, and if we mix up these levels we create opposing ideas. A priori we don't know what models or scales apply to linguistic changes. We can't know until we have tried to apply them, that is test them against the data. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From artabanos at mail.utexas.edu Wed Nov 11 16:34:52 1998 From: artabanos at mail.utexas.edu (Tom Wier) Date: Wed, 11 Nov 1998 11:34:52 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- "D. Anthony Tschetter-Breed" wrote: > Forgive me for being anectdotal, but when topic comes up of sound changes > from X to Y and back to X, I always think of the English: > > askan > ask > ax (in some dialects, for example Black American English) > > One could propose that this is not an example of X>Y>X either on the > grounds that the "ax" form is not an innovation but rather a retention of > an earlier form, or perhaps that it's not systematic. I don't know. Any > thoughts? The form in question I believe is actually connected to other dialects in the South which themselves go back to nonstandard dialects in England. OE had both "bscian" and "bcsian", and dialects in America were not homogenously from Southern England around London (on which dialect the standard was based) or anything, so I find it far more likely that it's merely a retention of the earlier form. ======================================================= Tom Wier ICQ#: 4315704 AIM: Deuterotom Website: "Cogito ergo sum, sed credo ergo ero." "The heart is deceitful above all things, and desperately corrupt; who can know it?" Jeremiah 17:9 ======================================================== From ratcliff at fs.tufs.ac.jp Wed Nov 11 16:34:32 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Wed, 11 Nov 1998 11:34:32 EST Subject: s > r (Iberian) correction Message-ID: ----------------------------Original message---------------------------- bwald wrote: > In a separate message, Robert Ratcliffe wrote: > > >But in general I do not see any reason to assume > >that no sound change is reversible given enough time. > This wasn't me. But Mark Hubey responding to something I wrote. Have to be careful with these interlinear discussions. My point was that some directions of change are more or less probable not that any are necessarily impossible. But some changes ARE irreversible, as Larry Trask pointed out earlier in this thread-- namely phoneme loss, mergers, and I would add assimilation in clusters (okto >> otto type changes). -- +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From hubeyh at montclair.edu Wed Nov 11 16:32:26 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Wed, 11 Nov 1998 11:32:26 EST Subject: Hawaiian meli Message-ID: ----------------------------Original message---------------------------- Lyle Campbell wrote: > > ----------------------------Original message---------------------------- > Ross Clark's example of Hawaiian -meli- 'honey' as a loanword from Greek > via classically-educated missionaries translating the Bible is a very nice > one, as Larry Trask pooints out. Larry notes that it ruins one of his > favorite examples of chance resemblance, but becomes a particularly > interesting example of borrowing. I suppose you could sort of raise the > ante on interestaing cases of chance resemblance and borrowing by throwing > Maori -mieri- 'honey' into the mix. A comparison of Hawaiian -meli- and > Maori -mieri- (bother the -i- vowel difference) might seem to suggest a > Polynesian cognate set (throw in Niuean -meli- 'heney' as well, also > apparently from Greek), but the Maori word is actually a French loanword > (from French -miel- 'honey'), courtesy apparently of early French Catholic > missionizing activities in New Zealand, which soon faded in the country. > (There are not many French loans in Maori, but a favorite is -wi:wi:- > 'French' < French -oui- 'yes'). As Larry Traks points out, this -meli- / > -mieri- false cognate is no longer a case of sheer accidental similarity, > in that both are from Indo-European languages, but we still have accident > to thank for it in a way, in that by sheer happenstance Hawiian ended up > with a Greek form and Maori with a French one (which happen to be related > languages), not something that would have been expected. These are nice examples. But that is like the fisherman's game. The one that got away was the biggest of them all. How are we to take into account all of those words that are said to be cognate in IE when it is quite possible that many (most) might be left over a substratum that was living in that neighborhood for many thousands of years and had stabilized so that changes were taking place very slowly? Of course, this is supposed to point out that false matches can occur. True. IT is also possible that we might miss out on real cognates. Maybe they cancel each other. That is why statistics is for. YOu take N measurements and average them, and the random deviations cancel. It also tells you how much confidence you can have in your results. That is a side benefit. > Lyle Campbell -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From johanna at uclink.berkeley.edu Thu Nov 12 01:27:35 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Wed, 11 Nov 1998 20:27:35 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36463306.60B6E2F@montclair.edu> Message-ID: ----------------------------Original message---------------------------- Mark Hubey writes: >----------------------------Original message---------------------------- > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), >emesal (female speech in Sumerian), emcek (breasts, udder), meme >(breast), >emzirik, etc etc. > >Do you understand the probabilistic implication of such patterns? > I'll take the liberty of commenting, as I understand the probabilistic implication of such patterns. What are the chances of finding, in any language, a word consisting of a vowel (any vowel? or just a non-round vowel?) followed by [m], with or without any following material of any shape, having one or the other of the set of 7 glosses cited above plus presumably any other meaning that might have to do with females? That boils down to: what are the chances of finding a sequence Vm(-) in any of a wide (or open-ended) set of meanings? Success is virtually guaranteed, since you get to keep looking until you find something that fits. I've computed the chance of finding resemblances in which two similar consonants occur in the same order in words with similar senses. If you're allowed: (1) a couple of distinctive features' leeway in defining generic consonants (so e.g. p, b, and f are all taken to match) (2) and much phonotactic leeway (so e.g. epte, fad, upatha, bdezolg, pet, puot, etc., etc. are all taken to match because they all have a p, f, or b followed later in the word by a t, d, or th, and no intervening consonant) (3) and up to five senses' leeway (e.g. 'black', 'night', 'dark', 'soot', 'shadow' or any other set of five meanings you consider related; or e.g. 'fingernail', 'finger', 'hand', 'arm', 'claw') then the event probability of such a match is 0.04, and 25 such matches out of a pre-specified 100-word list are required to reach the 95% confidence level on a binary test of two languages. (That's the conventional minimum level for deciding that the number and degree of resemblances are not random. I got the number of 25 from a binomial probability table for an event probability of 0.04 and 100 trials.) For one-consonant sets like Mark's, over half of the 100-word list would have to consist of matches in order to reach 95% confidence. All this is if you prespecify the 100 glosses, prespecify the range of 5 for each, and prespecify the generic consonants. And choose in advance the two languages you want to compare. If you get to look through all words (i.e. entire dictionaries) of any languages, then the required numbers of matches go up. So the probabilistic implication of such patterns is nil, unless you have over 50 of them out of some standard 100-word list. Such are the hazards of doing open-ended searches. (This brings us to Yeniseian and Na-Dene, of which more anon.) The computation of probability in limited searches is described (briefly) in my paper 'The comparative method as heuristic' in M. Durie and M. Ross, eds., The Comparative Method Reviewed (Oxford UP, 1996). I'm working on a fuller explanation. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From johanna at uclink.berkeley.edu Thu Nov 12 01:28:08 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Wed, 11 Nov 1998 20:28:08 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Steven Schaufele writes: ... a Reuters story (posted on www.cnn.com) about an article by Merritt Ruhlen appearing this week in the Proceedings of the National Academy of Sciences, claiming an affiliation between Ket, the sole surviving Yeniseian language, and the Na-Dene family of North America. ... I'm wanting to know, does anybody subscribing to this EBB know anything about this? Is there anything to this proposed affiliation? And if not, is anybody doing anything about clarifying the issue for the general public? - - - - - - - - - - - - - - - - - - - - - - - - - I have seen the article and discussed it with a couple of journalists who contacted me about it. Here is the gist of what I said: The burden of proof is on Ruhlen to demonstrate that his resemblances exceed what is expected by chance. There are well-known tests, standards, criteria, etc. for doing this in linguistics and in statistics. He has not invoked or applied any such criteria, so his findings must be assumed to be chance resemblances until and unless he shows otherwise. Ruhlen lists 36 word sets that he considers resemblant and indicative of genetic relatedness between Yeniseian and Na-Dene. In fact these 36 sets are not numerous enough and not closely enough resemblant to exceed the range of chance. I've worked out the chances of finding words with two similar consonants in the same order, with similar but not necessarily identical meanings, in two languages. Out of a fixed list of 100 meanings chosen in advance (this is analogous to asking "what is the probability that similar forms will mean 'water' in both languages?", and so on for another 99 glosses), this is how many resemblant sets it takes to exceed the range of chance and show that relatedness is likely: 2-consonant words with the very same meaning: 7 2-consonant words with similar meanings (modeling this as a search that allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or whatever; these must also be specified in advance): 25 1-consonant words (or 2-consonant words with one resemblant consonant and one non-resemblant one) with the same meaning: 27 1-consonant words with 5 senses' leeway each: over 50 Ruhlen has: 15 words with two resemblant consonants in the same order 2 words with two resemblant consonants but in different orders or with other consonants intervening ('children' and 'foot') 18 words with one resemblant consonant 1 word with perhaps no resemblant consonant ('name', where the initial glottal stop in the Proto-Yeniseian form is found in no daughter language and must there be phonetic detail of the pronunciation of initial vowels rather than a true structural consonant; and where similarly there doesn't seem to be any evidence for reconstructing a glottal stop as initial, or even as non-initial but the first consonant in the root, for Na-Dene). For simplicity and to give Ruhlen the benefit of the doubt these can be described as 17 two-consonant resemblances and 19 one-consonant resemblances. 11 of Ruhlen's sets have considerable semantic leeway; most have some leeway; for only a few is the sense really the same. So this can be modeled as a search of up to perhaps five senses. In practice this means the researcher finds a word in one of the languages and gets to cast about in the other language looking for a word of similar form in a meaning plausibly connected to that of the first word. Occasionally s/he is lucky and finds a resemblance in the very same sense; usually s/he has to search through a few close senses; and sometimes s/he has to search longer. So even if Ruhlen had started out with a strict list of 100 glosses and a list of 5 senses' worth of leeway for each, he would need about 50 sets of the kind he has. (I'm not calculating this with any precision. I figure he has about half of the 25 2-consonant words he needs, so he also needs to have about half of the 50+ 1-consonant words he needs, and the sum of about half of 27 and about half of over 50 is somewhere around 50.) But he didn't set up a closed list of glosses. His wordlist is not 100 but some open-ended number. At least the total number of words found in the source on Yeniseian he draws on, or maybe the total of those plus any additional meanings found in his various Na-Dene sources. Or maybe it's the average total number of root words per language on earth. I don't know quite how to model this, but he clearly has a wordlist of well over 100. So he needs correspondingly more than the about 50 words he would need if he used a pre-specified closed list of 100. My calculation is for comparisons of actual daughter languages, while Ruhlen compares protolanguages. A protoform has a greater chance of occuring in the protolanguage than in any daughter language. (This is because a protoform occurs in the protolanguage by definition, while for any cognate set it's likely that one or another daughter language will happen to lack a cognate.) Therefore, for comparisons of protolanguages larger numbers of resemblant words will be needed than for daughter languages. For instance, Ruhlen's set 'hunger' has Haida as its only Na-Dene representative, and 'river' has Haida as the only form with a two-consonant resemblance. Some of the sets have no Haida cognate, e.g. 'foot', 'birch bark', 'boat'. From the point of view of methodology, this means he allowed himself to adduce evidence from either Haida or from Eyak-Athabaskan-Tlingit, and the chances of finding a resemblant in one or another family are greater than finding a resemblant in one particular one. Again, this can be modeled as a search in which if the researcher doesn't find a resemblant in Eyak-Athabaskan-Tlingit s/he gets to look for one in Haida. In some sets an ejective consonant in Na-Dene corresponds to a sequence of consonant plus vowel plus glottal stop in Yeniseian (e.g. 'birch bark'); in others (e.g. 'clay') an ejective just corresponds to a Yeniseian consonant with no glottal stop elsewhere in the word; in still others there is a Yeniseian glottal stop but no ejective anywhere in Na-Dene (e.g. 'lake', 'word'). Thus Ruhlen seems to have allowed himself to scan consonants in whatever way maximized resemblances. For all of these reasons, Ruhlen's resemblances are no more numerous and no closer than would be expected to come up by chance. They are exactly the sort of thing one finds when looking through dictionaries and casting about to find resemblances with few constraints on how the search is to be conducted. This is all assuming that the linguistic analysis and the protoforms are sound, matters I cannot judge. Any errors in data, analysis, or reconstruction increase the number of resemblances that must be adduced if they are to be regarded as non-accidental. The paper is not up to date on human genetics and archeology. It mentions only the Greenberg-Turner-Zegura (1986) theory of three American settlements, not mentioning all the work on mitochondrial DNA which has been done since 1987. The claim that "the first migration of the Amerinds [occurred] about 11,000 years ago" is no longer current; the well-dated, well-accepted Monte Verde site in southern Chile at ca. 12,500 years ago shows that the first Americans entered well over 11,000 years ago. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From bwald at HUMnet.UCLA.EDU Thu Nov 12 01:32:44 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Wed, 11 Nov 1998 20:32:44 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- D. Anthony Tschetter-Breed writes: >Forgive me for being anectdotal, but when topic comes up of sound changes >from X to Y and back to X, I always think of the English: >askan > ask > ax (in some dialects, for example Black American English) This example is problematic because OE asc(i)an and acs(i)an have both continued UNINTERRUPTED into current dialects of English, e.g., there are still British dialects that have "aks" (it's common in some working class areas of urban Mersey and Lancashire, for example). It cannot be assumed that Black English "reinvented" ask > aks, rather than received it from relevant British dialects, and has preserved it while most American dialects have not. This is NOT a clear case of reversal of a *sound change* that you want. Meanwhile, it turns out that there indeed have been spontaneous reversals ask > aks. For example, many New York City working and lower class speakers use "aks". It is clear that the metathesis was spontaneous in this case, for the following reasons. In New York City, "short a" (as in "*a*sk") is RAISED to [E] (a mid to high front tense vowel with some inglide for its final off-glide). However, this raising is highly conditioned. It occurs in closed syllables before voiceless fricatives, but NOT before voiceless stops. Thus, "short a" as in "back, bat, tap, tax", etc. remains LOW. The NYC "aks" speakers, however, RAISE "short a" in "ask", i.e., they pronounce it "Eks", and it contrasts with the low vowel is "axe". From this we realise that FIRST, a > E /_s(k) etc (as in "past", "task", "bath", "raft", even "bash" etc) and THEN sk > ks / #E_# (ONLY IN THIS WORD, not, for example, in "task" vs. "tax", etc.) I gave an example of a real reversed sound change, for what it's worth, in my last message. Latin > Romance s > z /V_V and LATER, Central Spanish z > s (but without intervocalic conditioning). That was NOT a matter of -s- surviving from Latin in some dialects of Spanish-to-be and then rediffusing. On the contrary, it was a later sound change that happened to reverse the previous and much older sound change. No mysterious "invisible hand" guided Spanish to reverse s > z BACK to s (and indeed the conditioning is different). It just happened -- because both directions are possible (under certain conditions -- certainly NOT ****z > s /V_V). Essentially, such "reverse" changes have no more intrinsic interest than to serve as a warning that in more time-compressed and less documented cases that *s = s etc does not necessarily mean that there NEVER were any intervening stages between a reconstruction and "unchanged" documented reflex. I don't remember why this issue arose. Maybe in the context of somebody wanting to pose borrowing from some language into another language when the reflex cannot be accounted for in terms of any attested stages of either donor or recipient languages. P.S. One comment on one of Hubey's dyspeptic replies to Trask: Larry wrote: > Mr. Hubey, are you suggesting that one need not spend years studying > Sumerian in order to know Sumerian? Mark replied: >In order to produce a list of cognates all you need is a dictionary. Unh Unh. Wrong! As Larry implied in his message about wrong morpheme cuts, you also need a GRAMMAR (which dictionaries usually don't supply, at least not in sufficient detail). Dictionaries do not generally analyse WORDS into MORPHEMES, esp if the morphemes are derivational affixes of varying degrees of obscurity. Amateurs (and even some professionals) produce a lot of crap by using dictionaries (of languages whose grammars they are ignorant of), and arbitrarily inserting morpheme boundaries to fit preconceived cognates into procrustean molds by getting rid of the "bad" stuff. Such practices conceitedly ape the more judicious use of the technique by more RESPONSIBLE linguists, who nevertheless sometimes make mistakes with that reconstructive technique. At least when responsible linguists indulge in the technique (responsible amateurs included) they continue to take seriously the issue of the historical significance of the stuff they snipped off. From hubeyh at Montclair.edu Thu Nov 12 01:33:22 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Wed, 11 Nov 1998 20:33:22 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), emzirik, etc etc. > > This illustrates the fact that a little knowledge of the languages > involved is never a bad thing. Larry should have been slightly more > precise by indicating that Basque was not borrowed from vanilla > Occitan, but from Gascon/Bearnais. Then it immediately becomes clear > that the word is simply Latin FEMINA, which regularly becomes hemna > (hemm@ ~ hemno) in Gascon (f > h). The ultimate root is IE *dheH1- > "to suck" (*dheH1-mHn-oH2 > femina). The coincidence with PTurkic > *eme "woman" is entirely coincidental. And so is, with even more I don't know that PT 'eme' is 'women'. I only pointed out that 'em' is an old word in Turkic, and might well be one of those that belongs in PW (protoworld) for those who believe in it. > > reason, Sumerian where the element doesn't mean > "woman" at all, but "language, tongue". The word means "thin, > refined" (it used to be thought that was "woman", and one can > still find that in older books, but in fact the Sumerian word was > ). Related words again. This is one of the reasons why the creation of a standard "semantic" space is so important for historical linguistics. All the powerful tools of math are unavailable to linguists because of that, and probably because of this, articles like those Times, etc will keep appearing despite the fact that some hate it. If there is no standard and no impartial referee, that is the way it will always be. If sociologists and psychologists are now fighting over correlation coefficients it is time for linguists to bite the bullet. > > >Do you understand the probabilistic implication of such patterns? > > Yes. In short: "Coincidences happen" and "Garbage in, garbage out". That is another one of those things linguists say. I will refrain from further comments. There are not really too many things to do in the face of uncertainty. Everyone from rocket scientists, engineers, economists, biologists, to computer scientists, workers in speech recognition, synthesis, phonetics and psychologists does the same thing. Except of course linguists. But then again, nobody writes articles in Times that the moon is full of green cheese, or that the world is supported on the back of turtles. > > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From hubeyh at Montclair.edu Thu Nov 12 01:34:26 1998 From: hubeyh at Montclair.edu (H. M. Hubey) Date: Wed, 11 Nov 1998 20:34:26 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > ----------------------------Original message---------------------------- > At the risk of getting caught in the crossfire, I'd like to interject > myself in the debate between Larry Trask and Mark Hubey. I'm in > agreement with about 90% of what LT has to say, and have to > say (with no malice intended) that MH's postings reveal a great > ignorance of the field of linguistics. Nonetheless I do think that MH > I think what is necessary is to get going discussing what is being done and how, and something good might come out of it. > is asking some legitimate questions and deserves a better response from > the profession than the old refrain that the esoteric knowledge is only > revealed to those who have joined the secret brotherhood. > That is the real reason I post what I do. There are computer programs that paint and compose music. Is it really that hard to believe that linguistic reconstruction is no less structured? I bring this up, because for a long time the anti-AI crowd used arguments similar to those offered often on linguistics lists for why AI would be impossible. I will try to reply to the rest of the comments later from home. > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe -- M. Hubey Email: hubeyh at Montclair.edu Backup:hubeyh at alpha.montclair.edu WWW Page: http://www.csam.montclair.edu/Faculty/Hubey.html From delancey at darkwing.uoregon.edu Thu Nov 12 01:34:44 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Wed, 11 Nov 1998 20:34:44 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Ralf-Stefan Georg wrote: > Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially > Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as > I remember this was based on typological resemblances only), Sumerian has Pretty much, yes--their both having tone weighed heavily in the hypothesis (since we now know that in both families tone is a secondary development in the languages and branches where it occurs this loses any force it might ever have had), and Sapir saw a fundamentally monosyllabic structure to both. (That may not be as completely far-fetched a way of thinking about Athabaskan as it sounds). But he did dig up a handful of actual resemblants. Some can be found in Victor Golla's _Sapir-Kroeber Correspondence_, especially letter 332: Tlingit k'a 'surface' : Navaho k'a 'surface' : Tibetan k'a 'surface' Chinese t'an 'charcoal' : Haida s-t'An 'charcoal' OChin ti 'this' : Ath. di 'this' OChin ti 'pheasant' : Ath di 'partridge' Nadene k'u 'hole' : Indo-Chinese [sic] k'u 'hole', with several supporting forms from different N-D and S-T languages. Pretty standard "long-ranger" stuff. In an addendum to that letter he has a very nice word-family comparison involving words having to do with 'tie', 'twist', 'rope', 'snare', 'trap', which is the only evidence I've ever seen presented for Sino-Na-Dene that made me feel even for a minute like I'd like to see more. I had heard that in his notebooks he had more extensive comparisons, so one day while I was at the APS looking for Chinookan stuff I spent an hour looking for ST-ND stuff, but didn't find any. > been compared to pretty much everything, so has Basque, but I think it is > safe to say that a Basque - (unspecified) Caucasian connection has probably > lured more early researchers than anything else (probably originally > instigated by ergativity, which once made up for a quite exotic > look-and-feel of a language - which is hardly the case today). I wish I could remember, or had saved, the citation. I can't even remember the journal--but I did once see an article, by some English individual, from the 1940's or so I think, claiming a Tibetan-Basque relationship, pretty much entirely on the grounds that they were both ergative. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From drc at antnov1.auckland.ac.nz Thu Nov 12 12:46:10 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 12 Nov 1998 07:46:10 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Wed, 11 Nov 1998 12:22:12 -0500 > From: "H.M.Hubey" > Reply-to: hubeyh at montclair.edu > Organization: Montclair State University > To: Ross Clark > Cc: HISTLING at VM.SC.EDU > Subject: Re: rhotacism from Ray Hickey > Ross Clark wrote: > > > > > Aren't morphological paradigms also part of 'regular sound change"? > > > > No, they're not. > > > > But more to the point, contrary to what Mr Hubey seems to be > > suggesting, the little couplet above is not an instance of circular > > reasoning. It's merely the same statement phrased two different ways. > > Or rather, 2. is a statement of a general principle, of which 1. is a specific application. > > > > What Mr Hubey may be trying to articulate is the superficially > > circular-looking: > > > > 1. A,B,C are a language family ===> A,B,C have regular sound > > correspondences. > > > > 2. A,B,C have regular sound correspondences ===> A,B,C are a language > > family. > > > > Perhaps only some people use it circularly, or many linguists use it > circularly, but it is in use. Could you cite an example or two? I don't mean of 1. or 2. in use separately, but of the two used together as a fallacious syllogism. > > > But the attribution of circularity rests on a misreading of the > > relations between the propositions in 1. and 2. as the same. 1. is a > > causal relation -- regular sound correspondences result from the > > definition of a language family, plus the fact that sound change is > > regular. > > But borrowings also create regular sound correspondences. Yes, just as things other than measles can produce spots on the face. We need to take such things into consideration if we want to raise our competence in historical linguistics (or medical diagnosis) from this very rudimentary level. > > 2. is a progression from evidence to inference -- we observe > > regular sound correspondences, from which we conclude these languages > > are a family. (Whether this is an accurate account of what we > > actually do is not the question here.) It's no more circular than: > > > > 1. Patient has measles ===> patient has spots on face. > > 2. Patient has spots on face ===> patient has measles. > > This rests on something different. > > 1. Patient has measles (definition comes from some other place, but may > include spots on the face). These days the defn would come from being > able > to culture the bacteria. Then the spots on the face and measles > correlate. > The time depth is short and one can see a non-measles person get it, get > sick, etc. > > Because of the correlation of measles and spots, 2 then becomes an > implication. > > > But that does not work so in historical linguistics because we never had > a record of any language family (knowing its relatives, etc) but all of > it > rests on a larger theory of which regular sound correspondence must be a > part. Well, we do in fact have records of various language families. What are you trying to say here? > So it is a whole mess of correlations which lead towards that > conclusion. > > More to the point it is based on this reasoning. > > 1. These languages have too many things in common. IOW, there are many > words in > these languages which can be made to look like each other with similar > meanings > and which could not be due to chance. > > 2. If that is not due to chance then either they got these words from > each other > or the words are all descended from a common language. > > 3. We have plenty of evidence (what?) that these languages did not get > these > words from each other. > > 4. Therefore these words in these languages must all come from an > earlier common > source. > > This is how it is supposed to work, but you can see rather easily how > and where > problems crop up, and where arguments occur. > > Isn't this basically right? Yes, problems crop up and arguments occur, to be sure. I recognize 1-4 as a rough outline of the reasoning by which one arrives at a hypothesis of genetic relatedness among languages. Rather than argue about details, I'd like to know where you're going with it. Are we finished with the idea that it's logically circular? Ross Clark From hubeyh at montclair.edu Thu Nov 12 12:47:06 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:06 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- bwald wrote: > Mark replied: > >In order to produce a list of cognates all you need is a dictionary. > > Unh Unh. Wrong! As Larry implied in his message about wrong morpheme > cuts, you also need a GRAMMAR (which dictionaries usually don't supply, at > least not in sufficient detail). Dictionaries do not generally analyse > WORDS into MORPHEMES, esp if the morphemes are derivational affixes of > varying degrees of obscurity. Amateurs (and even some professionals) > produce a lot of crap by using dictionaries (of languages whose grammars > they are ignorant of), and arbitrarily inserting morpheme boundaries to fit > preconceived cognates into procrustean molds by getting rid of the "bad" > stuff. Such practices conceitedly ape the more judicious use of the > technique by more RESPONSIBLE linguists, who nevertheless sometimes make > mistakes with that reconstructive technique. At least when responsible > linguists indulge in the technique (responsible amateurs included) they > continue to take seriously the issue of the historical significance of the > stuff they snipped off. Let me rephrase it. It should be obvious to everyone by now that Starostin is way ahead of the game than the 99% of the subscribers. It should be clear by now that if he has already built up a database of lexicons of various languages with their meanings, writeable in ASCII, Unicode etc, he has singlehandedly done what should have been done by the linguistics community. But that is not all. If there are people who are writing programs to paint (yes, produce art), and compose music, it takes no genius to see that even if Starostin only wrote (or got a student to write) a brute-force, dumb program on a commodity grade PC, he can uncover relationships that many humans cannot do, even if they collaborate. The reason for this will take too long to explain. But given a set of words (and their meanings) even a brute-force program can keep cranking 24 hours a day to produce cognates via regular sound changes, clusters, and things that a typical linguist does not even know exists. It's too bad that the attitude of most linguists is, in fact, the most damaging to themselves and their own professions. But, that is the way evolution is. Short term goals and intuition only go so far. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:47:23 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:23 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > I've worked out the chances of finding words with two similar consonants in > the same order, with similar but not necessarily identical meanings, in two > languages. Out of a fixed list of 100 meanings chosen in advance (this is > analogous to asking "what is the probability that similar forms will mean > 'water' in both languages?", and so on for another 99 glosses), this is how > many resemblant sets it takes to exceed the range of chance and show that > relatedness is likely: > > 2-consonant words with the very same meaning: 7 > 2-consonant words with similar meanings (modeling this as a search that > allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or > whatever; these must also be specified in advance): 25 > 1-consonant words (or 2-consonant words with one resemblant consonant and > one non-resemblant one) with the same meaning: 27 > 1-consonant words with 5 senses' leeway each: over 50 I erased the rest not because it is not important but because I want to ask about this. Is it not true that the most important consideration in probability theory is knowing the sample space? In other words, when "matches" due to chance are being calculated, should not the fact that the two languages have (or seem to have) the same set of phonemes enter into the calculation? In other words, the sample space should consist of the phonemes that the languages could have had (along with the phonemes that they do have) but do not? The calculations should involve conditional probabilities. No? Secondly, I also made some calculations. But mine is not for phonemes and does not take into account phonemes for the reason that they cause more complications, and do not take into account that the same speech space available for humanity is divided up differently and into different number of chunks (phonemes) in different languages. The fact that out of possible M phonemes if languages seem to have a particular set of N phonemes that in itself has to be accounted for. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:47:59 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:47:59 EST Subject: Doing historical linguistics (part 1) Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), > >emzirik, etc etc. > > > >Do you understand the probabilistic implication of such patterns? > > > > I'll take the liberty of commenting, as I understand the probabilistic > implication of such patterns. What are the chances of finding, in any > language, a word consisting of a vowel (any vowel? or just a non-round > vowel?) followed by [m], with or without any following material of any > shape, having one or the other of the set of 7 glosses cited above plus > presumably any other meaning that might have to do with females? No. I was thinking of something more. If words (phonological shapes and their meanings) did not cluster there would be no such thing as etymology. If some word X or its reflexes shows up in a family in random scatter but is strongly represented in another family, what does that imply? Secondly, if words of two different languages were generated independently of each other, the if there are chance occurences it is not only the number of them that matters but also the patterns. But counting up the numbers shows no indication of patterns. Word generation is a Markov process so the tests should be run on some model that purports to be a model of language evolution/development. > That boils down to: what are the chances of finding a sequence Vm(-) in > any of a wide (or open-ended) set of meanings? > > Success is virtually guaranteed, since you get to keep looking until you > find something that fits. Sure, it is too easy. > > I've computed the chance of finding resemblances in which two similar > consonants occur in the same order in words with similar senses. If you're > allowed: > > (1) a couple of distinctive features' leeway in defining generic > consonants (so e.g. p, b, and f are all taken to match) > > (2) and much phonotactic leeway (so e.g. epte, fad, upatha, > bdezolg, pet, puot, etc., etc. are all taken to match because they all have > a p, f, or b followed later in the word by a t, d, or th, and no > intervening consonant) > > (3) and up to five senses' leeway (e.g. 'black', 'night', 'dark', > 'soot', 'shadow' or any other set of five meanings you consider related; or > e.g. 'fingernail', 'finger', 'hand', 'arm', 'claw') > > then the event probability of such a match is 0.04, and 25 such matches out > of a pre-specified 100-word list are required to reach the 95% confidence > level on a binary test of two languages. (That's the conventional minimum > level for deciding that the number and degree of resemblances are not > random. I got the number of 25 from a binomial probability table for an > event probability of 0.04 and 100 trials.) For one-consonant sets like > Mark's, over half of the 100-word list would have to consist of matches in > order to reach 95% confidence. All this is if you prespecify the 100 > glosses, prespecify the range of 5 for each, and prespecify the generic > consonants. And choose in advance the two languages you want to compare. > If you get to look through all words (i.e. entire dictionaries) of any > languages, then the required numbers of matches go up. Some of these computations were also done on Language. And that list is mainly for quantitative approaches. Yes, but this only takes into account number and not pattern. There are lots of ways of testing things. > So the probabilistic implication of such patterns is nil, unless you have > over 50 of them out of some standard 100-word list. Such are the hazards > of doing open-ended searches. (This brings us to Yeniseian and Na-Dene, of > which more anon.) I was talking about the implication of patterns of words in a given language. One of the biggest problems that I encounter all the time is that neutral evidence works for the advantage of the dominant theory. For example, most of the IE words could be due to the substratum which could have been a family. One can always insist that the reason why IE words resemble each other is because they are all left over from a previous language which was spread out over the same region. > The computation of probability in limited searches is described (briefly) > in my paper 'The comparative method as heuristic' in M. Durie and M. Ross, > eds., The Comparative Method Reviewed (Oxford UP, 1996). I'm working on a > fuller explanation. Yes, I read it about a year ago. I think you had an article on comparison of Hittite with others. -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:48:25 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:48:25 EST Subject: probability & sound change (nee rhotacism from R.H.) Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > but I am more sure now that some of what we see is not really > > universal but rather due to some specific local phenomena. > > This is not a matter of opinion, or something that can be decided simply > by thinking about it. It is a matter of evidence, and that is what I > meant by 'empirical'. Labov's conclusion, based on case studies of some > twenty languages, is: ".. there are no directions of vowel shifting that > are forbidden to speakers of human languages, but ... some directions > are taken far more than others" (Principles of Linguistic Change 1994, > p.116). Specifically he proposes that in chain shifts, long vowels > rise, short vowels fall, and back vowels move to the front. I have nothing about empiricism. But these directions must be local because if it were not, we would not have had those which are supposed to disappear. > Labov's sample (mistake in title, should be Prinicples of Linguistic > Change, sorry) is small by the standards of typology, and most of the > languages are European. Further work in the diachronic typology of Even if all the languages of the world were in the sample, it would still be local (temporally) we do not know what it was like before we recorded these languages. > phonological systems is certainly desirable. Nonetheless, as empirical > scientists we have to prefer a theory consistent with some evidence > (directional sound change) over a theory (random sound change) which has > not been shown to apply to any evidence. It is certainly possible that > directionality effects are part of a short term cycle which evens out to > zero in the long run, but I haven't seen any evidence which would > support this view. Yes, I understand all that. But even if someone produced an equation of evolution, it still cannot be universal because it is still local in time. A time evolution cannot be created from a short sample in time. > > There is a similar problem in probability theory, that of stationarity > > of a signal. It's impossible to prove. No matter how many languages > > you > > look into you can only look at a specific time interval. It's hard to > > say if the sound changes are due to universals or to a particular > > combination of sounds and phonotactics. > > This passage is a bit cryptic. I haven't talked about universals, but > only about emprically observed statistical tendencies, and haven't > discussed explanatory factors at all. It would certainly be meaningless > to say that something is "due to" a statistical tendency. The length of > the time interval is only limited by the historical record, but even so, > it isn't necessarily a problem. If directional effects are only part of > a long term cycle X>Y>X (a cycle longer than the historical record), > this should still be demonstrable, based on the fact that not all > languages should be at the same point in the cycle. In some langues X>Y > should be the trend in the attested period, in others Y>X. Breadth of > the sample can substitute for depth. YEs, given a long enough time and a large enough sample we can say that such and such changes have a tendency to occur. But is it a tendency that is unconditioned or is it a tendency that arose from some reasons which are lost in time? To attack the problem a model is needed, a model of change over time. > You're right, of course. Directional trends relate only to what is > probable in sound change, not to what is possible or impossible. On the > other hand it occurs to me that there is one type of sound change which > IS irreversible, that is X>0, as found in mergers, phoneme loss, or > assimilation in clusters. Don't languages also add vowels and consonants? Japanese suffixes vowels to many borrowed words. Ditto for Finnish. Turkish prefixes vowels to initial consonant clusters. The general idea is that the phonotactics decides. That means some phonotactical rule of some language (even the same language at a different time) can create the conditions which can result in the addition of a vowel. But then again, a model is needed of language evolution. Lass talks about a lot of this in his book. There are things we can think about just from the way things behave in nature. Gell-Mann discusses this in the Santa Fe book. > Directionality does not mean teleology. To say that sound change moves > in a non-random direction is not to say that it is going anywhere in > particular. One might propose (although I didn't) that over time Maybe there is. Dynamical systems have behaviors which can be used as models. For example, the concept of stability. Languages might change very slowly after they have essentially reached stability. That could explain why some languages seem to change fast and others slowly. > phoneme inventories get smaller, or that 'marked' phonemes tend to be > lost in favor of unmarked. I have yet to see a teleological proposal of > this type which can survive the test of the evidence. The directionality > trends discovered by Labov appear ultimately due to physiological > constraints both articulatory and auditory. If that can be shown to follow from physical and auditory causes so much the better. In other words the direction > of sound change is constrained by the present not by the future. I > suppose this is what you mean by talking about local versus universal > (i.e. short term vs. long term). There are no long term directional > trends (as far as the evidence now goes). But we cannot conclude from > this that the sum of short term directional trends adds up to zero or no > directionality or random directionality. No, local means restricted to a small region of space or time. If humans have been speaking for 500,000 years, a 5,000 year span is very small. > The prequisite to a realistic probalistic model of language change > is a systematic research program in diachronic typology. We have to > establish the probable direction of change on the basis of statistical > data from observed changes-- not on the basis of a priori reasoning. Of, course. But even statistics is built on models. > suspect the ultimate model will look something like those typhoon maps > we see here (don't know if you have them in other countries): There is a > circle showing where the storm is and from the outer edges of the circle > there is a wedge shaped projection showing the area into which the storm > is likely to move. At the big end of the wedge there is another bigger > circle showing where the storm is likely to be at the end of a certain > period (24 hrs, e.g.). From this big circle another wedge is projected, > with a bigger mouth showing where the storm should go in the next > period, and so on. Well, that is dynamics already. And there are models (equations) that will do those things. The zeroth order approximation is the asexual family tree model. Something beyond this seems desirable. What kind of a model should it be? > > > The same applies to linguistics changes.There are many scales at which > > changes occur, and if we mix up these levels we create opposing ideas. > > A priori we don't know what models or scales apply to linguistic > changes. We can't know until we have tried to apply them, that is test > them against the data. Yes, and if the present model says that the subtratum has no effect then there's no way to test it against data. The reason is that the neutral data is interpreted to always confirm the status quo. It is circular. Substratum has no effect. The change was internal. Proof: no words from the substratum. If we don't know the substratum how can we tell if there was no effect. Similarly, how do we reach the conclusion that the subtratum had no effect, if we do not consider any change to be due to the substratum but always internal? > +++++++++++++++++++++++++++++++++++++++++++++++++ > Robert R. Ratcliffe > Senior Lecturer, Arabic and Linguistics, > Dept. of Linguistics and Information Science > Tokyo University of Foreign Studies > Nishigahara 4-51-21, Kita-ku > Tokyo 114 Japan -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:48:55 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:48:55 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > Could you cite an example or two? I don't mean of 1. or 2. in use > separately, but of the two used together as a fallacious syllogism. There are too many cases. I don't feel like embarrasing people and making it worse. > > But borrowings also create regular sound correspondences. > > Yes, just as things other than measles can produce spots on the face. > We need to take such things into consideration if we want to raise > our competence in historical linguistics (or medical diagnosis) from > this very rudimentary level. But that is not all. Human family members resemble each other. That does not mean that unrelated people cannot resemble each other. And despite the fact that we know both we still consider two people who resemble each other to be related unless there's proof to the contrary. We arrive at this through experience. We see families (which we can confirm) resemble each other and therefore create a general inductive rule. For measles, doctors know very well what healty people look like. How many language families has any human experienced? I do not mean the purported/alleged language families. If a human could be created who could live 100,000 years or more and if we can send him to the past to learn dozens of languages, then he would have 'experienced' language families like human families and the way doctors (and others) have the experience of knowing what measles does. No such thing can be done in linguistics so all of it is based on analogy to models from the rest of the world, such as Linnaean trees, etc. That is coupled with some intuitive calculation of whether the resemblence is due to chance. > Well, we do in fact have records of various language families. What > are you trying to say here? There we go again. Do we know these families like we know human families or is this based on some calculation that the occurences cannot be due to chance? > > > 1. These languages have too many things in common. IOW, there are many > > words in > > these languages which can be made to look like each other with similar > > meanings > > and which could not be due to chance. > > > > 2. If that is not due to chance then either they got these words from > > each other > > or the words are all descended from a common language. > > > > 3. We have plenty of evidence (what?) that these languages did not get > > these > > words from each other. > > > > 4. Therefore these words in these languages must all come from an > > earlier common > > source. > Yes, problems crop up and arguments occur, to be sure. > > I recognize 1-4 as a rough outline of the reasoning by which one > arrives at a hypothesis of genetic relatedness among languages. > Rather than argue about details, I'd like to know where you're going > with it. Are we finished with the idea that it's logically circular? But it is not finished. The key here is that we have to know what is due to chance and what is not. Otherwise we can be creating an argument like this: Well, this mathematical method says that X and Y are related but I know that they are not, so the mathematical method is wrong. There are people (yes, real people, and linguists too) who do this. In fact one of the superstarts of sci.lang and linguistics actually argued exactly like this in email to me. > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hubeyh at montclair.edu Thu Nov 12 12:49:23 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Thu, 12 Nov 1998 07:49:23 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Robert R. Ratcliffe wrote: > > > My view: > > That "reconstruction" is (or even could be) a matter of algorithm, or > rule is a widely held misunderstanding outside the field, no doubt > supported by a misinterpretation of the technical term "reconstruction" > in its ordinary sense. We cannot literally "reconstruct" anything. A > "reconstruction" is a hypothesis, neither more nor less. Like any > Of course, there is no proof even in physics. I state this regularly. hypothesis in any field of science we arrive at it by guesswork, by > intuition, by imagination, by accident. There is no path of deductive > reasoning, no "discovery procedure", no algorithm which leads from the > data to the hypothesis. The rigor in historical linguistics as in the > natural sciences is not in the way in which hypotheses are reached but > in the way in which the hypotheses are submitted to the test of the > data. A good reconstruction should be testable in principle-- it should > be a specific hypothetical prediction about the way a particular > language was spoken by a particular group of people at a particular > point in time. A reconstruction is virtually never (directly) testable > in practice-- it is so only on those extremely rare occasions when new > texts of ancient languages are unearthed. That is lots of evidence for correctness but not proof. If some new language that was unearthed was remarkably like someone's prediction, that is cause for celebration. How do we test it then?-- By > implication. Each reconstruction (of a proto-phoneme for example) has > implications for the whole system of the proto language (the whole > phonological system, eg), for the development path leading from the > proto-language to the attested languages (the sequence of sound changes, > eg), and for the forms of the reflexes in the descendant languages. > Only the last is directly observable, of course, and only this real data > can be used to rule out a proposed reconstruction absolutely. But here is where the iteration comes in. The first attempt at reconstruction of a protolanguage *X will be based on N languages. If we add the (N+1)st language then *X might have to be changed. We might find another language y to add to the family. How many correspondences do we need? Even worse, if the similarity of the language to other languages is not considered, it will be added to the most established, largest family, and it will continue to snowball. If you want to see if A is more like B or like C you have to have both the compare. This is the "forced binary discrimination" test which is often used (in phonology and phonetics in lingistics). But then those languages that got writing first keep piling up everything because of these factors. > proposed development path implied by the reconstruction cannot be tested > as right or wrong but only as plausible or implausible based on the > statistical frequency of observed patterns of change. To me that says that we need numbers so we can compare languages to others. That means standardization. > implications which a reconstruction has for the whole system can only be > tested under the assumption of the "uniformitarian principle" - the > assumption that prehistoric languages were not fundamentally different > in kind from attested languages and hence should not show structural > anomalies of a type not found in attested languages (languages shouldn't > be reconstructed with no vowels, or all verbs, for example). > Probability and statistics is involved in all of this, and much could be > gained from making it more explicit. Much could also gained if > historical linguists made an effort to base assumptions about plausible > directions of changes on an explicit, accesible, statistically > analyzable body of evidence of changes, rather than on implicit personal > knowledge. Amen! > -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mccay at redestb.es Thu Nov 12 12:50:05 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 12 Nov 1998 07:50:05 EST Subject: s > r (Iberian) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: >The phenomenon does extend to "apitxat" Valencian Catalan, i.e. the >central Valencian dialects of Valencia city and province (not N and S >Valencian of Castello' and Alacant provinces). I'm glad you have introduced some information about "apitxat". As it happens, my companion is a speaker of apitxat, but a non-native one (she is Basque but grew up in the city of Valencia), and she never gives me clearcut data because she is insufficiently confident in her competence in it, and I don't know any other apitxat speakers!!! Also, I get the idea there is some sort of stigma attached to apitxat speech within Catalan, which could be producing unwanted interference. So anyway, we add one dialect of Catalan to the list of varieties without /z/ in the peninsula (the others being the whole of Castilian and (virtually) all of Galician). (And Asturian?) Now: are there any interesting sibilant-transforming phenomena in apitxat, I wonder? (1) Does /s/ voice before voiced consonants? (In Castilian and Galician this is the only context in which we find [z]. My money is on the same thing happening in apitxat.) (2) Does /s/ rhotacize? Does it aspirate? Does it drop? (I'm guessing no to all three...) In fact, the only phenomenon of this type that I have heard of anywhere in Catalan is rhotacism in varieties of Mallorcan (as pointed out by Carrasquer earlier). Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From wbehr at rullet.leidenuniv.nl Thu Nov 12 12:50:35 1998 From: wbehr at rullet.leidenuniv.nl (WB (in Frankfurt today)) Date: Thu, 12 Nov 1998 07:50:35 EST Subject: Ket-Na-Dene affiliation? Message-ID: ----------------------------Original message---------------------------- Just a small note of caution, re: >The Yenisseyan-Na-Dene affiliation is subscribed to mainly by Sergei >Starostin and his Moscow-based school of Nostraticists. (Stefan Georg) My understanding from personal conversations with Starostin last year is that he himself remains non-committal as far as the proposed Ketic-Na-Dene affiliation is concerned, and that he was not particularly amused when Ruhlen arbitrarily added Na-Dene comparanda to his translation of Starostin's original 1982 _Ketskij Sbornik_ article in _On the Origin of Languages_. Ditto for Sumerian, Kusunda and Burushaski, which are all ad- ditions by Bengtson, Blazhek, Ruhlen, Boisson (?) et al., not by Starostin, as far as I can see. It is true though, that Starostin's collaborator on NC, S.L. Nikolaev subscribes to the existence of "Sino-Caucasian Languages in America" (see his articles in Shevoroshkin ed. 1989, 1991) but that doesn't necessarily mean, of course, that _all_ "long-rangers" at Humanities U. in Moscow would agree upon that. For detailed background information on Sapir's view of the Na-Dene -- ST relationship, cf. Alan S. Kaye, "Distant genetic relation- ship and Edward Sapir", _Semiotica_ 91 (3-4): 273-300 (see also J.D. Bengtson, "Edward Sapir and the 'Sino-Dene' Hypothesis",_Anthro- pological Science_ 102 (1994): 207-230). Judging from Kaye's careful account, it would seem that Sapir's initial enthusiasm for the Sino-Athapaskan connection in the early 20ies gradually bleached, possibly under the influence of his student Li Fang-kuei, who was certainly more qualified than any other scholar around to judge _both_ sides of the comparison. (Indeed, one letter to Sapir suggests that the main motivation for accepting FKL as his student was precisely Sapir's interest in verifying the ND-ST hypothesis!). A decade later, in a letter to Nick Bodman (July 21, 1933), Sapir had already become slightly more cautious, saying: "It is true that some years ago I announced that the Nadene group of American Indian languages might be remotely related to Sinitic (Indo-Chinese) group of languages. I still believe this is true but have so far not prepared my notes for publication". Notice however, that Sapir's original motivation for the assumption of a genetic relationship was _not_ exclusively lexical/typlogical but _morphological_. In a letter to Berthold Laufer of 1921 he writes "I have at last found what I had been looking for some time now, namely that in Nadene as in Indo-Chinese there is an alternation of unaspirated and aspirated consonants, the latter of which have causative value." It would thus be interesting to hear more about the diachronic sources of this particular causative formation from a specialist in "Na-Dene" (as well as the current state of opinions about a "Na-Dene" +/- Tlingit itself etc.), especially since the Sino-Tibetan distinction is now com- monly accepted to go back to *s-prefixation of the root. In any case, I really wish someone could edit Sapir's manuscript "Atha- baskan Dictionary, part 2, Sino-Nadene" (preserved in the American Philosophical Sociery library), since his letters to Berthold Laufer, most of which have been published by Hartmut Walravens in the meantime, do not go beyond some rather superficial remarks. Finally I wonder if someone has ever read Laurence Farget's M.A. thesis _Na-Dene and Sino- Tibetan: Historical linguistics and new data towards establishing genetic relationship_ (Lyon 1986), or would know, how to get a copy of it. Best wishes, Wolfgang Behr Kaye quotes It forms part of a >much larger grouping, called Dene-Caucasian, comprising no less than the >following languages and families: Basque, Iberian (yes, Iberian, though >nobody knows anything about Iberian !), Burushaski, Yenisseyan, >"North-Caucasian" (comprising NW- and NE-Caucasian which are not generally >grouped together by Caucasianists), Hurro-Urartean (lumped with N-Cauc.), >Kusunda (an extinct language of Nepal, on which little is known; Bengtson, >another supporter of this grouping, recently informed me that Kusunda has >been taken out of the grouping; I'm inclined to call this a step in the >right direction ;-) and Na-Dene (I'm currently unsure whether further North >American lgs. have made it into the family yet), and, sorry I forgot, the >whole of Sino-Tibetan, and of course Sumerian. >It is clear that one of the goals of this grouping is to hoover up most >languages of the Old World, which are currently thought to be isolates. >Larry Trask has shown on numerous occasions that (the supporters of) this >theory treat(s) Basque data in a less than competent manner; the same can >be said about Yenisseyan and much of the Tibeto-Burman (part of >Sino-Tibetan) data I've seen in connection with this theory. >Parts of this giganto-macro-grouping have some history, though: a >Yenisseyan-Sino-Tibetan connection has been en vogue in the earlier days of >Yenisseyology with investigators like Donner, Bouda and the outsider >Ramstedt, comparing Na-Dene languages with Sino-Tibetan (especially >Tibetan) has a Sapirian pedigree (I await to stand corrected, but as far as >I remember this was based on typological resemblances only), Sumerian has >been compared to pretty much everything, so has Basque, but I think it is >safe to say that a Basque - (unspecified) Caucasian connection has probably >lured more early researchers than anything else (probably originally >instigated by ergativity, which once made up for a quite exotic >look-and-feel of a language - which is hardly the case today). >Readers may already have inferred from my slightly ironic tone that I'm >personally disinclined to buy much of this (I have working experience with >Tibeto-Burman, Yenisseyan, NW-Caucasian, NE-Caucasian and a bit of >Burushaski). > >St.G. > >Stefan Georg >Heerstrasse 7 >D-53111 Bonn >FRG >+49-228-69-13-32 > From mccay at redestb.es Thu Nov 12 12:50:55 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 12 Nov 1998 07:50:55 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- B. Wald says: >It just happened -- because both directions >are possible (under certain conditions -- certainly NOT ****z > s /V_V). That provoked me to try and think of counterexamples to Benji's latter assertion here. I couldn't come up with an example of z > s /V_V, but I did find one of +voice > -voice /V_V. Furthermore, if I am right in thinking that the "devoicing" is an innovation (and thus did happen) rather than a survival (i.e. unless there was really no change; see the end), then it provides as good an example of X>Y>X as what Benji called "Central Spanish" s > z > s. (Okay, I'm going to be a pain and say it: just what is "Central Spanish" supposed to mean? The phenomenon in question includes the WHOLE of Castilian (i.e. Spanish) in its territorial extension. Are you implicitly calling Portuguese, Catalan and perhaps Basque "peripheral Spanish"?? Sorry, I couldn't hold that one back. May I suggest that perhaps Benji meant something like "Central Ibero-Romance"?) What I have in mind is a dialectal phenomenon within modern Welsh. First the relevant background information: In modern Welsh word stress is normally on the penultimate syllable, and the phenomenon I will describe occurs in the consonant following the stress when another syllable follows (as it usually does except in stressed monosyllables, of course), i.e. normally at the boundary before the last two syllables of a word, for example where the /d/ is in: /'kadair/, /'kader/ or /'kadar/, orthographically , 'chair' < Latin "cat(h)edra" etc. In this post-stress intervocalic position, in standard Welsh, the consonant may be phonetically simple, as in standardly pronounced ['ka:der] (or in slow speech, ['ka:dajr]) or geminate, as in: /'etto/ 'yet, again' The distribution of simple and geminate consonants in this environment is largely predictable (e.g. "voiced" stops are simple, "voiceless" stops are geminate), as is the distribution of short and long vowels in the stressed syllable (long preceding simple consonants, short preceding geminate ones), for which reason the modern orthography reflects neither consonant length nor vowel quantity in these cases. But there are some minimal contrasts too, e.g. 'sing' versus 'whiten'. (Caveat: I have put quotes around "voiced" and "voiceless" because phonetically all stops in Welsh tend to be voiceless (or at least voicing is not critical) and the contrast is realized principally in terms of tenseness and aspiration. To simplify the exposition I shall henceforth largely ignore that fact in the transcriptions and terminology used.) That is in STANDARD Welsh. Now in Northern spoken Welsh, this system has been altered, in that nearly all consonants in the intervocalic post-stress position are pronounced geminate (and correspondingly, all stressed vowels followed by a consonant in non-final syllables are short - but I'm going to focus on the consonants here). So corresponding to standard ['ka:dajr] we will find ['kaddar] in northern Welsh, while ['etto] shows no change since the consonant is already geminate. Long vowels in northern Welsh are found mainly in monosyllabic words (/ka:n/ 'song') and in penultimate syllables when in hiatus with the following vowel (/'diod/ [di:od] 'drink'); otherwise all vowels are short. In northern Welsh, then, and are both ['kannI]. Of these two systems, the standard and the northern, the standard one must represent a diachronically prior stage, for a number of reasons that I won't go into, but some of which are already fairly obvious. Now: apart from this gemination and shortening, but possibly related to it, we find some varieties of Welsh in which, corresponding to consonants like the /d/ in , if the consonants in question are "voiced" stops in standard Welsh, we find "voiceless" stops instead: ['katar] etc. Assuming once again that the standard form is diachronically prior, we then have d > t (or possibly d > dd > t) in post-stress intervocalic position, and similarly b > p and g > k. Irritatingly, this phenomenon has been noted in the south, in an area not contiguous to the northern "geminating" varieties. By the way, we might assume that the same varieties would have had z > s intervocalically, if only Welsh possessed a /z/ phoneme, which it doesn't; such is life! Then again, maybe z > s wouldn't have happened, since the voiced fricatives /v/ and /D/ do NOT undergo devoicing in this context. Tant pis! (Actually, these consonants also seem to resist gemination in the "geminating" northern dialects, even though the preceding stressed vowel is STILL short. That's the reason why I said "nearly all" three paragraphs back.) As may have been noticed, however, the standard Welsh /d/ which in some dialects has become /t/ (by the account I just gave) itself derives ultimately from an earlier /t/. In the evolution from proto-Celtic (or Latin, in the case of loanwords) to modern Welsh, intervocalic voiceless stops were regularly voiced, much as in western Romance languages. The modern intervocalic voiceless stops come from original geminate stops - which explains their gemination in the modern language, as well as the lack of it in the voiced counterparts in standard Welsh. So in an earlier stage of the development of Welsh we have approximately: VttV > VttV VtV > VdV and similarly for /p/ and /k/. In northern Welsh ['kaddar] we then have a further step: 'VCV > 'VCCV (hence 'VdV > 'VddV) while in the varieties that now pronounce ['katar] we presumably have a later reversal of t > d: 'VdV > 'VtV Unless, of course, 'VtV in these varieties is actually a SURVIVAL of Celtic/Latin VtV. Would anyone more expert than myself like to take up the discussion from here? NOTE: For anyone who wants to know the detailed diachronic background of Welsh phonology (in general), I recommend the classic historical grammar by Morris-Jones. From jhewson at morgan.ucs.mun.ca Thu Nov 12 21:20:10 1998 From: jhewson at morgan.ucs.mun.ca (John Hewson) Date: Thu, 12 Nov 1998 16:20:10 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Tue, 10 Nov 1998, Larry Trask wrote: > Oh, there are certainly rules, but there are no algorithms. For > example, given the observed variant forms of the Basque word for `ear', > there is only one reconstruction that obeys all the rules. > > > My original question is/was why there is no algorithm for producing > > protoforms. Either what you practice is a science or it is not. If > > it is magic you don't have to explain it. If it is science it should > > be possible to see it in writing in some book. Actually there is an algorithm that a Systems Analyst and I developed to produce my _Computer-generated Dictionary of Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be stated in a single sentence, as on p.iv of the above: From the data of the daughter languages generate all possible protoforms, then sort alphabetically, and examine all sets of identical protoforms collocated by the sort. The methodology has now been streamlined by John Lowe and Martine Mazaudon, who reported on it in an article in the Bulletin de la Societe Linguistique de Paris. Their "Reconstruction Engine" even reconstructs tones, and can be adapted to the use of any language family. The existence of these algorithms is a massive demonstration of the regularity of sound change. But the final product still has to be worked on by the linguist, given the nature of human languages. There may be more than one possible reconstruction thrown up by the algorithm, where perhaps one or more definitive item of evidence is lacking. If this further evidence can not be obtained by subsequent research, the reconstruction must remain tentative (the same is true for all reconstructions, of course, machine-generated or not). John Hewson, FRSC tel: (709)737-8131 Henrietta Harvey Professor of Linguistics fax: (709)737-4000 Memorial University of Newfoundland St. John's NF, CANADA A1B 3X9 From delancey at darkwing.uoregon.edu Thu Nov 12 21:19:33 1998 From: delancey at darkwing.uoregon.edu (Scott DeLancey) Date: Thu, 12 Nov 1998 16:19:33 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, WB (in Frankfurt today) wrote: > do not go beyond some rather superficial remarks. Finally I wonder if > someone has ever read Laurence Farget's M.A. thesis _Na-Dene and Sino- > Tibetan: Historical linguistics and new data towards establishing genetic > relationship_ (Lyon 1986), or would know, how to get a copy of it. I have a 13-page "abrege" of the thesis ("Na-Dene et sino-tibetain: synthese et nouvelles donnees pour une eventualle parente"), but it is just a brief literature review, with no new evidence presented. I too would like to see the complete thesis. For some reason she seems to attach some importance to Benedict's reconstruction of a two-tone system for Proto-Sino-Tibetan, which is not a very popular idea within the field--I think most Sino-Tibetanists currently reconstruct PST without tone. Scott DeLancey Department of Linguistics University of Oregon Eugene, OR 97403, USA delancey at darkwing.uoregon.edu http://www.uoregon.edu/~delancey/prohp.html From larryt at cogs.susx.ac.uk Thu Nov 12 21:19:10 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:19:10 EST Subject: Hawaiian meli In-Reply-To: <3648FDB6.2C7D5309@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, H.M.Hubey wrote: > How are we to take into account all of those words that are said to > be cognate in IE when it is quite possible that many (most) might be > left over a substratum that was living in that neighborhood for many > thousands of years and had stabilized so that changes were taking > place very slowly? First of all, languages do not "stabilize". Every living language is always changing, and the only language that doesn't change is a dead one. Now, if some substrate words had managed to enter the ancestor of PIE, that wouldn't matter. A PIE word is a PIE word, regardless of its ultimate origin, and all we are usually trying to do is to reconstruct PIE, and not to go further back. Only if we *do* want to go further back do we need to worry about loan words into Pre-PIE. As for loan words into an already existing PIE, well, that's a problem familiar to all historical linguists. We always have to worry about the possibility that some words in any given language might be loan words. But PIE happens to be a more than averagely convenient language in this respect. PIE words tend rather strongly to adhere to certain patterns of formation, perhaps most typically Root-Formative-Suffix, with all three elements recurring in other words. There is no reason to expect a loan into PIE to conform to such patterns, and hence we perhaps have a better than average chance of spotting loans into PIE. As for the suggestion that PIE consisted mostly of elements from another language, well, that's rather reminiscent of the famous claim that the Iliad and the Odyssey were not written by Homer, but by another poet of the same name. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From sally at isp.pitt.edu Thu Nov 12 21:18:48 1998 From: sally at isp.pitt.edu (Sarah G. Thomason) Date: Thu, 12 Nov 1998 16:18:48 EST Subject: Ket-Na-Dene affiliation? In-Reply-To: Your message of "Wed, 11 Nov 1998 13:20:34 EST." Message-ID: ----------------------------Original message---------------------------- One additional problem with the Na-Dene comparison, of course, is that the inclusion of Haida in Na-Dene is (to put it mildly) not universally accepted by specialists. See, for instance, Robert Levine's article in IJAL on the subject some years back -- an article that was attacked by Greenberg in a whole chapter, in his 1987 book on classification of Native American languages. But Levine's article is good nevertheless, and Greenberg's arguments against Levine aren't impressive. -- Sally From larryt at cogs.susx.ac.uk Thu Nov 12 21:18:21 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:18:21 EST Subject: rhotacism from Ray Hickey In-Reply-To: <364969B6.511AF413@eucmos.sim.ucm.es> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Javier Martinez wrote: [LT] > > But some sound changes are quite irreversible. Consider loss. In the > > ancestor of Greek, prevocalic */s/ was lenited to /h/, and the resulting > > /h/ was later lost. I predict confidently that the Greeks will never > > reverse this change by re-introducing those long-gone /s/s, > > yes. as a product of analogy, see the -s- futures etc. Yes, agreed, except that specialists do not seem to be sure whether /s/ was first lost from futures like (from `loosen') and then restored by analogy, or whether it was never lost in the first place because of paradigmatic pressure to retain it. As usual, we cannot tell without textual evidence whether we are looking at an instance of analogical restoration or at an instance of analogical preservation. Analogical preservation is quite well attested. For example, some of the Finnic languages have lost word-final /n/ in all cases *except* where that /n/ was the sole marker of a grammatical category, in which case it has been retained. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Thu Nov 12 21:18:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 16:18:09 EST Subject: Doing historical linguistics (part 1) In-Reply-To: <36557252.64845784@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, Miguel Carrasquer Vidal wrote: OK. I'm going to try to catch up a little here. > "H.M.Hubey" wrote: > >Larry Trask wrote: > >> `female' (which itself is borrowed from Occitan), plus <-kume> > >And where is that borrowed from? It is not borrowed at all, but native. Basque means `child, offspring'. This is a native word, reconstructed as *, and apparently attested in the Aquitanian ancestor of Basque as both OMBE- and VMME. This word often occurs as the second element in compounds, in which position it occurs as <-kume>, with an initial /k/. Nobody knows how this /k/ arises, but it is not unique to this morpheme. Very commonly, a vowel-initial Basque word acquires an initial /k/ or /t/ when it serves as the second element in a compound. The origin of this plosive is only partly understood, but in any case it is normal in Basque, and the /k/ in requires no additional explanation. This <-kume> is frequent: `cat', `kitten' `sheep', `lamb' `pigeon', `baby pigeon' And so on, for many examples. The other phonological developments seen here are all regular. The word `woman' is seemingly of rather recent origin in Basque: it has only a single attestation before the 18th century, since when it has become the usual word for `woman', except in the Salazarese dialect, in which it means `girl', which we believe to have been its original meaning, since the etymology is * `female offspring'. The earlier word for `woman' was , which today has been specialized to `wife', except in Salazrese, where it still means `woman'. It will surprise no one to learn that both and are compounds built upon the word `female', whose regular combining form is (look at two of the examples above). We suspect an etymology * `young female'. And, as Miguel C V has pointed out, this is borrowed from Gascon , itself from Latin . Basque contains many loans from Gascon. The /mn/ cluster has been intolerable in Basque at all periods, and, in loan words containing it, it is always reduced to either /m/ or /n/. > >em (to suck), am (cunt in Turkish), amma (mother), amcik (pussy), > >emesal (female speech in Sumerian), emcek (breasts, udder), meme > >(breast), emzirik, etc etc. The Sumerian word is not relevant, and it has been effectively disposed of by Miguel. Turkish does not mean `mother'; it means `but'. The Turkish word for `mother' is in Anatolian Turkish but in standard Istanbul Turkish, this last apparently being an expressive variant of . Turkish `nipple' is not available for comparison. One of the best-known facts about Turkish is that native Turkish lexical items do not begin with /m/ unless they are imitative words or nursery words. In all likelihood, we are looking at a nursery word here, and nursery words cannot be cited as comparanda, because they are so often created independently. The stem `suck' is the source of the derivatives `nipple', `nipple', and `suckle'; these can only be counted as one word. Finally, is merely a diminutive of `vulva', containing the usual Turkish diminutive suffix <-cik>. Hence all we have here for Turkish is a verb-stem `suck' and a noun `vulva'. And nothing whatever can be concluded from this. It's every bit as impressive as English `ear' and `hear'. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 21:14:45 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 16:14:45 EST Subject: intervocalic DEvoicing can also happen / X > Y > X In-Reply-To: <3.0.3.32.19981112103615.006eaf48@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >As may have been noticed, however, the standard Welsh /d/ which in some >dialects has become /t/ (by the account I just gave) itself derives >ultimately from an earlier /t/. In the evolution from proto-Celtic (or >Latin, in the case of loanwords) to modern Welsh, intervocalic voiceless >stops were regularly voiced, much as in western Romance languages. The >modern intervocalic voiceless stops come from original geminate stops - >which explains their gemination in the modern language, as well as the lack >of it in the voiced counterparts in standard Welsh. So in an earlier stage >of the development of Welsh we have approximately: > >VttV > VttV >VtV > VdV > >and similarly for /p/ and /k/. In northern Welsh ['kaddar] we then have a >further step: > >'VCV > 'VCCV (hence 'VdV > 'VddV) > >while in the varieties that now pronounce ['katar] we presumably have a >later reversal of t > d: > >'VdV > 'VtV > >Unless, of course, 'VtV in these varieties is actually a SURVIVAL of >Celtic/Latin VtV. I'm far from an expert on Welsh, and I'm not even sure I have the basic facts right. What I gather from the above and from Paul Russel's "An introduction to the Celtic languages" is that: -d- > -D- Adam > Addaf -t- > -d- carita:tem > cardod -tt- > -T- cattus > cath Apparently, new geminates -tt- arose from syncope, giving -d- (cadair) and -tt- (eto /etto/) as the only allowed intervocalic stops. Now the lenitions of -d- to /D/ and -t- to /d/ seem to be the oldest. It is also reasonable to suppose that after they had taken place, geminate /tt/ in /kattu(s)/ was simplified to /katu(s)/ (or directly spirantized to /kaTu(s)/), and did in any case not merge with the new /tt/. If -t- in the /katar/ varieties is a survival of Celtic/Latin -t-, then that must mean that *tt did not pass through a stage *t, but spirantized directly, at least in those varieties: d > D t > t tt > T -- > tt As opposed to the rest of Welsh: SW NW d > D D D t > d d dd tt > t > T T T -- > tt tt tt But that would put the /katar/ varieties of Welsh above Cornish and Breton in the branching tree, which doesn't seem very likely. So it must be: SW NW tW d > D D D D t > d d dd t tt > t > T T T T -- > tt tt tt tt What's missing in the above is a variety of Welsh that has eliminated geminates: D, d, T, t. One would expect one. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From johanna at uclink.berkeley.edu Thu Nov 12 21:14:14 1998 From: johanna at uclink.berkeley.edu (Johanna Nichols) Date: Thu, 12 Nov 1998 16:14:14 EST Subject: Yeniseian and Na-Dene In-Reply-To: <364A801F.95E8163A@montclair.edu> Message-ID: ----------------------------Original message---------------------------- Mark Hubey writes: >Johanna Nichols wrote: >> >> I've worked out the chances of finding words with two similar consonants in >> the same order, with similar but not necessarily identical meanings, in two >> languages. Out of a fixed list of 100 meanings chosen in advance (this is >> analogous to asking "what is the probability that similar forms will mean >> 'water' in both languages?", and so on for another 99 glosses), this is how >> many resemblant sets it takes to exceed the range of chance and show that >> relatedness is likely: >> >> 2-consonant words with the very same meaning: 7 >> 2-consonant words with similar meanings (modeling this as a search that >> allows up to 5 senses' leeway, e.g. for 'fly' also 'flee', 'wing', or >> whatever; these must also be specified in advance): 25 >> 1-consonant words (or 2-consonant words with one resemblant consonant and >> one non-resemblant one) with the same meaning: 27 >> 1-consonant words with 5 senses' leeway each: over 50 > > >I erased the rest not because it is not important but because I want to >ask about this. > >Is it not true that the most important consideration in probability >theory is knowing the sample space? > >In other words, when "matches" due to chance are being calculated, >should not the fact that the two languages have (or seem to have) >the same set of phonemes enter into the calculation? In other words, >the sample space should consist of the phonemes that the languages >could have had (along with the phonemes that they do have) but do not? > >The calculations should involve conditional probabilities. No? > >Secondly, I also made some calculations. But mine is not for phonemes >and does not take into account phonemes for the reason that they cause >more complications, and do not take into account that the same speech >space available for humanity is divided up differently and into >different >number of chunks (phonemes) in different languages. The fact that out of >possible M phonemes if languages seem to have a particular set of N >phonemes that in itself has to be accounted for. > I have two ways of computing the probability of a generic consonant. (1) Languages (as sampled in my database) have an average of about 20 consonant phonemes. One of them has, on average, 0.05 chance of occurring in a randomly chosen position in a randomly chosen form. This is the probability of a specific consonant. Allowing three trials (allowing a search through two to three distinctive features' space, or about three phones) yields a probability of 0.143. (2) Whatever the consonant inventories of the languages under comparison, divide each of them into 7 phonetically coherent spaces. (Variant: divide them into 6 coherent spaces, and count lack of any consonant -- e.g. initial V rather than C -- as a seventh possibility. This makes it possible to accommodate Rotokas, with its 6-consonant system.) This way too we get a probability of 0.143 (1/7 = 0.143). Neither of these procedures guarantees fair coverage of vastly different frequencies of different consonants, language-specific or family-specific preferences of different consonants or consonant classes for different phonotactic positions, and the like. I hope that some of these differences get ironed out by putting consonants together in phonetic groupings. Still, the metric is only approximate. It enables us to point out that 36 resemblant sets, half of them with only one resemblant consonant, isn't enough to indicate genetic relatedness unless a very small wordlist was specified in advance. Johanna Nichols * * * * * * * * * * * * * * * * * * * * * Johanna Nichols Professor Department of Slavic Languages Mailcode 2979 University of California, Berkeley Berkeley, CA 94720, USA Phone: (1) (510) 642-1097 (direct) (1) (510) 642-2979 (messages) Fax: (1) (510) 642-6220 (departmental) * * * * * * * * * * * * * * * * * * * * * From larryt at cogs.susx.ac.uk Thu Nov 12 16:37:20 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 11:37:20 EST Subject: X>Y>X In-Reply-To: <364A7E48.2595E67@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, H.M.Hubey wrote: > If there are people who are writing programs to paint (yes, produce > art), and compose music, it takes no genius to see that even if > Starostin only wrote (or got a student to write) a brute-force, dumb > program on a commodity grade PC, he can uncover relationships that > many humans cannot do, even if they collaborate. The reason for > this will take too long to explain. But given a set of words (and > their meanings) even a brute-force program can keep cranking 24 > hours a day to produce cognates via regular sound changes, clusters, > and things that a typical linguist does not even know exists. No; this is not remotely so. The fundamental problem here, I think, is this. Such a dumb brute-force approach is obliged to treat all data on an equal footing. But the first thing you learn when you take up historical linguistics is that you *cannot* treat all data on an equal footing. Anyway, I might point out that just such brute-force programs already exist, that they have already been developed to a certain level of sophistication beyond the maximally dumb, and that they have already been applied to a number of individual cases. The ones I know most about are those developed at Cambridge, and these are interesting. However, these programs, interesting as they are, have certain inherent weaknesses. First, they cannot prove a linguistic relationship. At best, they can conclude that a genetic relationship is likely at the confidence level of 95%, or 99%, or whatever. And even these impressive-looking levels generally only arise in cases in which linguists have already established that a genetic link exists. Second, and more seriously, they cannot distinguish relatedness from non-relatedness. If you feed in data from, say, English, Dutch, French and Chinese, what you get is a tree in which English and Dutch are the two closest languages, French is somewhat more distantly connected, and Chinese is more distantly connected still. That is, the programs cannot distinguish an unrelated language from a distantly related language. One more thing. One of these programs has the curious habit of reporting a strong link between French and Hungarian at the 95% confidence level or above. Not IE and Hungarian, you understand: just French and Hungarian. Brute force or not, a mere dumb program is capable of reaching conclusions which any knowledgeable linguist knows are just plain wrong. Properly designed programs, in the hands of skilful linguists, are potentially capable of becoming a useful tool -- but certainly not a replacement for ordinary work in historical linguists. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 16:36:38 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 11:36:38 EST Subject: s > r (Iberian) In-Reply-To: <3.0.3.32.19981112015451.006d35bc@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >Now: are there any interesting sibilant-transforming phenomena in apitxat, >I wonder? > >(1) Does /s/ voice before voiced consonants? (In Castilian and Galician >this is the only context in which we find [z]. My money is on the same >thing happening in apitxat.) > >(2) Does /s/ rhotacize? Does it aspirate? Does it drop? (I'm guessing no >to all three...) My direct experience with apitxat is minimal, but I guess your guesses are correct. I'm still wondering where my non-voicing of /s/ in Castilian before /n/, /m/, /l/, /r/ comes from. Aragonese? That's where my father's from. I've just remembered a case of Catalan zetacism: the word "cupboard", vulgo [az'mari]. Could be hypercorrection, could be dissimilation, as opposed to assimilation in *ciresa > cirera "cherry". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From larryt at cogs.susx.ac.uk Thu Nov 12 16:36:09 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Thu, 12 Nov 1998 11:36:09 EST Subject: the Trask-Hubey debate In-Reply-To: <364A34D2.5287B5D3@Montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Wed, 11 Nov 1998, H. M. Hubey wrote: > That is the real reason I post what I do. There are computer programs that > paint and compose music. Is it really that hard to believe that linguistic > reconstruction is no less structured? There is no comparison. *Any* piece of paintwork we might produce counts as a painting. And *any* piece of music we might produce counts as music. But it is *not* true that any "reconstruction" we might produce counts as a reconstruction. A linguistic reconstruction is an attempt at recovering a real but unrecorded piece of prehistory, and it is not an attempt at producing an original work of art. I have little doubt that it would be possible to write a computer program that would chomp its way through any pile of linguistic data we chose to dump into it and spit out some kind of result, according to its instructions. But I see no reason to suppose that such a result would be anything but meaningless. > I bring this up, because for a long time the anti-AI crowd used > arguments similar to those offered often on linguistics lists for > why AI would be impossible. I do not believe this is true. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at wxs.nl Thu Nov 12 16:14:33 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Thu, 12 Nov 1998 11:14:33 EST Subject: s > r (Iberian) In-Reply-To: <3.0.3.32.19981111105818.006e93e8@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >the intervocalic >s > z development, general in Romance (all? most?), was later reversed in >Castilian and Galician, as opposed to most other Romance languages. It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, -g-, which must mean that the change was general in most Western Romance, but did not occur in Eastern Romance (S. Italian and Romanian). The exceptions in Western Romance are Mozarabic (partially, and as far as this can be determined) and Aragonese, or at least part of it. In High Aragon we see cases of maintained intervocalic -p-, -t- and -k-. Intervocalic -s- is /s/ now of course, but I wonder whether we have to assume a phase with /z/ for those Aragonese dialects that conserve -p-, -t- and -k-. The Ribargorc,a dialect of Catalan, adjacent to the Aragonese area, has intervocalic /s/, but it may be safely assumed that this is a recent phenomenon. Ribagorc,a` does not maintain intervocalic -p-, -t- and -k-, and there's even some evidence for rhotacism (-z- > -r-) in toponyms of the Pyrinean area (Glosianes > Glorianes). The standard language also has CERESIA > cirera "cherry", an admittedly isolated case of pure intervocalic rhotacism that hadn't been mentioned yet. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From drc at antnov1.auckland.ac.nz Fri Nov 13 01:17:16 1998 From: drc at antnov1.auckland.ac.nz (Ross Clark) Date: Thu, 12 Nov 1998 20:17:16 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- > Date: Thu, 12 Nov 1998 07:48:55 EST > Reply-to: hubeyh at montclair.edu > From: "H.M.Hubey" > Organization: Montclair State University > Subject: Re: rhotacism from Ray Hickey > To: HISTLING at VM.SC.EDU > ----------------------------Original message---------------------------- > Ross Clark wrote: > > > > Could you cite an example or two? I don't mean of 1. or 2. in use > > separately, but of the two used together as a fallacious syllogism. > > There are too many cases. I don't feel like embarrasing people > and making it worse. > I must insist. I deny that anybody uses 1. and 2. in this way. Prove me wrong. Embarrass somebody. > > > > > But borrowings also create regular sound correspondences. > > > > Yes, just as things other than measles can produce spots on the face. > > We need to take such things into consideration if we want to raise > > our competence in historical linguistics (or medical diagnosis) from > > this very rudimentary level. > > But that is not all. > > Human family members resemble each other. That does not mean that > unrelated > people cannot resemble each other. And despite the fact that we know > both > we still consider two people who resemble each other to be related > unless > there's proof to the contrary. We do? We arrive at this through experience. We > see > families (which we can confirm) resemble each other and therefore create > a general inductive rule. > > For measles, doctors know very well what healty people look like. > > How many language families has any human experienced? I do not mean the > purported/alleged language families. Since I don't know which entities qualify in your mind as "language families" as opposed to "purported/alleged language families", I can't answer this. However, on my own understanding of what a language family is, any competent historical linguist has experienced (has some knowledge of) a variety of language families and of various languages not known to be related. If a human could be created who > could > live 100,000 years or more and if we can send him to the past to learn > dozens of languages, then he would have 'experienced' language families > like human families and the way doctors (and others) have the experience > of > knowing what measles does. > > No such thing can be done in linguistics so all of it is based on > analogy > to models from the rest of the world, such as Linnaean trees, etc. That > is > coupled with some intuitive calculation of whether the resemblence is > due > to chance. You seem to be suggesting that the empirical base of historical linguistics is too small. Unless you have some realistic suggestion as to how it could be signficantly enlarged, we have to live with it. If that disqualifies it as "science" in your opinion, too bad. > > > > Well, we do in fact have records of various language families. What > > are you trying to say here? > > There we go again. Do we know these families like we know human > families or is this based on some calculation that the occurences > cannot be due to chance? Calculation of whether resemblances could be due to chance or not becomes relevant in distant relationships or borderline cases, about which so much argument goes on now. I'm talking about families even Lyle Campbell believes in, where there is no argument. In some cases (eg Latin and Romance) we have the proto-language through direct documentation. We also have recorded histories of many individual languages which tell us a lot about how languages change. > > > > > > 1. These languages have too many things in common. IOW, there are many > > > words in > > > these languages which can be made to look like each other with similar > > > meanings > > > and which could not be due to chance. > > > > > > 2. If that is not due to chance then either they got these words from > > > each other > > > or the words are all descended from a common language. > > > > > > 3. We have plenty of evidence (what?) that these languages did not get > > > these > > > words from each other. > > > > > > 4. Therefore these words in these languages must all come from an > > > earlier common > > > source. > > > Yes, problems crop up and arguments occur, to be sure. > > > > I recognize 1-4 as a rough outline of the reasoning by which one > > arrives at a hypothesis of genetic relatedness among languages. > > Rather than argue about details, I'd like to know where you're going > > with it. Are we finished with the idea that it's logically circular? > > But it is not finished. The key here is that we have to know what is > due to chance and what is not. Otherwise we can be creating an argument > like this: Well, this mathematical method says that X and Y are related > but I know that they are not, so the mathematical method is wrong. There > are people (yes, real people, and linguists too) who do this. In fact > one > of the superstarts of sci.lang and linguistics actually argued exactly > like this in email to me. This is not circularity. Mathematical methods can be wrong. A number of different proposals have been made for calculating the probability of accidental linguistic resemblances. They give different results. Therefore they can't all be right. They may be free of mathematical error, but they do not necessarily yield historically correct conclusions. On the other hand, one would want to ask the linguist in question just what was the basis of his/her certainty that X and Y are _not_ related. Ross Clark From hubeyh at montclair.edu Fri Nov 13 13:21:42 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:21:42 EST Subject: the Trask-Hubey debate Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > On Wed, 11 Nov 1998, H. M. Hubey wrote: > > > That is the real reason I post what I do. There are computer programs that > > paint and compose music. Is it really that hard to believe that linguistic > > reconstruction is no less structured? > > There is no comparison. > > *Any* piece of paintwork we might produce counts as a painting. And > *any* piece of music we might produce counts as music. But it is *not* > true that any "reconstruction" we might produce counts as a > reconstruction. Any piece of paintwork does not count as painting. The painters have their own black magic guild which decides what is art and what is not. Does that ring a bell? Besides, art and music were almost the last refuge for anti-AI crowds who thought they finally found something computers could not do but which "intellectuals" could do. Of course, they also deemed that doing math (and physics, and engineering, and compsci and ....) was mechanistic, rote, deterministic, rule-following, algorithmic stuff that computers could do, and "real intellectual activity" was things like playing chess, painting, writing poetry, composing music etc. Now they have no place to hide because what we see is that they are also about rules. And now during the last hurrah of anti-AI frenzy you are picking up the mantle and expect to convince me with decades old arguments which have already been blasted to smithereens. > A linguistic reconstruction is an attempt at recovering a real but > unrecorded piece of prehistory, and it is not an attempt at producing an > original work of art. > > I have little doubt that it would be possible to write a computer > program that would chomp its way through any pile of linguistic data we > chose to dump into it and spit out some kind of result, according to its > instructions. But I see no reason to suppose that such a result would > be anything but meaningless. I have no doubt that I can write a program to do exactly what you claim can't be done, just as I had no doubt that music creation or art creation could be automated. I also have no doubt that people like Starostin already have made much progress with his electronic and machine readable database of languages and etymologies. I also have no doubt that soon lots of others will be doing it. I also have very small doubts that I can write a program that can take n words from language A, and m from language B and write a program that can change one set into the other (at least enough of them to dumbfound the skeptics and force them to having N new looks into the "comparative method") using only regular sound changes. The only problem is that I don't have the time or energy to put into such a useless demonstration. It will probably be done by someone (like Starostin) or someone else who is upset at the way linguists spurn statistics and math. After all, it the only thing that will make people sit up and take notice there is no better way than to demolish their toy :-) > > I bring this up, because for a long time the anti-AI crowd used > > arguments similar to those offered often on linguistics lists for > > why AI would be impossible. > > I do not believe this is true. See, Hubey, H.M. (1996) "Topology of Thought", CC-AI: Journal for the Integrated STudy of Artificial Intelligence, Cognitive Science, and Applied Epistomology", vol 13, No.2-3, pp.225-292. > Larry Trask > COGS > University of Sussex > Brighton BN1 9QH > UK > > larryt at cogs.susx.ac.uk -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mcv at wxs.nl Fri Nov 13 13:22:15 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 08:22:15 EST Subject: Doing historical linguistics In-Reply-To: <364A7E48.2595E67@montclair.edu> Message-ID: ----------------------------Original message---------------------------- "H.M.Hubey" wrote: >It should be obvious to everyone by now that Starostin is way ahead of >the game than the 99% of the subscribers. It should be clear by now >that if he has already built up a database of lexicons of various >languages with their meanings, writeable in ASCII, Unicode etc, he has >singlehandedly done what should have been done by the linguistics community. I certainly agree that Starostin's databases are a great resource. But that is not all. >But that is not all. > >If there are people who are writing programs to paint (yes, produce >art), and compose music, it takes no genius to see that even if Starostin >only wrote (or got a student to write) a brute-force, dumb program on a >commodity grade PC, he can uncover relationships that many humans cannot do, >even if they collaborate. The reason for this will take too long to explain. Well, why don't you write (or get a student to write) a brute-force program to explain it? We *should* use information technology to assist research in historical linguistics. If only to nip in the bud attempts to relate Bq. "dry" with something in Caucasian ("Warning 23: oldest attested semantics: "barren, sterile"") or Bq. with words for "woman" sounding like /kwVn-/ ~ /kwVm-/ or Sum. with words for "woman" sounding like /em(e)/ ("Error 09: operation not commutative"). Certainly there are people using programs to paint and make music. But the value of the result depends entirely on the human input parameters, and on human selection/rejection/editing of the output. Even computer art is a craft. It takes a specialist (or a gifted person) to get interesting results. A program (brute-force or otherwise) to assist in reconstruction of proto-languages can be made, but surely to be useful it should incorporate existing knowledge about linguistics in its programming or configuration parameters (likely phonological developments and semantic shifts) and it should be fed reliable and complete data, including morphological information, otherwise it's GIGO. It is an illusion to think that we can throw away two centuries of comparative linguistic practice, and let "logic, probability theory and fuzzy set theory" do all the work. Mathematicians aren't going to be replaced by theorem proving automata any time soon, and neither are historical linguists scheduled to be replaced with "proto-language construction programs". ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From hubeyh at montclair.edu Fri Nov 13 13:23:02 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:23:02 EST Subject: domina/womina, haber, bad Message-ID: ----------------------------Original message---------------------------- Larry Trask wrote: > > dialect, in which it means `girl', which we believe to have been its > original meaning, since the etymology is * `female offspring'. > > The earlier word for `woman' was , which today has been > specialized to `wife', except in Salazrese, where it still means > `woman'. > > It will surprise no one to learn that both and are > compounds built upon the word `female', whose regular combining > form is (look at two of the examples above). We suspect an I don't recall writing that it was borrowed into Basque. What I did write, however was that Lahovary (Dravidian and the West) writes about the pre-IE and pre-AA languages of the "Mediterranean" and tries to connect them with Dravidian and Caucasian. That 'em' word shows up in them, with related meanings. von Soden, points out that pre-IE and pre-AA languages of the region confused the liquids /r/ and /l/ (especially in the beginnings of words), etc. Obviously, the languages which do not have words beginning with liquids and which confuse them are Dravidian and Altaic. Obviously all of this matters to some people. > etymology * `young female'. And, as Miguel C V has pointed > out, this is borrowed from Gascon , itself from Latin > . Basque contains many loans from Gascon. The /mn/ cluster kap>capere, ulu/uru > ululare, karnash > carnal, tes > seks, oghlan > clan, and now fem > hem > em, etc. > The Sumerian word is not relevant, and it has been effectively disposed > of by Miguel. Turkish does not mean `mother'; it means `but'. Turkish yes, Turkic no. Karachay-Balkar uses ata, ana, atta, anna, appa, akka, amma,... and most of these words can be found in Turkic languages. > The Turkish word for `mother' is in Anatolian Turkish but > in standard Istanbul Turkish, this last apparently being an expressive > variant of . It is also a part of similar changes. I think Chuvash is also anne or ane. > Turkish `nipple' is not available for comparison. One of the > best-known facts about Turkish is that native Turkish lexical items do > not begin with /m/ unless they are imitative words or nursery words. In > all likelihood, we are looking at a nursery word here, and nursery words > cannot be cited as comparanda, because they are so often created > independently. This statement seems to fly in the face of what is known. Babies can create and do babble every sound there is. That means that there is some reason why parents pick up certain types of words. The most likely explanation is that those words already exist in the language and the parents are happy to hear those words. This means that your argument is backwards (ahistorical accretion) and one (or the only) purpose for its existence is the fact that it makes the IE theory look good and the proto-world look bad. Just recently you mentioned that there are 6,000+ languages. In how many of them are "nursery" words present? The list posted to sci.lang (or some other usenet list) by Miguel listed less than 100 languages. That is 100/6,000 which is 1/60. Since when does 1/60 become the standard? By whose arithmetic? > The stem `suck' is the source of the derivatives `nipple', > `nipple', and `suckle'; these can only be counted as > one word. So there is some value to doing clustering analysis and semantic distances, and Swadesh-like lists after all. > Finally, is merely a diminutive of `vulva', containing the > usual Turkish diminutive suffix <-cik>. > > Hence all we have here for Turkish is a verb-stem `suck' and a > noun `vulva'. And nothing whatever can be concluded from this. > It's every bit as impressive as English `ear' and `hear'. That once again shows the classic method. Take each possible cognate one at a time, and ignore them one at a time. When someone spends from 1947 to 1990 patiently collecting words (Dr. Tuna) and presents it all at once, then what? Change the Sumerian words? Once again this gets us in the thick of things. How many such words are needed? If it happens all the time, why don't linguists list them someplace instead of listing the perennial 'bad' (English,Farsi), haber (Latin, German), domina/womina (Italian,Japenese) etc. Maybe this is the real big reason I want things quantified. IT reminds me of the joke about the accountant. At a job interview he is asked "How much is 2+2?". He says "How much do you want it to be?" Is this how science is supposed to be done, or is this how linguistics "science" supposed to be done (according to you)? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From mccay at redestb.es Fri Nov 13 13:23:43 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 13 Nov 1998 08:23:43 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer wrote: >Apparently, new geminates -tt- arose from syncope, giving -d- >(cadair) and -tt- (eto /etto/) as the only allowed intervocalic >stops. I accept the correction. I was writing from memory, without the relevant manuals to hand. But I don't think this affects the gist of my argument. >But that would put the /katar/ varieties of Welsh above Cornish and >Breton in the branching tree, which doesn't seem very likely. I think it's pretty obvious that /kadar/ > /katar/ is an innovation, not a survival. I was merely saying that it has to be one or the other (and is, in fact, the one, not the other). >What's missing in the above is a variety of Welsh that has eliminated >geminates: D, d, T, t. One would expect one. I'll say what I know (or think I know) about this. As I said before, except in certain cases (e.g. "canu" versus "cannu"), it is predictable in modern Welsh whether a given intervocalic consonant preceded by a stressed vowel will be geminate or simple, and in the case of stops, the standard rule is voiced -> simple, and voiceless -> geminate. So I think it can be argued that gemination may not be phonemic, except for those "certain cases". Thus [t] and [tt] would be allophones of /t/, etc. Given that situation, the issue seems to belong to the domain of surface phonetics. The difference between southern and northern dialects in this respect would then be that in the south only voiceless stop phonemes have the two allophones (simple and geminate) while in the north this pattern has been extended to include voiced stops. SOUTH (more conservative in this respect): /t/ -> [t], [tt] /d/ -> [d] NORTH: /t/ -> [t], [tt] /d/ -> [d], [dd] hence e.g. southern [etto], [ka:der] vs. northern [etto], [kaddar] for /eto/, /kadVr/. Not being a phonetician nor a native Welsh speaker (but I do speak Welsh), my IMPRESSION from contact with speakers of different varieties of Welsh, and reading on the subject, is that: in the NORTHERN dialects, the acoustic effect of both the gemination of the stops and the shortening of the preceding stressed vowels is very striking and "vigorous". At least in many speakers, in a word like /kadar/, the intervocalic consonant seems to "last" considerably longer than the preceding stressed vowel (sic!), which sounds "clipped" to my foreign ear. For the stop to "last" that long, I believe the flow of breath must be interrupted completely in the middle, and that is what it sounds like: [kad] (pause) [dar], probably more accurately [kat] (pause) [tar], with no aspiration. (In my experience, speakers from the Iberian peninsula who learn Welsh tend to identify this as [katar].) in the SOUTHERN dialects, where gemination affects a smaller range of consonants, my general impression is that the distinction is maintained but in a phonetically more low-key manner, and since the distinction is phonetic rather than phonemic, my guess is that there is probably considerable variation in how these consonants are realized. On the other hand, for the southern dialects, the distinction, while not phonemic, retains *phonological* significance in that it correlates with a length contrast in the preceding vowel ([ka:der] versus [et(t)o]), and the vowel length distinction *is* observed (very noticeably if you're more used to hearing northern Welsh). Thus while the simple/geminate contrast is not by itself distinctive in most cases (it not only correlates with length of the adjacent vowel but is furthermore itself usually predictable from other features, such as voice), the fact that it does correlate with (clearly perceptible) vowel quantity probably helps to reinforce the speaker's *competence* regarding consonant length, even if structurally superfluous. Then there are the cases where consonant gemination remains phonemic in southern Welsh, e.g. /ka:ni/ 'sing' versus /kanni/ 'whiten' (orthographically, "canu" and "cannu", and both /kannI/ in northern Welsh). It would be interesting to know whether some southern dialects actually confuse such pairs, but this seems unlikely because the vowel "helps", even if the gemination contrast is weakened. In any case, I have not read or heard of southern dialects in which the gemination contrast is lost completely, both phonologically and phonetically. Alan From mccay at redestb.es Fri Nov 13 13:24:12 1998 From: mccay at redestb.es (Alan R. King) Date: Fri, 13 Nov 1998 08:24:12 EST Subject: s > r (Romance) Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer writes: >It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, >-g-, which must mean that the change was general in most Western >Romance, but did not occur in Eastern Romance (S. Italian and >Romanian). In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > -z- has taken place. I would have guessed (from my position of overall ignorance about the Italian dialects) that this probably also reflected roughly the situation in dialectal Italian south of the Spezia-Rimini line (north of the line, -s- is also voiced, of course, but so are the stops). Doesn't this establish a precedent clearly opposed to assuming (in the absence of further data, I mean) that other Romance dialects which didn't lenite the stops couldn't have lenited -s-? I would have thought so. It is easy enough to see why the (apparent or real) parallel between voicing of intervocalic stops, on the one hand, and of -s-, on the other, suggest the conclusion that a single sound-change event is involved, rather than two separate phenomena, but that is surely not a foregone conclusion a priori? Among non-Romance languages in which regular voicing of (some) intervocalic consonants has occurred, I believe there are cases of voicing of stops but not of -s-, and of course there are languages with the opposite, voicing of -s- but not of stops (e.g. Old English). An example of the former combination may be Welsh in its development from proto-Celtic via proto-Brittonic. Intervocalic stops were voiced, as I remarked in my previous post and is anyway well known. Intervocalic s was in general aspirated, never voiced. My doubt is as to which happened first. If s aspirated first, it could then be argued that the reason why s wasn't voiced together with the stops is that at the time the stops were voiced, there were no intervocalic sibiliants around (only h's). Even if the Welsh example is dubious (as I said, that would seem to depend on the relative chronology of the voicing of stops and aspiration of -s-), perhaps there are examples in other languages? Coming back to Romance, though, and getting somewhat aprioristic myself, I would have thought that a weightier factor capable, potentially, of influencing the way -s- developed into -z-, would be, not what happened with the intervocalic stops, but what happened with *geminate* consonants in general, or at any rate with -ss- in particular. I'm thinking now in terms of structural economy. In dialects where geminates were simplified, non-geminate -s- either had to evolve to -z- (or at least to something other than -s-) or else be confused with -ss- > -s-. Where, on the other hand, geminates are maintained, there can have been no such pressure. The same argument applies, of course, to the lenition of intervocalic voiceless stops. Thus, if such principles of structural economy operate, we would indeed expect -s- and voiceless stops to voice in the same dialects and in the same contexts, unless other factors intervened to differentiate, because both developments would have been responses to similar pressures. But that *is* aprioristic. Returning to facts, isn't Italian a counter-example? Alan From hubeyh at montclair.edu Fri Nov 13 13:24:48 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:24:48 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- Ross Clark wrote: > > I must insist. I deny that anybody uses 1. and 2. in this way. Prove > me wrong. Embarrass somebody. Lots of people do, and some of them even have PhDs. > > Human family members resemble each other. That does not mean that > > unrelated > > people cannot resemble each other. And despite the fact that we know > > both > > we still consider two people who resemble each other to be related > > unless > > there's proof to the contrary. > > We do? Of course we do. WE make lots of decisions about the world based on a kind of a "theory" of everything that we have created about us based on our life experiences from the moment of birth. Unless we know the people it is similarity that propels us to thinking such things. > > How many language families has any human experienced? I do not mean the > > purported/alleged language families. > > Since I don't know which entities qualify in your mind as "language > families" as opposed to "purported/alleged language families", I > can't answer this. However, on my own understanding of what a > language family is, any competent historical linguist has experienced > (has some knowledge of) a variety of language families and of various > languages not known to be related. But that is not what I asked nor commented on. There lies the circularity again. How do we know that they constitute a family? > You seem to be suggesting that the empirical base of historical > linguistics is too small. Unless you have some realistic suggestion > as to how it could be signficantly enlarged, we have to live with it. > If that disqualifies it as "science" in your opinion, too bad. No, I am not suggesting that. I am merely stating the reasons for belief that language families exist, and that we can come to know them. > Calculation of whether resemblances could be due to chance or not > becomes relevant in distant relationships or borderline cases, about > which so much argument goes on now. I'm talking about families even > Lyle Campbell believes in, where there is no argument. In some cases > It is always relevant. The fact that some people's beliefs can be based on analogy and resemblance to physics or biological families or imitation does not change the facts. (eg Latin and Romance) we have the proto-language through direct > documentation. We also have recorded histories of many individual > languages which tell us a lot about how languages change. We have Latin at some stage and its descendants. The others are then created using analogies, which is what at least some linguists deny. But they also deny that the reasoning (about all those resemblances not being due to chance) is implicitly probabilistic. So then what is left? It's not analogy, it's not induction, it's not probability! > This is not circularity. Mathematical methods can be wrong. A number This is worse than pseudoinduction. Which mathematical method is wrong? No axiomatic system can ever be wrong by definition. > of different proposals have been made for calculating the probability > of accidental linguistic resemblances. They give different results. > Therefore they can't all be right. They may be free of mathematical > error, but they do not necessarily yield historically correct > conclusions. It is not the method, it is the person who uses it. If you bash your finger using a hammer is it the hammer's fault? There are incompetent people in many places. > On the other hand, one would want to ask the linguist in question > just what was the basis of his/her certainty that X and Y are _not_ > related. Why ask a person when there are dozens of books which purport to explain it? > > Ross Clark -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From larryt at cogs.susx.ac.uk Fri Nov 13 13:27:20 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Fri, 13 Nov 1998 08:27:20 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Thu, 12 Nov 1998, John Hewson wrote: > Actually there is an algorithm that a Systems Analyst and I > developed to produce my _Computer-generated Dictionary of > Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be > stated in a single sentence, as on p.iv of the above: From the data of the > daughter languages generate all possible protoforms, then sort > alphabetically, and examine all sets of identical protoforms collocated by > the sort. Very interesting, but I'm amazed. What particularly catches my attention is the expression "all possible protoforms". How on earth can this label be fleshed out? How do you know that your program generates all possible protoforms? Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From hubeyh at montclair.edu Fri Nov 13 13:27:46 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 08:27:46 EST Subject: Yeniseian and Na-Dene Message-ID: ----------------------------Original message---------------------------- Johanna Nichols wrote: > > Neither of these procedures guarantees fair coverage of vastly different > frequencies of different consonants, language-specific or family-specific > preferences of different consonants or consonant classes for different > phonotactic positions, and the like. I hope that some of these differences > get ironed out by putting consonants together in phonetic groupings. How about simulation results? Have you done any ? > Still, the metric is only approximate. It enables us to point out that 36 > resemblant sets, half of them with only one resemblant consonant, isn't > enough to indicate genetic relatedness unless a very small wordlist was > specified in advance. What is your conclusion on how many such pairs are needed? Is this number for the whole language or just using a Swadesh like list? -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From Roger.Wright at liverpool.ac.uk Fri Nov 13 13:33:46 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 13 Nov 1998 08:33:46 EST Subject: s > r (Iberian) In-Reply-To: <364adb3b.1343704@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- In fact, the supposed non-voicing of intervocalic plosives in Mozarabic (Ibero-Romance spoken in Moslem Spain) was probably a chimera caused by the Arabic alphabet's application to Ibero-Romance; I think most specialists are rather dubious about it by now. And the Aragonese referred to is high up in the Pyrenees, and not characteristic of the whole of Aragon - RW On Thu, 12 Nov 1998, Miguel Carrasquer Vidal wrote: >It's hard to see -s- > -z- separate from -p-, -t-, -k- > -b-, -d-, >-g-, which must mean that the change was general in most Western >Romance, but did not occur in Eastern Romance (S. Italian and >Romanian). > >The exceptions in Western Romance are Mozarabic (partially, and as >far as this can be determined) and Aragonese, or at least part of it. From Roger.Wright at liverpool.ac.uk Fri Nov 13 13:35:14 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 13 Nov 1998 08:35:14 EST Subject: Stabilized languages In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Larry said: >First of all, languages do not "stabilize". Every living language is >always changing, and the only language that doesn't change is a dead >one. Change is at some times slower than at other times, though. And it just isn't true that "dead" languages don't change. Medieval Latin varied astonishingly widely in time and space, Renaissance Latin was recognizably different again, and even the modern Latin used by the Vatican has been continually acquiring new vocabulary. RW From mcv at wxs.nl Fri Nov 13 21:27:04 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 16:27:04 EST Subject: s > r (Iberian) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Roger Wright wrote: >And the Aragonese >referred to is high up in the Pyrenees, and not characteristic of the >whole of Aragon. Not anymore, although a word like still covers the whole Aragonese area. Place names and medieval documents prove that retention of -p-, -t- and -k- was also wide-spread in Lower Aragon until the 15th century or so. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From mcv at wxs.nl Fri Nov 13 21:26:49 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 13 Nov 1998 16:26:49 EST Subject: s > r (Romance) In-Reply-To: <3.0.3.32.19981113084536.006e4c00@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- "Alan R. King" wrote: >In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > >-z- has taken place. I would have guessed (from my position of overall >ignorance about the Italian dialects) that this probably also reflected >roughly the situation in dialectal Italian south of the Spezia-Rimini line >(north of the line, -s- is also voiced, of course, but so are the stops). It's not that simple. Standard Italian has plenty of cases of voiced stops (Rohlfs gives: ago, drago, lago, sugo, spiga, briga, segale, luogo, lattuga, aguzzo, pagare, segare, pregare, sfogare, affogare, fregare, annegare, frugare, intrigare; spada, strada, malgrado, contrada, rugiada, contado, scudo, lido, padella, badia, badessa, badile, gradire, scodella, gridare, medaglia, stadera, mortadella, podesta`, budello, mudare, podere; povero, vescovo, arrivare, ricevere, cavezza, ricoverare, rimproverare, sceverare, scovolo). As to -s-, it is pronouned voiceless /s/ in asino, cosa, casa, mese, fuso, peso, naso, the suffixes -oso, -ese (except cortese, francese, marchese, palese, paese), the verbal endings -esi, -isi, -osi, -usi, -eso, -iso, -uso. Voiced /z/ is found in: base, battesimo, bisogno, caso, chiesa, crisi, cristianesimo, deserto, dose, fantasia, fase, fisica, lasagna, lesina, medesimo, misero, musica, osare, pausa, posa, paradiso, rosa, quaresima, quasi, spasimre, sposo, scusare, te`si, uso, usare, vaso, ventesimo, viso (from : esame, esatto, esempio, esemplare, esiguo, esigere, esercito, esente, eseredare). According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in approxiamte order of frequency) are not native Tuscan forms, but words imported from the north. What *is* native Tuscan is the development -k- > -h- ("gorgia toscana"), and also (in a smaller area) -t- > -T- and -p- > -P-. As in the Goidelic lenitions, this also occurs across word boundaries. To the south of Tuscany, an equally "sandhinista" area is Lazio (as well as Corsica and Sardinia), but this time the parallel is with Brythonic (-p- > -b- [here we can distinguish between Northern imports and local lenitions], -t- > -d-, -k- > -g-). Apparently -s- participates in Corsica (/u zale/ "the salt"), but I'm not sure about Lazio. There is no "aspiration" of /s/ in Tuscany, whatever that might mean. So these *are* counterexamples. However, the Eastern Romance non-lenition (south of La Spezia-Rimini) applies both to the stops and to -s-. Of course the imbalance in the system (3/4-way for the stops /d/, [/dd/,] /t/, /tt/, 2-way for the sibilant /s/, /ss/) may have favoured the influx of northern words containing /z/. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From hubeyh at montclair.edu Fri Nov 13 17:03:33 1998 From: hubeyh at montclair.edu (H.M.Hubey) Date: Fri, 13 Nov 1998 12:03:33 EST Subject: Doing historical linguistics Message-ID: ----------------------------Original message---------------------------- Miguel Carrasquer Vidal wrote: > Well, why don't you write (or get a student to write) a brute-force > program to explain it? Students are not what they used to be :-) > We *should* use information technology to assist research in > historical linguistics. If only to nip in the bud attempts to relate > Bq. "dry" with something in Caucasian ("Warning 23: oldest > attested semantics: "barren, sterile"") or Bq. with words > for "woman" sounding like /kwVn-/ ~ /kwVm-/ or Sum. with > words for "woman" sounding like /em(e)/ ("Error 09: operation not > commutative"). As they say "it all comes out in the wash". That is why averages are used in statistics. The fact is that if there are lots of errors in the present method, there's no way to find out, exactly as in the case of using stats. For example, without Sumerian, no book like Tuna's would have been possible. That means that many Turkic words could have always been attributed to Tocharian, Iranian etc where now one may be able to point to roots earlier than IE as in Hittie etc. That means the comparative method also suffers from the same problems as any stat method, but that cannot be discovered. It may be that looking at thousands of languages employing brute-force methods might create results which themselves could be useful for standard practice. But at the same time clear models are needed. After all, what we want in the final analysis is if there exists any pattern that is not likely due to chance. > Certainly there are people using programs to paint and make music. > But the value of the result depends entirely on the human input > parameters, and on human selection/rejection/editing of the output. > Even computer art is a craft. It takes a specialist (or a gifted > person) to get interesting results. There's no doubt that humans created computers and wrote the programs. There's also no doubt that computers can now do things that great mathematicians used to do. So there should also be no doubt that they can do what linguists do. > A program (brute-force or otherwise) to assist in reconstruction of > proto-languages can be made, but surely to be useful it should > incorporate existing knowledge about linguistics in its programming > or configuration parameters (likely phonological developments and > semantic shifts) and it should be fed reliable and complete data, > including morphological information, otherwise it's GIGO. Existing knowledge, yes, existing dogma no. There needs to be clear models not lumping things constantly ignoring possible errors. If nothing else familiarity with some aspects of probability theory will restrain linguists from making the kinds of blanket/certain statements which no respectable scientist would ever make. In this case I am referring particularly to kinds of statements which Larry Trask often makes, probably thinking that he is discussing chemistry probably via analogy. > It is an illusion to think that we can throw away two centuries of > comparative linguistic practice, and let "logic, probability theory > and fuzzy set theory" do all the work. Mathematicians aren't going > to be replaced by theorem proving automata any time soon, and neither > are historical linguists scheduled to be replaced with > "proto-language construction programs". Mathematicians might be replaced by computers. About a year ago a program which was more or less brute-force proved a theorem and when people (mathematicians) examined the theorem, they thought it was rather ingenious. Let us not forget about Big Blue. I will post the music article separately. The reason all this happens is because people make intuition errors. Intuition (creativity, etc) is not all what it is cracked up to be, and it is mostly wrong to rely on it. > > ======================= > Miguel Carrasquer Vidal > mcv at wxs.nl > Amsterdam -- Best Regards, Mark -==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= hubeyh at montclair.edu =-=-=-= http://www.csam.montclair.edu/~hubey =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From jhewson at morgan.ucs.mun.ca Fri Nov 13 17:02:28 1998 From: jhewson at morgan.ucs.mun.ca (John Hewson) Date: Fri, 13 Nov 1998 12:02:28 EST Subject: Doing historical linguistics (part 2) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, Larry Trask wrote: > On Thu, 12 Nov 1998, John Hewson wrote: > > > Actually there is an algorithm that a Systems Analyst and I > > developed to produce my _Computer-generated Dictionary of > > Proto-Algonquian_ (Ottawa: National Museums of Canada, 1993). It may be > > stated in a single sentence, as on p.iv of the above: From the data of the > > daughter languages generate all possible protoforms, then sort > > alphabetically, and examine all sets of identical protoforms collocated by > > the sort. > > Very interesting, but I'm amazed. What particularly catches my > attention is the expression "all possible protoforms". How on earth can > this label be fleshed out? How do you know that your program generates > all possible protoforms? Let me try to answer with a simple example. An /s/ in Fox can come from 6 different Proto-Algonkian sources. The program generates all 6 from every /s/ in a Fox word. Let us call this the productivity of Fox /s/. The total of all possible protoforms for a Fox word is based on the productivity of all the phonemes in the word. Since Algonkian words often contain several morphs, the number of possible protoforms can be very high. In the first batch we did, the average number of possible protoforms for every item of input data was over 20. Given the length of Algonkian words we were able to simplify by deleting the vowels and working from the consonant frameworks only. Otherwise the program would probably have bombed, even on the mainframe on which we were working at that time. Lowe and Mazaudon, by contrast, worked on monosyllables that had tone. Such systems are for the reconstruction of the vocabulary once all the correspondences and reflexes of the relevant daughter languages have been worked out. It saves the worker the donkey work of pounding through dictionaries looking for items that may not exist. It cannot replace the linguist, it is merely a tool to aid the linguist. Algorithms may operate mechanically, but they are not created mechanically. To create them requires imagination, and to use them requires a professional level of understanding of the comparative method. John Hewson, FRSC tel: (709)737-8131 Henrietta Harvey Professor of Linguistics fax: (709)737-4000 Memorial University of Newfoundland St. John's NF, CANADA A1B 3X9 From scott at math.csuohio.edu Fri Nov 13 17:00:24 1998 From: scott at math.csuohio.edu (Brian M. Scott) Date: Fri, 13 Nov 1998 12:00:24 EST Subject: rhotacism from Ray Hickey Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > Ross Clark wrote: > > This is not circularity. Mathematical methods can be wrong. A number > > This is worse than pseudoinduction. Which mathematical method is wrong? > No axiomatic system can ever be wrong by definition. It may, however, be inapplicable. If a mathematical method produces a historically incorrect conclusion, then obviously the assumptions on which it is based do not obtain, and it is wrong, i.e., the wrong method. Brian M. Scott Dept. of Mathematics Cleveland State Univ. From cravens at macc.wisc.edu Sat Nov 14 15:15:45 1998 From: cravens at macc.wisc.edu (Thomas D. Cravens) Date: Sat, 14 Nov 1998 10:15:45 EST Subject: s > z (etc.) (Italy, Spain) Message-ID: ----------------------------Original message---------------------------- Comments below (Tom Cravens) (Note: I [TC] mistakenly sent this directly to Miguel, rather than the list. Perhaps the discussion of weakening by class is of some general interest.) At 04:26 PM 11/13/98 EST, you wrote: >----------------------------Original message---------------------------- >"Alan R. King" wrote: > >>In Standard Italian intervocalic stops are (mostly) not voiced, but -s- > >>-z- has taken place. I would have guessed (from my position of overall >>ignorance about the Italian dialects) that this probably also reflected >>roughly the situation in dialectal Italian south of the Spezia-Rimini line >>(north of the line, -s- is also voiced, of course, but so are the stops). > >It's not that simple. Standard Italian has plenty of cases of voiced >stops (Rohlfs gives: ago, drago, lago, sugo, spiga, briga, segale, >luogo, lattuga, aguzzo, pagare, segare, pregare, sfogare, affogare, >fregare, annegare, frugare, intrigare; spada, strada, malgrado, >contrada, rugiada, contado, scudo, lido, padella, badia, badessa, >badile, gradire, scodella, gridare, medaglia, stadera, mortadella, >podesta`, budello, mudare, podere; povero, vescovo, arrivare, >ricevere, cavezza, ricoverare, rimproverare, sceverare, scovolo). >As to -s-, it is pronouned voiceless /s/ in asino, cosa, casa, mese, >fuso, peso, naso, the suffixes -oso, -ese (except cortese, francese, >marchese, palese, paese), the verbal endings -esi, -isi, -osi, -usi, >-eso, -iso, -uso. Voiced /z/ is found in: base, battesimo, bisogno, >caso, chiesa, crisi, cristianesimo, deserto, dose, fantasia, fase, >fisica, lasagna, lesina, medesimo, misero, musica, osare, pausa, >posa, paradiso, rosa, quaresima, quasi, spasimre, sposo, scusare, >te`si, uso, usare, vaso, ventesimo, viso (from : esame, esatto, >esempio, esemplare, esiguo, esigere, esercito, esente, eseredare). Unless I'm mistaken (I don't have the volume at hand), Rohlfs is talking about Tuscan in that passage, not Standard Italian. The two are distinct, although they match up fairly well for the examples he gives (but not always across the lexicon: Mod. Florentine /letikare/ vs. It. /litigare/ 'argue, fight'). In any case, data pulled out of Standard Italian (assuming we can identify what that is) are not the most felicitous for constructing or testing theories, given that the Standard was presumably no one's native language until relatively recently, and even in the far-off days of trying to derive a standard from Tuscan, that standard was based on literary models, rather than actual speech, and ultimately was and has been subject to a good amount of winnowing. [It's of interest how much Tuscan detail ("tutte helle hosine" someone once told me) is absent from Italian.] >According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in >approxiamte order of frequency) are not native Tuscan forms, but >words imported from the north. Yes, Rohlfs says so, but this bit isn't that simple, either. Restructured voicing of /p t k/ is not lexically consistent within Tuscany. More importantly, it affects forms which are not northern (e.g. codesto), some which did not voice in N. Italy or even in Spain (PAUCU > W. Tuscan /pogo/, AVICA > W. Tuscan /oga/), and lots of local toponyms, presumably not borrowed. The latest work on this suggests strongly that surface-level (allophonic) voicing was once much more widespread in Tuscany than it is today, and that much of the /p t k/ > /b-v d g/ now found may be the result of reinterpretation of surface forms during rule competition (gorgia coming in to compete with voicing). It's still going on. Relevant to the current thread is that /p t k/ and /s/ appear to have been affected by most of the same complex of rules, i.e. that Tuscan doesn't provide a good counterexample to the hypothesis that /p t k/ and /s/ are expected to undergo restructuring en bloc. The gorgia mentioned below does, though, and draws attention to the expectation that sound change (accepting that introduction of allophony is a change in sound) will proceed member by member within classes, with nothing to suggest that it must spread outside the innovating class. It can, of course, as Corsican /s/ --> [z] along with voicing of /p t k/ shows, but it needn't be the case. Alto Aragones illustrates this as well with regard to degemination, or at least did fifty years ago when Badia Margarit was investigating the speech of Bielsa. Working on memory again, but I seem to recall that in the midst of the massive degemination that one would expect in that part of the Romance world, there was maintenance of length not only for the expected /rr/, but also for /ll/ and /nn/ (and possibly even /mm/, but there were very few examples, one of which was [tammjen], suspect as a recent, and perhaps surface-only, assimilation). In any case, Belsetan by that time had degeminated all of /pp tt kk/, /bb dd gg/, etc, but was hanging on to the last remnants of /nn/, /ll/, /rr/. Marie-Jose' Dalbera-Stefanaggi has found similar phenomena in Northern Corsica: liquids and nasals are the last to lose the possibility of manifesting surface length. In neither type did degemination of the stops imply (immediate) degemination outside their class. >What *is* native Tuscan is the development -k- > -h- ("gorgia >toscana"), and also (in a smaller area) -t- > -T- and -p- > -P-. As >in the Goidelic lenitions, this also occurs across word boundaries. >To the south of Tuscany, an equally "sandhinista" area is Lazio (as >well as Corsica and Sardinia), but this time the parallel is with >Brythonic (-p- > -b- [here we can distinguish between Northern >imports and local lenitions], -t- > -d-, -k- > -g-). Apparently -s- >participates in Corsica (/u zale/ "the salt"), but I'm not sure about >Lazio. There is no "aspiration" of /s/ in Tuscany, whatever that >might mean. So these *are* counterexamples. Yes. I think Lazio is typically [s], but the voicing of /p t k/ is extensive enough to cause spelling problems a` la ~ in Am. E. >However, the Eastern Romance non-lenition (south of La Spezia-Rimini) >applies both to the stops and to -s-. Of course the imbalance in the >system (3/4-way for the stops /d/, [/dd/,] /t/, /tt/, 2-way for the >sibilant /s/, /ss/) may have favoured the influx of northern words >containing /z/. Maybe. But there's still the gap of no /zz/, and no phonologically "breve" counterpart of long palatal laterals and nasals, i.e. no *[aljo] possible to contrast with [alj:o] 'garlic', no *[onji] can contrast with [onj:i] 'each' (apologies for the fake IPA). Tom Cravens University of Wisconsin-Madison cravens at macc.wisc.edu >======================= >Miguel Carrasquer Vidal >mcv at wxs.nl >Amsterdam From fcosw5 at mail.scu.edu.tw Sat Nov 14 20:33:59 1998 From: fcosw5 at mail.scu.edu.tw (Steven Schaufele) Date: Sat, 14 Nov 1998 15:33:59 EST Subject: concerning the algorithmizability of historical linguistics Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > I also have very small doubts that I can write a program that can take > n words from language A, and m from language B and write a program that > can change one set into the other (at least enough of them to dumbfound > the skeptics and force them to having N new looks into the "comparative > method") using only regular sound changes. The only problem is that I > don't have the time or energy to put into such a useless demonstration. > It will probably be done by someone (like Starostin) or someone else > who is upset at the way linguists spurn statistics and math. After all, > it the only thing that will make people sit up and take notice there is > no better way than to demolish their toy :-) This is precisely where i feel we are taking leave of scientific responsibility. If you can write a program that can take n words -- *any* n words -- from language A -- *any* A -- and m words -- *any* m words -- from language B -- *any* B -- and change one set into the other, how is such a result falsifiable? Such a program sounds to me like one that is guaranteed to find evidence of affiliation no matter what data are fed to it. And i'm quite sure that is not desirable. Should we ever discover extraterrestrial, or otherwise non-human, language, such an approach as Mark seems to be suggesting here would surely be able to `prove' that such a language is nevertheless affiliated somehow with any given human language. Which would definitely be a major step in the wrong direction, since at least IMHO the discovery of such a language ought precisely to provide us with an opportunity to get beyond studying merely human language (presumably a manifestation of species-specific characteristics of the human brain, etc.) and onto the path of examining the nature of Language itself in the abstract. I'm concerned that Mark's approach to developing AI programs for historical linguistics might also lend itself to developing programs that would quite happily `demonstrate' that, e.g., fish, ichthyosaurs, and whales are all more closely related to each other than any of them are to salamanders, turtles, or cows, or that dragonflies, eagles, and bats are more closely related to each other than any of them are to spiders, penguins, or shrews. As, e.g., Larry and Miguel have noted, AI systems can be extremely helpful in doing a lot of the comparative work our science is based on, and perhaps even draw our attention to relationships and affiliations we might not otherwise have considered. But the results of such analyses still need to be subjected to expert peer review. Maybe a certain amount of that sort of thing can be automated too, but i very much doubt that all of it can be, at least in the rather elementary, simple-minded manner of reducing it to an algorithm. There's a lot of judgment involved, which judgment needs to be based on a lot of experience which is definitely not algorithmizable. To what extent that kind of experience can be taught to a computer, or a computer programmed to learn it as a living, flesh-and-blood historical linguist does, i confess i have no idea. I am also bothered, as is Larry, by Mark's continued (apparent) tendency to equate scientific research in a field such as (historical and comparative) linguistics with artistic endeavours such as painting pictures, composing music, or playing chess. A freshly-painted picture, a new piece of music, a game of chess is a new creation and derives some of whatever value it has from that novelty; the same is true, of course, of a freshly-crafted, previously-unused and -unheard-of sentence. Or a mathematical theorem, for that matter. A reconstructed protoform, on the other hand, is a hypothesis concerning something (a linguistic expression) that is supposed to have really existed in the objective universe at some point in the past. Its validity ultimately stands or falls on how closely it approximates objective, historical reality, not on its solipsistic elegance. Best, Steven -- Steven Schaufele, Ph.D., Asst. Prof. of Linguistics, English Department Soochow University, Waishuanghsi Campus, Taipei 11102, Taiwan, ROC (886)(02)2881-9471 ext. 6504 fcosw5 at mail.scu.edu.tw http://www.prairienet.org/~fcosws/homepage.html ***O syntagmata linguarum liberemini humanarum!*** ***Nihil vestris privari nisi obicibus potestis!*** From bwald at HUMnet.UCLA.EDU Sat Nov 14 20:34:50 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sat, 14 Nov 1998 15:34:50 EST Subject: X>Y>X Message-ID: ----------------------------Original message---------------------------- Mark Hubey makes an interesting, visionary and somewhat naive point -- as follows: >If there are people who are writing programs to paint (yes, produce >art), and compose music, it takes no genius to see that even if Starostin only >wrote (or got a student to write) a brute-force, dumb program on a >commodity grade PC, he can uncover relationships that many humans cannot >do, even >if they collaborate. The reason for this will take too long to explain. >But given a set of words (and their meanings) even a brute-force >program can keep cranking 24 hours a day to produce cognates via regular >sound changes, clusters, and things that a typical linguist does not >even know exists. In principle, once someone has established cognate relationships, they can indeed write an algorithm to derive their original data from their reconstruction. But that is the REVERSE of what Mark envisions. Thinking about the nature of the discovery procedure imagined by Mark we take all the words (from dictionaries -- but shouldn't it be morphemes?) in all the languages of the world and run all possible correspondences on them (whatever "all" means in this context, e.g., k =t, k= p , k = a, k = @@ etc etc etc) and then choose the ones that work out best (semantically, phonetically and in terms of number of cognates per language pair -- or whatever, AND apply apppropriate statistical tests to make sure the "best" is above chance, AND fight like hell against anyone who dares claim "borrowing", "wander-words" and all that esoterica, in cases where statistical tests are not overwhelmingly decisive etc etc) and VOILA. I wonder if Mark realises what is involved in such a program and how pitifully simple the chess program that beat the world champion is in comparison (with its evaluation of 3 gazillion positions per minute or whatever it was). Because of what can be done a posteriori, I don't dismiss the vision out of hand -- but in view of what's really involved, we'll be arguing the way we do now for generations to come, before anything remotely resembling Mark's vision emerges for any uses other than cryptography. He ends with the usual premise of science fiction writers, and pop accounts about scientist-heros and their breakthroughs: >It's too bad that the attitude of most linguists is, in fact, the most >damaging to themselves and their own professions. But, that is the way >evolution is. Short term goals and intuition only go so far. In other words, the "establishment" are the usual and familiar small-minded villains that hinder progress, and are too vested in their own positions to recognise a good idea when it is shoved up their .... Not me, Mark. I'm open-minded. So, explain how this brute-force program works better than I just did above, and how it addresses all those issues I compressed into my description. Can it also decipher Linear A etc. (or do I mean "translate")? P.S. Anybody. What is the promise of programmed cryptographic methods for discovering "non-obvious" sound correspondences and-so-on among the lexica of different languages P.P.S. As opposed to cryptography, doesn't the programmed "cognate" search have to start with semantic features (and then the gazillions of brute attempts at sound correspondences)? How should a semantic feature search be organised for big-time time-depths (cf. Johanna's remarks on "five" different meanings for a cognate, e.g., "fly", "wing", "feather", "fur", etc. "five" is a very rough count. Cognates or assumed cognates exhibit different DEGREES of semantic resemblance, from obvious to "far-fetched", e.g., "night", "bump into things in the dark", "get knocked out", "sleep", "lie down", "sexual intercourse", "AIDS", "dirty needles", "gay", "happy", etc. That also has to be modelled for statistical measurement, e.g., "wing" is "closer" to "feather" than to "fur", etc. -- but "fur flies", at least in English.) PPPS. Mark wrote: >...most of the IE words could be due to the substratum which >could have been a family. One can always insist that the reason why IE >words resemble each other is because they are all left over from a >previous language which was spread out over the same region. Isn't that like saying: The "Odyssey" wasn't composed by Homer but by somebody else with the same name (or whom we choose to call Homer -- yeah, I know, but he wasn't blind). If not that, then Mark is proposing an extremely complicated and unlikely hypothesis, in which case in the absence of contrary evidence we choose the simpler one ("science" actually makes decisions like that -- in the absence of contrary evidence, I said.) I'm not even sure that I can imagine what the complicated hypothesis is in this case. Maybe something like, IE had fragmented into Germanic and Slavic etc, but the current languages classified as Gmic, Slvic etc don't descend from them but just happened to borrow most of their vocabulary and grammar from them. So, then, BY WHAT CRITERIA did these languages borrow so much from their IE neighbors rather than give up their own presumably non-IE languages and ADOPT those IE languages, so that they are INDEED descendants of IE. From bwald at HUMnet.UCLA.EDU Sat Nov 14 20:35:18 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Sat, 14 Nov 1998 15:35:18 EST Subject: intervocalic DEvoicing can also happen / X > Y > X Message-ID: ----------------------------Original message---------------------------- I welcome Alan King's message. On his first point: >B. Wald says: It just happened -- because both directions >>are possible (under certain conditions -- certainly NOT ****z > s /V_V). >That provoked me to try and think of counterexamples to Benji's latter >assertion here. Good. If I'm wrong, I'd like to know as soon as possible, so that I don't keep on thinking that and misleading other people. That's my best pay-off for my interventions on the list. He continues: >I couldn't come up with an example of z > s /V_V, but I >did find one of +voice > -voice /V_V. Just as good. I would have thought that if such a change occurs it can't be due to the intervocalic position (just as in the z > s "reversal" case), and would expect it to occur in that position only as part of a more general voicing shift (which the intervocalic position is not able to resist at the time). A short digression. He asked: >(Okay, I'm going to be a pain and say it: just what is "Central Spanish" >supposed to mean? The phenomenon in question includes the WHOLE of >Castilian (i.e. Spanish) in its territorial extension. I don't know. In my insecurity as a non-expert, and without checking, I wanted to rule out some non-Central (northern) dialects, without being sure which ones, maybe Leonese or Aragonese or whatever, where maybe z > s did not occur, at least at first (since I remembered it is said to have been observed first in the "Central" dialects like Castille -- and not to have taken effect until, say, after the Jews were expelled, so that Ladino still has-z- in "casa" etc.) To get back to the main point, Alan proposes an example from some (southern) Welsh dialects. I'm not sure I understood the entire discussion, but the situation seems to be as follows: >I have put quotes around "voiced" and "voiceless" because >phonetically all stops in Welsh tend to be voiceless (or at least voicing >is not critical) and the contrast is realized principally in terms of >tenseness and aspiration. To simplify the exposition I shall henceforth >largely ignore that fact in the transcriptions and terminology used.) Alan is referring to STANDARD Welsh here, and that by "critical" he must mean relevantly that the degree of voicing of the "tense" stops varies. The issue then becomes whether this is a novel situation or reflects an older situation in which there were consistently voiced stops that have now come to vary in voicing as they shift to tenseness as their primary (or distinctive) feature. The issue is also whether this shift, if it is so, is ONLY intervocalic (post-stress) or whether it is general to the stop paradigm (e.g., pre-stress, initial, etc.). He continues: >... in Northern spoken Welsh, this system has >been altered, in that nearly all consonants in the intervocalic post-stress >position are pronounced geminate (and correspondingly, all stressed vowels >followed by a consonant in non-final syllables are short - but I'm going to >focus on the consonants here). So corresponding to standard ['ka:dajr] we >will find ['kaddar] in northern Welsh, while ['etto] shows no change since >the consonant is already geminate. Very interesting. I would assume that the gemination of voiced stops represents an interruption of voicing between vowels, and that even the shortening of a long vowel before a geminate voiced stop anticipates such an interruption. In any case, that does seem to me a mechanism by which intervocalic voiced stops could devoice -- and to the extent that gemination is strictly an intervocalic phenomenon in Welsh the stage is indeed set for devoicing of the intervocalic stop. Finally, he notes: >Now: apart from this gemination and shortening, but possibly related to it, >we find some varieties of Welsh in which, corresponding to consonants like >the /d/ in , if the consonants in question are "voiced" stops in >standard Welsh, we find "voiceless" stops instead: ['katar] etc. Assuming >once again that the standard form is diachronically prior, we then have d > >t (or possibly d > dd > t) in post-stress intervocalic position, and >similarly b > p and g > k. Irritatingly, this phenomenon has been noted in >the south, in an area not contiguous to the northern "geminating" varieties. In which case, it is commendable that Alan suggests that the route assumed may have been a continuation of the one still evident in the north. I need to reflect further on this case, but it seems plausible to me on the face of it. What gives me most pause is whether this "counts" as STRICTLY intervocalic position, if there is the assumed geminate intermediate stage, and it just so happens that geminates only occur intervocalically -- so that it might be interpreted as a shift of voicing in GEMINATES, regardless of their position -- which just happens to be intervocalic. I'm also not sure I followed the entire thing. For example, since there were also geminate VOICELESS intervocalic stops in these dialects of Welsh, was there a merger, or did the "voiced" geminates de-geminate before "fully" devoicing? i.e., -tt- vs. d > dd > *td* > t (where somehow *td* is something that avoided merger with -tt- as it devoiced.) The phenomenon is certainly interesting in its own right, and deserves further attention, possibly further dialectal research to examine what can be discovered about the details of the path. This is particularly important if the result is -tt- vs. -t- (< -tt- vs. -d-), since Alan's suggestion seemed to be that gemination had something to do with the devoicing. So far the evidence in the south does not indicate that, but only that, for some reason, -d- came to move into a space parallel to -tt- by devoicing. If that's the case, it is indeed a counter-example to the notion that stops cannot devoice intervocalically without first going through some other changes (like gemination and associated metric effects on syllabification). I appreciate Alan's example, and hope that further clarity will emerge on the historical changes represented in it. PS. I forgot whether Alan mentioned this, but if a form like "chair" had a long vowel before the stop originally, as in the standard, and shortened it before gemination in the north, then the shortening in the south may also suggest prior gemination. Obviously we need to know a whole lot more about the history of the relevant Welsh dialects before we can draw secure conclusions. From larryt at cogs.susx.ac.uk Sat Nov 14 20:35:44 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Sat, 14 Nov 1998 15:35:44 EST Subject: the Trask-Hubey debate In-Reply-To: <364BC8D4.743917C0@montclair.edu> Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, H.M.Hubey wrote: > Any piece of paintwork does not count as painting. The painters have > their own black magic guild which decides what is art and what is > not. Does that ring a bell? Nope. What we try to reconstruct is not individual miscellaneous items, but rather the whole system of a vanished language, or as much of that as we can. A proposed reconstruction is accepted by other linguists to the extent that it (a) is internally coherent and (b) accounts successfully for the data. Aesthetic factors like symmetry and economy are not negligible, but they must take a back seat to our major concern: how well does the proposed reconstruction account for the data in a principled manner? > I also have very small doubts that I can write a program that can take > n words from language A, and m from language B and write a program that > can change one set into the other (at least enough of them to dumbfound > the skeptics and force them to having N new looks into the "comparative > method") using only regular sound changes. The only problem is that I > don't have the time or energy to put into such a useless demonstration. I think "useless" is a very apt description. No doubt one could write a cutesy program that would convert, say, a set of English words into Zulu words, but such activity bears no discoverable resemblance to comparative reconstruction or to any other aspect of historical linguistics. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at cogs.susx.ac.uk Sat Nov 14 20:36:08 1998 From: larryt at cogs.susx.ac.uk (Larry Trask) Date: Sat, 14 Nov 1998 15:36:08 EST Subject: Stabilized languages In-Reply-To: Message-ID: ----------------------------Original message---------------------------- On Fri, 13 Nov 1998, Roger Wright wrote: > Change is at some times slower than at other times, though. Agreed, of course, but it never vanishes in a living language. > And it just isn't true that "dead" languages don't change. Medieval > Latin varied astonishingly widely in time and space, Renaissance > Latin was recognizably different again, and even the modern Latin > used by the Vatican has been continually acquiring new vocabulary. Agreed, but I wasn't claiming that a dead language can never change, but rather that absence of change is only possible in a dead language. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mccay at redestb.es Sat Nov 14 20:36:37 1998 From: mccay at redestb.es (Alan R. King) Date: Sat, 14 Nov 1998 15:36:37 EST Subject: help with Finnish, please? Message-ID: ----------------------------Original message---------------------------- This may or may not be deemed a suitable question for HISTLING, but I'm going to try anyway. I am translating/editing an article on Basque language standardization, and would like to double-check on the accuracy of a few statements by the original author concerning Finnish words and their etymologies. Anyone want to (and able to) help? I will acknowledge help (if ultimately used) in a footnote to the translated version. The author says that: Finnish PEILI "mirror" is the assimilated form of a loanword SPEKLI; Finnish ROUVA "lady" similarly < FROUWA; Finnish ROPAKANTA "propaganda" < PROPAGANDA; Finnish RANTA "?" < STRAND; Finnish STRUKTUURI < STRUKTURA. The author does not state what language(s) the second forms given came from, but I would like to know (I'm thinking here of SPEKLI and FROUWA in particular; is this some sort of proto-Germanic, or what? should they be starred? and of course, are they accurate?). Kiitos! Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From mccay at redestb.es Sat Nov 14 20:38:59 1998 From: mccay at redestb.es (Alan R. King) Date: Sat, 14 Nov 1998 15:38:59 EST Subject: intervocalic devoicing in Welsh (?) Message-ID: ----------------------------Original message---------------------------- In response to Benji Wald in response to me on a possible example of intervocalic devoicing in certain varieties of Welsh, I unfortunately cannot add very much as that is about the limit of my knowledge on the particular subject (and I do hope the information I gave is largely accurate). From here on one would need the help of a Welsh *specialist* (I just happen to know some Welsh and have an inquisitive mind, but have not studied the subject in detail). So regarding Benji's PS at the end: >PS. I forgot whether Alan mentioned this, but if a form like "chair" had a >long vowel before the stop originally, as in the standard, and shortened it >before gemination in the north, then the shortening in the south may also >suggest prior gemination. Obviously we need to know a whole lot more about >the history of the relevant Welsh dialects before we can draw secure >conclusions. I can only agree with the second sentence, and point out with regard to the first that I don't know, off-hand, what length the vowel has in the first syllable of the word for "chair" (standard CADAIR) in varieties with the devoicing phenomenon; if the transcription I gave implied that it is short, that was a slip-up. Since southern dialects in general maintain vowel quantity distinctions, my initial assumption must be that it is probably long. Furthermore, when a Welsh specialist is found, she or he should be questioned on the exact dialectal distribution of the devoicing phenomenon. I know it does occur in the southeast, but can't remember whether or not it is also found further north, in which case there might be a dialect with both devoicing and vowel shortening. Bearing in mind my disclaimers, I'll be glad to give any related information about Welsh if needed and I feel competent to do so. Hwyl, Alan From mcv at wxs.nl Sat Nov 14 20:39:38 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Sat, 14 Nov 1998 15:39:38 EST Subject: s > z (etc.) (Italy, Spain) In-Reply-To: <28111321133870@vms2.macc.wisc.edu> Message-ID: ----------------------------Original message---------------------------- Tom Cravens wrote: >Comments below (Tom Cravens) >(Note: I [TC] mistakenly sent this directly to Miguel, rather than the list. >Perhaps the discussion of weakening by class is of some general interest.) And this is how I replied, and what Tom replied to my replies while we were sorting out where the message should have gone... I trust it's allright with Tom to compress things this way, as his final reply to me (though private in principle) does not contain information more sensitive than that he attended to a party in Perugia in his student days (who hasn't?), and does contain a couple of useful references of general interest. [TC:] >>Unless I'm mistaken (I don't have the volume at hand), Rohlfs is talking >>about Tuscan in that passage, not Standard Italian. [mcv:] >For v, d, g he's talking about "la lingua nazionale" (and then >separately about Toscana, N, C and S). For s, he starts discussing >Tuscan right away. Since the written language does not distinguish >/VsV/ from /VzV/, both written , in a sense there is no >"standard" pronunciation. [TC:] Well, there is, sort of. At least prescriptivists try to enforce it (more or less vainly). But you're right, it's certainly not consistent. A striking anecdotal-but-real indication that (some) speakers don't worry about it, though: many years ago while a student in Perugia, I was at a party where there was a young woman addressed by some as Lui[s]a, others as Lui[z]a. My Italian was very elementary at the time, but I had a good ear, so I finally asked her if her name was Lui[s]a or Lui[z]a. The answer was the classic Italian "E` lo stesso." [mcv:] >At least the spelling doesn't enforce any, >like it does in the case of the stops, so there's logic in that. [TC:] Right. [mcv:] >>>According to Rohlfs, the cases of /v/, /d/, /g/ and /z/ (listed in >>>approxiamte order of frequency) are not native Tuscan forms, but >>>words imported from the north. [TC:] >>Yes, Rohlfs says so, but this bit isn't that simple, either. [mcv:] >I thought as much. I didn't stop to re-read the whole chapter called >"Ricapitolazione critica dello sviluppo de -k-, -p-, -t-, -s- in >Toscana", but I noted a touch of controversy. Usually happens when >things are not that simple. [TC:] Yes, Rohlfs and a few others were once pitted against Merlo and very few others, in trying to enforce a strict Neogrammarian all-or-none interpretation. The effects linger, with the Rohlfs school having gained the upper hand. But a detailed investigation suggests that something is very much awry. Martin Maiden has a nice, if very brief, treatment in his Linguistic history of Italian, and Luciano Giannelli and I address this rather superficially in Maiden and Parry (eds). 1997. The dialects of Italy. Routledge. [TC:] >[Alto Aragones geminate /nn/, /ll/...] [mcv:] Interesting. I had completely overlooked that (it is mentioned, although very briefly, in my copy of "Dialectologma espaqola" by Alonso Zamora Vicente). [and I have nothing further to add at this time] ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From ratcliff at fs.tufs.ac.jp Sat Nov 14 20:47:39 1998 From: ratcliff at fs.tufs.ac.jp (Robert R. Ratcliffe) Date: Sat, 14 Nov 1998 15:47:39 EST Subject: reconstruction methodology Message-ID: ----------------------------Original message---------------------------- H.M.Hubey wrote: > ----------------------------Original > message---------------------------- > Robert R. Ratcliffe wrote: > > > > How do we test it then?-- By > > implication. Each reconstruction (of a proto-phoneme for example) > has > > implications for the whole system of the proto language (the whole > > phonological system, eg), for the development path leading from the > > proto-language to the attested languages (the sequence of sound > changes, > > eg), and for the forms of the reflexes in the descendant languages. > > Only the last is directly observable, of course, and only this real > data > > can be used to rule out a proposed reconstruction absolutely. > > But here is where the iteration comes in. The first attempt at > reconstruction of a protolanguage *X will be based on N languages. > If we add the (N+1)st language then *X might have to be changed. That's right. It happens all the time. I'll quote something to the point:"We find in this development an exact parallelism to the manner in which scientific ideas generally arise, develop and change. They are created to point out the common part in a variety of observed phenomena... At first almost any idea will do.. afterwards, the inconsistencies of the first trial make themselves felt; the first idea is then changed to meet better the new requirements. For a shorter or longer time the facts and ideas may remain in accord, but the uninterrupted increase in empirical knowledge involves sooner or later new fundamental alterations of the general idea, and in this way there is a never-ending process of adaptation of ideas to facts." (Wilhelm Ostwald, from the article on chemical element in the Encyclopedia Brittanica, 1911) > > We might find another language y to add to the family. How many > correspondences do we need? > > Even worse, if the similarity of the language to other languages is > not considered, it will be added to the most established, largest > family, and it will continue to snowball. I don't believe this happens very often. It generally seems to take a long time before specialists in an established family are willing to accept newly described or newly discovered languages as members of the family. It was a while before Hittite was accepted as IE. And it has certainly taken a long time for Semitists to recognize that their family is simply part of a larger family whose other members are found in Africa. +++++++++++++++++++++++++++++++++++++++++++++++++ Robert R. Ratcliffe Senior Lecturer, Arabic and Linguistics, Dept. of Linguistics and Information Science Tokyo University of Foreign Studies Nishigahara 4-51-21, Kita-ku Tokyo 114 Japan From DISTERH at UNIVSCVM.SC.EDU Sat Nov 14 21:24:33 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Sat, 14 Nov 1998 16:24:33 EST Subject: discussion ended Message-ID: Dear Colleagues, By now the discussion about phonological reconstruction and mathematical models has reached the limits of its usefulness to the entire list. I will no longer post discussions on this topic. However, the few people that are involved in this discussion may wish to continue it off-list. I apologize for any inconvenience this may cause you. Dorothy Disterheft Moderator, HISTLING From bwald at HUMnet.UCLA.EDU Tue Nov 17 00:18:57 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Mon, 16 Nov 1998 19:18:57 EST Subject: intervocalic devoicing in Welsh (?) Message-ID: ----------------------------Original message---------------------------- In agreement with Alan King's last message on Welsh intervocalic devoicing, I hope that some Welsh experts will indeed enlighten us further on the situation. Meanwhile, the possibility of the phenomenon raised a question of a more general nature in my mind. I embedded it in my last message, but I would like to restate it more elegantly. It concerns the phonological conditions on sound change in the following way. Suppose, for the sake of argument, that one or more intervocalic voiced stops geminate. Let's express this as: [+voiced] > [+geminate] / V [+stop,__] V Maybe that's already odd, if [+geminate] implies an INTERRUPTION of voicing ONLY in intervocalic position. Presumably, however, this change is attested in Northern Welsh. Thus, odd or not, it becomes an interesting fact about possible sound change under some conditions. It's really the next stage that raises the issue that I'm most interested in. Suppose that next, the voiced geminate devoices (somehow avoiding merger with a previously existing voiceless geminate -- that raises all kinds of questions, but that's not my main concern here). My main concern is that since (or if) geminates only occur in intervocalic position, then we can express the sound change simply as: [+geminate] > [-voiced] The point is that there is no need for a condition. This could be a trick of economical notation, OR it could be a claim that if there WERE non-intervocalic geminates they would also devoice. I suggest that the sound change is the latter (it's not a notational trick). In that case, it is not "intervocalic" devoicing, where "intervocalic" implies a conditioning factor. That is, intervocalic position has nothing to do with the sound change, even though the change only has the OPPORTUNITY to occur in intervocalic position. There are plenty other cases of such a sequence of sound changes, abstractly: 1. X > Y /W__Z conditioned 2. Y > Q unconditioned But what other cases are there of sound changes where a. Y only exists in context W__Z AND b. X > Q / W__Z is an UNEXPECTED sound change. From mccay at redestb.es Tue Nov 17 13:01:29 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 17 Nov 1998 08:01:29 EST Subject: intervocalic devoicing in Welsh (?) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- >In agreement with Alan King's last message on Welsh intervocalic devoicing, >I hope that some Welsh experts will indeed enlighten us further on the >situation. I indeed hope so, but if nobody volunteers and the matter is of sufficient interest I *could* dig into my books and/or bother a colleague or too to get some more information. I would rather not put so much time into this, though. >Meanwhile, the possibility of the phenomenon raised a question >of a more general nature in my mind. I embedded it in my last message, but >I would like to restate it more elegantly. It concerns the phonological >conditions on sound change in the following way. > >Suppose, for the sake of argument, that one or more intervocalic voiced >stops geminate. Let's express this as: [+voiced] > [+geminate] / V >[+stop,__] V Actually if we're looking at the northern Welsh model, the [+voiced] bit is strictly unnecessary. The change can be represented simply as: [+stop] > [+geminate] / V __ V, which I should think is intuitively more plausible. Of course, this occurred under the circumstance that voiceless stops in the specified position were already geminate (for diachronic reasons that we've already been into). Again, relevant to the Welsh phenomenon is the fact that the gemination probably only occurs when the preceding vowel is stressed, so another modification is necessary: [+stop] > [+geminate] / [V, +stress] __ V. I would suggest there are two ways of "understanding" this change. One is just in the terms implied by the foregoing notation: northern Welsh speakers "decided" to geminated all stops between a preceding stressed and a following (unstressed) vowel. "For fun", so to speak. The other view might be that since voiceless stops in that position were already always geminate, they "decided" to *simplify* their phonological system by extending the "gemination habit" to all stops, regardless of voicing. Actually they went further than that. Gemination in the said position was extended to *most* consonants; notably including /n/, /l/ and /r/. At this point we might say that they were really getting "carried away", since length in these particular consonants (only) had until that time been phonemic, as it still is in southern Welsh (although phonetically most of the work to distinguish short and long liquids is probably performed by the preceding stressed vowel through a compensatory length contrast). So far from doing such a strange thing as geminating intervocalic voiced stops for no apparent reason, we could say that they pretty much had a "gemination party", geminating nearly everything they found in the position / [V, +stress] __ V. Why? I don't know if it's a motivation or merely an effect, but the outcome of this development is a very characteristic *staccato, almost syncopated* rhythm to northern Welsh speech, since most stressed syllables are pronounced with a short, rather energetic vowel followed by a long and also energetically articulated consonant which seems to go implosion-interruption-explosion. Interestingly, in some speakers at least, the intensity of the stressed and the posttonic unstressed vowels doesn't seem to contrast as much in this "articulatory style" as in, say, English or even Spanish (so that foreign ears may have difficulty interpreting which is the stressed syllable), and it now occurs to me that this may be explained if we assume that the "staccato-ey" features I've described have taken over the function of identifying the position of the stress. In particular it is common for the pitch of the posttonic syllable to be higher than that of stressed syllable. As I already said previously, stops in general tend to be strictly voiceless in pronunciation, with other features such as aspiration doing most of the work of differentiating the voiced and voiceless series. It seems to me that in this form of speech intervocalic geminate stops are *particularly* voiceless: i.e. voicing is interrupted along with everything else in their intervocalic articulation. I believe that the voiced and voiceless series nonetheless remain phonologically fully differentiated in this pronunciation (although some foreign speakers might be excused, again, for getting the signals wrong). But *phonetically* one might argue that you already have intervocalic devoicing here, at least as a subsidiary effect - assuming of course that some voicing is there in the voiced stop series to begin with. And if that is the case, it appears to me that we might describe the phenomenon as one of dissimilation. There are surely enough precedents for that is phonology? When intervocalic stops get voiced, that is assimilation to the voicing of the surrounding vowels, I take it; why then, when the opposite happens it is surely dissimilation. The motivation might be to heighten the contrast between adjacent segments: vowels voiced, consonants unvoiced. Maybe we should look for other instances of *that* phenomenon? Please remember, I repeat, that I am mostly describing my "non-rigorous" *impressions* in the preceding description. I am not by training a phonetician (nor a historical linguist, for that matter). Caveat emptor. Regards, Alan From eska at vtaix.cc.vt.edu Tue Nov 17 16:32:31 1998 From: eska at vtaix.cc.vt.edu (Joseph F. Eska) Date: Tue, 17 Nov 1998 11:32:31 EST Subject: On Welsh voicing Message-ID: ----------------------------Original message---------------------------- I don't have the time at the moment to enter the discussion fully, but one thing to bear in mind is that the phonetic realisation of phonemes is not identical across languages. In Welsh, broadly speaking, phonemic /p b/, for example, are [p^h p] phonetically, i.e., the contrast is one of phonetic aspiration, rather than voicing. There is no process of intervocalic devoicing in Welsh, but there is one of intervocalic de-aspiration for voiceless stops, which, in many dialects, is accompanied by partial voicing. Joe Eska eska at vtaix.cc.vt.edu From P.G.Honeybone at newcastle.ac.uk Wed Nov 18 21:06:33 1998 From: P.G.Honeybone at newcastle.ac.uk (Patrick Honeybone) Date: Wed, 18 Nov 1998 16:06:33 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- In connection with the recent discussion on the possibility of intervocalic devoicing in general, and whether this happened in some dialects of Welsh in particular, Joe Eska wrote: > I don't have the time at the moment to enter the discussion fully, > but one thing to bear in mind is that the phonetic realisation of > phonemes is not identical across languages. In Welsh, broadly > speaking, phonemic /p b/, for example, are [p^h p] phonetically, > i.e., the contrast is one of phonetic aspiration, rather than > voicing. There is no process of intervocalic devoicing in Welsh, but > there is one of intervocalic de-aspiration for voiceless stops, > which, in many dialects, is accompanied by partial voicing. I think that this is an important point, and although Alan King referred to it a couple of times in some of his contributions (e.g... > As I already said previously, stops in general tend to be strictly > voiceless in pronunciation, with other features such as aspiration > doing most of the work of differentiating the voiced and voiceless > series. ... the implications of it weren't taken up. So I thought I'd add my thoughts. If this is right (and I'm afraid I know woefully little about the phonology of Welsh myself, so can't comment) then it fits in with an approach to laryngeal features like [voice] and [aspiration] that's been firming up recently. This is the idea that in a language with a two-way distinction in stops (and other obstruents, too, probably, but most of the work seems to centre on stops), that distinction could be made in two ways phonologically. Either the language uses [voice] *or* it uses [aspiration]. This fits in well with phonological theories that use privative features, but could also be understood as using [+/- voice] *or* [+/- aspiration]. (I should perhaps say at this point that there is some variation among proponents of the idea in terms of the names that they give to the features; some use [aspiration], some [spread glottis], some [tense] and some (in Government Phonology) 'H' but I think this is not really a substantial difference). So, using a system like this, people tend to say that most standard Germanic languages (but not Dutch or even all the dialects of English or German) characterise the difference using [aspiration] (or whatever), whereas Romance languages and Slavic tend (again - not necessarily all of them) to use [voice] instead. The idea also has an obvious connection with the long tradition of using [fortis/lenis] as a feature (-pair) rather than [+/- voice] in the Germanic literature. (A good reference for all this is: Iverson, G. & Salmons, J. (1995). 'Aspiration and laryngeal representation in Germanic'. Phonology 12. 369-396). This means that the conventional transcriptions like /t/ and /d/ can be misleading, because in, say, German, /t/ involves [aspiration] and /d/ is unspecified (in a privative system) whereas in, say, French, /d/ involves [voice] and /t/ is unspecified. So /t/ vs. /d/ in German is not the same opposition as /t/ vs. /d/ in French. And if what Joe Eska and Alan King wrote about Welsh stops is right, then Welsh looks like it uses [aspiration] and not [voice]. Welsh /t/ > /d/ would not be quite the same kind of change as Romance /t/ > /d/, nor would a putative Welsh /d/ > /t/ be quite the same as a putative Romance /d/ > /t/ (which is sort of where we came in). If this is all on the right lines then the different kinds of stops in the different kinds of languages should have different phonological effects (and they do seem to) and might well allow different predictions for the kinds of lenition that the languages could undergo (and I think they do...). The point of what I'm trying to say is: maybe we should be careful about comparing changes in different languages - the symbols we use to transcribe sounds could be pulling the wool over our eyes. But I do have more - if I could just go on for a bit longer... This discussion on types of lenition has been very interesting and one key point is evident in the following exchange between Alan King and Benji Wald: >>B. Wald says: It just happened -- because both directions >>are possible (under certain conditions -- certainly NOT ****z > s >>/V_V). That provoked me to try and think of counterexamples to >>Benji's latter assertion here. > >Good. If I'm wrong, I'd like to know as soon as possible, so that I >don't keep on thinking that and misleading other people. That's my >best pay-off for my interventions on the list. > >He continues: > >>I couldn't come up with an example of z > s /V_V, but I >>did find one of +voice > -voice /V_V. > >Just as good. I would have thought that if such a change occurs it >can't be due to the intervocalic position (just as in the z > s >"reversal" case), and would expect it to occur in that position only >as part of a more general voicing shift (which the intervocalic >position is not able to resist at the time). I've just read an article by Ernst Jahr which seems relevant to this point [Jahr, E. H. (1989) 'Language planning and language change' In: Breivik, L. E. & E. H. Jahr 'Language Change. Contributions to the Study of its Causes'. Berlin: Mouton de Gruyter]. In it, Jahr reports on a change that looks like [+voice] > [-voice]/ V_V, or at least, [+voice] > [-voice]/V:_V, in 'upper-class' Oslo Norwegian. The basic details are: in 1880, in words such as the following, with a long vowel, stop, vowel sequence (and I think it basically really is just this environment), the stop was voiced in the pronunciation of this group of people, whereas now the stop is voiceless: 1880 now 'creator' [ska:b at r] [ska:p at r] 'know' [vi:d@] [vi:t@] 'realm' [ri:k@] [ri:k@] ('@' = schwa) So this looks like it might be an example of the 'unnatural' change that Benji Wald was looking for. But Jahr connects this with the language planning movement in Norway and the connected change in the spelling of the words (they used to have , and , now they have

, and ), so it wouldn't count as a 'normal', 'real' sound change. *And anyway* Norwegian, as far as I'm aware, has aspiration in 'voiceless' stops and little or no voicing in 'voiced' stops, so the symbols might be confusing, and ... I think what I'm trying to say is, even if we did find an example of [+voice] > [-voice]/ V_V it might be explicable in other ways and not be a counter-example to lenition tendencies. From manaster at umich.edu Thu Nov 19 18:08:38 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Thu, 19 Nov 1998 13:08:38 EST Subject: Arm. targal 'spoon' (fwd) Message-ID: ----------------------------Original message---------------------------- Dear historical linguists, I have as some of you may know proposed an Armenian sound law whereby inter alia PIE *w ends up ultimately as j^ after *r(H) or *l(H) in Armenian, e.g., olj^ is then simply *solwo-. The only counterexample that I am aware of is targal if this is really, as many think, from *drwaHlaH or the like. So the obvious question is can anyone think of an alternative etymology. One (maybe even the) possibility would be some proposal whereby targal is from *tarigal or *tarugal with the regular Armenian syncope of the relevant vowels. Any takers? AMR From mccay at redestb.es Thu Nov 19 17:55:05 1998 From: mccay at redestb.es (Alan R. King) Date: Thu, 19 Nov 1998 12:55:05 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- Having read Patrick Honeybone's latest contribution to the discussion I (rather unwittingly) triggered off, I have probably little of substance to add, but just a couple of related questions and thoughts. Just for clarification, if anyone cares: I am not a specialist in the subject of this discussion (as I've said several times already), and furthermore I haven't got a theory to defend, I'm more of an onlooker sitting on the sidelines and throwing in an idea or two, for what (if anything) it's worth, from time to time. And as such, if I make any silly mistakes, I expect to be corrected. On the question of aspiration versus voicing in Welsh: >I think that this is an important point, and although Alan King >referred to it a couple of times in some of his contributions (e.g... > >> As I already said previously, stops in general tend to be strictly >> voiceless in pronunciation, with other features such as aspiration >> doing most of the work of differentiating the voiced and voiceless >> series. > >... the implications of it weren't taken up. I confirm once again, that according to every phonological description I've ever seen, as well as my own acoustic impressions (the effect is unmistakable, at least if you know a minimum about phonetics and your ear is accustomed to hearing a language with a *voicing* opposition, as is my case): in Welsh stops there is definitely a contrast, and no doubt an opposition, in terms of aspiration, while voicing would seem to be secondary (if the contrast is there at all in phonological terms). In most contexts all stops sound pretty voiceless. The proof of the pudding is that there is, or was, an important colony of (originally, anyway) Welsh speakers, dating from a hundred years ago, in the Patagonia region of Argentina. Believe it or not (this is such a fascinating story, I can't resist telling it), the original emigrants are said to have moved in an organised attempt to ensure that the Welsh language remained alive *somewhere* in the world even if it disappeared in Wales, and had the permission and complicity of the Argentinean government to colonise a region of pampa (and some indigenous inhabitants, it should be added) to this end. Life being what it is, nowadays Spanish seems to have taken over most functions in Welsh Patagonia, which has been largely Argentineanized by now, but there is still a significant residue of maintenance of group identity and even some linguistic maintenance. Predictably, the Welsh back in Wales find all this very interesting, and some of the Patagonians are drawn to their ancestral homeland, so some ties exist. Okay, back to Welsh stops. I've never actually met a Patagonian, but from what I've heard, the Welsh in Wales have a notion of what a "Patagonian" accent is, and one of the most salient features by which "you can tell them apart immediately" is their pronunciation of the stops, since the more recent generations, who are Spanish speaking, apparently assimilated the Welsh system to the Castilian one. Presumably the younger Patagonians have no difficulty with proverbial obstacles to the acquisition of Welsh pronunciation such as its famous voiceless lateral fricative, but are unable to aspirate a /p/! (Or more likely, are unable to *not* voice a /b/ while still keeping it distinct from a voiceless aspirated /p/.) By several reports, native speakers of Romance languages who learn to speak Welsh fairly well all sound like "Patagonians" to the native Welsh. >And if what Joe Eska and Alan King wrote about Welsh stops is right, >then Welsh looks like it uses [aspiration] and not [voice]. Welsh /t/ >> /d/ would not be quite the same kind of change as Romance /t/ > /d/, >nor would a putative Welsh /d/ > /t/ be quite the same as a putative >Romance /d/ > /t/ (which is sort of where we came in). > >If this is all on the right lines then the different kinds of stops in >the different kinds of languages should have different phonological >effects (and they do seem to) and might well allow different >predictions for the kinds of lenition that the languages could undergo >(and I think they do...). Maybe we should try to be more explicit about the kinds of change we are predicting to be possible in each kind of phonological system. Let's take this set of four logical possibilities for the kind of distinction in stops and the theoretical possibility of intervocalic "lenition" (admitting, for convenience, that "lenition" might, until demonstrated otherwise, be either voicing or deaspiration) and "de-lenition" (with apologies: I just made the word up, to denote the reverse change): AC: voicing distinction in stops intervocalic stops may lenite AD: voicing distinction in stops intervocalic stops may de-lenite BC: aspiration distinction in stops intervocalic stops may lenite BD: aspiration distinction in stops intervocalic stops may de-lenite Which of these conditional relations between the type of distiction in stops and the possibility of a certain type of change are we hypothesizing to be true? If I understand him, Patrick Honeybone suggests the following: +AC, -AD, +BD (i.e., in voicing systems, stops may lenite but cannot de-lenite, while in aspiration systems, stops may de-lenite). He doesn't say anything about BC: in aspiration systems is lenition also admitted? I would think it both theoretically and methodologically interesting to encompass this question in such a hypothesis: it would certainly give you more predictions to test. On the other hand, I think there is a danger in being excessively categorical in matters like this, things may not be so black and white. One way in which this is true was pointed out by Patrick Honeybone in the last point in his message: >The basic details are: in 1880, in words such as the following, with a >long vowel, stop, vowel sequence (and I think it basically really is >just this environment), the stop was voiced in the pronunciation of >this group of people, whereas now the stop is voiceless: > > 1880 now >'creator' [ska:b at r] [ska:p at r] >'know' [vi:d@] [vi:t@] >'realm' [ri:k@] [ri:k@] > >('@' = schwa) > >So this looks like it might be an example of the 'unnatural' change >that Benji Wald was looking for. But Jahr connects this with the >language planning movement in Norway and the connected change in the >spelling of the words (they used to have , and , now they >have

, and ), so it wouldn't count as a 'normal', 'real' >sound change. *And anyway* Norwegian, as far as I'm aware, has >aspiration in 'voiceless' stops and little or no voicing in 'voiced' >stops, so the symbols might be confusing, and ... I think what I'm >trying to say is, even if we did find an example of [+voice] > >[-voice]/ V_V it might be explicable in other ways and not be a >counter-example to lenition tendencies. Precisely. Examples such as this one, which is very nice, give me much pause for thought. However, I would emphasise that this sort of problem is by no means limited to cases where language planning as such has intervened. I am reminded of the (only apparent) enigma in the evolution of modern English vowels. I don't remember the details, but around about he seventeenth (?) century there is every indication that the vowels in "sea" and "say" had merged, while that in "see" remained distinct; whereas in present-day English, the first two have miraculously become differentiated once again, and it is "sea" and "see" that are homophonous. Now as far as I know the only way to solve this paradox must involve the fact that languages are heterogeneous systems at many levels; in other words, historical sociology must be invoked of necessity. It seems fairly obvious that what happened is that there were several phonological varieties of English in coexistence and, at some point, competing for acceptance as the prestige and/or "official" model. My own sobering conclusion is that, in any language at any time, beneath the "calm" surface of a language's "general" description, there are undoubtedly undercurrents that may be far more turbulent (and which seriously affect the fishing, I expect!). So one question I'd like to ask is: okay, present-day Welsh stops (Patagonian Welsh excepted) have an "aspiration" system. But for how long has this been the case in the history of Welsh? I myself have no idea. What I do know, partly thanks to Patrick Honeybone, is that, for example, some Germanic languages are "aspirating", but some are "voicing"; it follows that in the history of the various Germanic languages, some must have shifted from one type to the other. In which languages, when, and what type do we think Proto-Germanic was? Furthermore, since changes occurred, some languages at some stages must have been in transition. More interestingly (or worryingly), the "transition" may have consisted of periods of heterogeneity and internal competition within the language. Extending that principle: if Welsh, most Germanic, etc. are "aspirating", but Romance, most Slavic, etc. are "voicing", then some of all these languages must also have shifted type at some stage or other. So my general question is: how confident can the historical linguist be in (a) asserting and (b) demonstrating constraints on language change such as these? Granting that the constraints on changes are true per se at a micro-level, it would seem that the cumulative factors that may have influenced the final outcome of an observable change in a language at the macro-level are always potentially capable of scrambling things to the extent that prediction cannot be confidently maintained. There is the "Patagonian factor": articulation habits may be altered quite suddenly, in structurally unpredictable ways. There is the "see-sea-say" factor: owing to "shifting undercurrents", a language can perform apparent about-turns, even in violation of historical linguistic "laws". There is the heterogeneity factor which explains this: generalizations may not be as generalizable as they appear across dialectal or sociolinguistic varieties. There is also the notational factor observed by Patrick Honeybone: "we should be careful about comparing changes in different languages - the symbols we use to transcribe sounds could be pulling the wool over our eyes". And so on. It is because I perceive such a mass of potentially conflicting factors in real-life linguistic evolution that I tend to balk at assertions in historical linguistics like that of Benji's which sparked off this debate: >It just happened -- because both directions >are possible (under certain conditions -- certainly NOT ****z > s >/V_V). and, when confronted by such, am usually led to a position of devil's advocate: >That provoked me to try and think of counterexamples to >Benji's latter assertion here. Perhaps, strictly speaking, I failed to come up with an instance of intervocalic DEVOICING. Nonetheless, I feel that the ensuing debate has brought to the forefront to what extent a priori diachronic laws probably require a great deal of qualification and specification, so much so that, I suspect, their empirical verifiability could be doubtful. Alan From bwald at HUMnet.UCLA.EDU Thu Nov 19 17:16:11 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Thu, 19 Nov 1998 12:16:11 EST Subject: On Welsh voicing Message-ID: ----------------------------Original message---------------------------- Joe Eska writes: > In Welsh, broadly >speaking, phonemic /p b/, for example, are [p^h p] phonetically, i.e., >the contrast is one of phonetic aspiration, rather than voicing. >There is no process of intervocalic devoicing in Welsh, but there is >one of intervocalic de-aspiration for voiceless stops, which, in many >dialects, is accompanied by partial voicing. I feel I am not risking too much to read into this message that he is suggesting that Welsh fits into "natural" sound change with respect to voicing in intervocalic position, i.e., that the Welsh varieties at issue are variably indulging in an EXPECTED sound change: [-asp] -> <[+voi]> / V_V (if this is indeed restricted to intervocalic position). Perhaps there are Welsh varieties for which the sound change is completed as well, but Patrick Honeybone's message warns against taking transcriptional practices or changes in orthography at face value (with regard to culling the literature on Welsh without listening to the sounds themselves). A deeper issue, then, implicates the situation in Proto-Celtic (whether there were true intervocalic voiced stops), and perhaps even ultimately whether I-E "glottalic" theory is significantly reflected in Celtic (or Proto-Celtic). That is, either voiced stops existed at some point intervocalically BEFORE Welsh emerged, in which case there was devoicing under some conditions (not made clear), or, as far back as we can go into the ancestry of Welsh there were NEVER (consistent) intervocalic voiced stops. Alan King's message is more complex, and contains many interesting ideas, some of which intersect with Honeybone's, where Patrick emphasises the issue of distinctive or "primary" vs. secondary features. To take Patrick's point first, the issue of the nature of sound change remains. As far as I know, the idea that [+stop] > [-voiced] /V_V (and ONLY in V_V) would be an "unnatural" change, REGARDLESS of whether [voiced] was a distinctive feature of Welsh or not. That seems to be because the common change, [+stop] > [+voiced] / V_V has a PHONETIC explanation which leads to unexpectedness of the reverse change on PHONETIC grounds alone. It is not clear to me to what extent distinctive vs. "redundant" features has ever been admitted among practicing historical linguists as a mitigation of the phonetic explanation. It's clearly an issue of basic and general interest. Can phonology (patterns of phonetic organisation in language) mitigate phonetically regular (and Neogrammarianly "blind" phonetic) changes? Without pausing to think of examples, I suppose so, but I don't know about such a case as intervocalic DEvoicing, with respect to this issue. Now, for Alan's comments and suggestions. >Actually if we're looking at the northern Welsh model, the [+voiced] bit is strictly unnecessary. The change can be represented simply as: [+stop] > [+geminate] / V __ V, which I should think is intuitively more plausible. Of course, this occurred under the circumstance that voiceless stops in the specified position were already geminate (for diachronic reasons that we've >already been into). This feeds into the issue of phonological mitigation of "blind" or strictly phonetically motivated change. It did give me pause in my last message, but I accepted it as a fact (provisionally). Next, Alan writes: >Again, relevant to the Welsh phenomenon is the fact that the gemination probably only occurs when the preceding vowel is stressed, so another >modification is necessary: [+stop] > [+geminate] / [V, +stress] __ V. Off the top of my head, the +stress condition may be important, since the following vowel is then less prominent, and hence its features, such as voicing, may play less of a role in inhibiting a devoicing motivated by other factors (such as "pattern pressure"). That is, devoicing following a stressed vowel may be more "natural" than preceding a stressed vowel (due to anticipation of its maximally prominent voicing). Still, I don't know that the thesis of "naturalness" of intervocalic voicing of stops has ever considered such a distinction in offering its explanation in terms of strictly phonetic motivation. Alan again: >I would suggest there are two ways of "understanding" this change. One is just in the terms implied by the foregoing notation: northern Welsh speakers "decided" to geminated all stops between a preceding stressed and >a following (unstressed) vowel. "For fun", so to speak. I don't get the point of the personification. I'll ignore it. Alan continues: > The other view might be that since voiceless stops in that position were already always geminate, they "decided" to *simplify* their phonological system by extending the "gemination habit" to all stops, regardless of voicing. That seems more plausible to me. The change is [stop] > [+geminate]. NO CONDITIONING. For maintaining distinctions, however, that seems to necessitate accepting that there were previously voiced stops, which goes against what Esko seemed to be suggesting. Actually, it doesn't if the distinction affects only an inventory of [+/- aspirate] stops. The aspirate dimension remains after the change, e.g., -tth- vs. -tt-. The further change -tt- > -dd- (> -d-), or whatever, can be seen as maximal differentiation of +asp and -asp, but intervocalic position might be unnecessary and remains problematic (so far) as a CONDITIONING factor. Next Alan writes: >Actually they went further than that. Gemination in the said position was extended to *most* consonants; notably including /n/, /l/ and /r/. At this point we might say that they were really getting "carried away", since length in these particular consonants (only) had until that time been phonemic, as it still is in southern Welsh (although phonetically most of the work to distinguish short and long liquids is probably performed by the >preceding stressed vowel through a compensatory length contrast). Generalisation to resonants (or whatever you call them) detracts from the "maximal differentiation" suggestion, UNLESS there were previously resonant aspirates, which do exist in some languages I'm familiar with, and therefore is not a priori implausible. (I suppose Welsh /ll/ might even count as a liquid "aspirate' -- I'm unfamiliar with its historical sources. It only now occurs to me that "devoicing" (aspiration?) of Spanish /rr/ in some varieties as opposed to /r/ may be another case of maximal differentiation, but I'm accustomed to viewing this tendency, where it exists, as part of the general Spanish tendency to devoice "fricatives", cf. Buenos Aires variable /zh/ > /sh/, significantly from /y/ (?and /ly/?), which becomes /dzh/ in much of Spanish, where prior existence of /ch/ might play a role in keeping devoicing of the voiced affricate at bay -- there's also common Spanish /ch/ > /sh/, in many areas, but that's another matter) Alan again: >So far from doing such a strange thing as geminating intervocalic voiced stops for no apparent reason, we could say that they pretty much had a "gemination party", geminating nearly everything they found in the position >/ [V, +stress] __ V. I need to be assured that the conditioning factor is indeed necessary. I had previously suggested (on the basis of incomplete data) that the gemination was unconditioned but, perhaps, has only been recognised in intervocalic position, i.e., it reflects a more general paradigmatic shift among stop series. Alan's next comments are most interesting: >Why? I don't know if it's a motivation or merely an effect, but the outcome of this development is a very characteristic *staccato, almost syncopated* rhythm to northern Welsh speech, since most stressed syllables are pronounced with a short, rather energetic vowel followed by a long and also energetically articulated consonant which seems to go >implosion-interruption-explosion. Implicit in this suggestion is a METRIC explanation for the change in terms of syllabic rhythm in words. This is certainly worthy of further serious discussion. There's a chicken-and-egg problem involved. It also somewhat (but not completely) overlaps with my earlier suggestion that Welsh stressed vowels may be much more prominent and influential on (preceding) consonants than unstressed vowels, relevantly with respect to voicing. It would indeed be interesting (and relevant) if Welsh unstressed vowels had any tendency to devoice. I've never heard anything along those lines for Welsh. Variable devoicing is a well known feature of Swahili (and some adjacent languages) POST-stress vowels (which are necessarily word-FINAL, and voicing alone is not a distinctive feature of Swahili consonants -- the stops written as voiced in Latin or Arabic letters are imploded in first-language Swahili, truly voiced stops are prenasalised -- some other languages in the vicinity borrow words with voiced stops by prenasalising them, e.g., English "soda" > Gikuyu "sonda", Swahili just implodes the "d" in such words. Gikuyu doesn't have imploded or simple "voiced" stops. Meanwhile Gikuyu has variable denasalisation of the prenasals, so the actual pronunciation can be "so(n)da" without reference to the English original, or the faintly possible Swahili intermediary.) Alan continues: >Interestingly, in some speakers at least, the intensity of the stressed and the posttonic unstressed vowels doesn't seem to contrast as much in this "articulatory style" as in, say, English or even Spanish (so that foreign ears may have difficulty interpreting which is the stressed syllable), and it now occurs to me that this may be explained if we assume that the "staccato-ey" features I've described have taken over the function of identifying the position of the stress. In particular it is common for the pitch of the posttonic syllable >to be higher than that of stressed syllable. That does not support what I just had in mind concerning the difference between stressed and post-tonic vowels, but continues the theme of metricity as a crucial feature of the relevant Welsh sound changes. Alan again: >As I already said previously, stops in general tend to be strictly voiceless in pronunciation, with other features such as aspiration doing most of the work of differentiating the voiced and voiceless series. It seems to me that in this form of speech intervocalic geminate stops are *particularly* voiceless: i.e. voicing is interrupted along with everything else in their intervocalic articulation. I believe that the voiced and voiceless series nonetheless remain phonologically fully differentiated in this pronunciation (although some foreign speakers might be excused, again, for getting the signals wrong). But *phonetically* one might argue that you already have intervocalic devoicing here, at least as a subsidiary effect - assuming of course that some voicing is there in the voiced stop >series to begin with. Again, Alan is presupposing an earlier change FROM voiced stops. For the moment this has to weighed against the contrary implication of Esko's remarks, which seem to leave open the possibility that where there is no voicing in Welsh there historically never was. Alan again: And if that is the case, it appears to me that we might describe the phenomenon as one of dissimilation. Exactly. But in the way I described it above as "maximal differentiation" [+asp] = [-voice] vs. [-asp] > [+voice]. I continued to question the conditioning factor, but Alan's comments on metricity leave me hesitant as to what is a more likely (and even more "natural") account. Alana again: >There are surely enough precedents for that is phonology? When intervocalic stops get voiced, that is assimilation to the voicing of the surrounding vowels, I take it; why then, when the opposite happens it is surely dissimilation. The motivation might be to heighten the contrast between adjacent segments: vowels voiced, >consonants unvoiced. The whole issue of dissimilation is complex and requires a lot of thought (I vaguely remember that a paper or a few have addressed this subject at length in general). To be sure, if the PHONETIC explanation of intervocalic voicing is (voicing) ASSIMILATION, then it is most expected that any attempt at explanation of a reverse phenomenon would begin with the notion of DISSIMILATION. I have my reservations, but the "maximal differentiation" notion fits into a general notion of dissimilation on the paradigmatic level. "Maximal differentiation" is a most well-known concept in accounting for the tendencies of VOWEL systems, but I know at least a few examples where this can be posited as an underlying motivation for changes in certain CONSONANTAL systems (even beyond dissimilation of coarticulation effects for consonants). EG, among "Swahili" varieties Bajuni exhibits the change t > ch and nd > ndr, where "ndr" is what English ears hear (and have commented on), but is simply an effect of the release of the retracted consonant and is not heard as an "r" release by Bajunis. The t > ch is a bona fide palatal affricate, but arose from the same retraction . Dissimilation is involved to the extent that these two sounds have become more distinct from the opposing dental stops t. and nd. (which, ironically, are reflexes of earlier *ch and *nj, still maintained in Southern and standard Swahili). Finally, "dissimilation" alone is not sufficient explanation for (problematic) intervocalic devoicing, but I hasten to add that Alan went a lot further than to suggest that it could be. From mcv at wxs.nl Fri Nov 20 12:08:23 1998 From: mcv at wxs.nl (Miguel Carrasquer Vidal) Date: Fri, 20 Nov 1998 07:08:23 EST Subject: Arm. targal 'spoon' (fwd) In-Reply-To: Message-ID: ----------------------------Original message---------------------------- manaster at umich.edu wrote: >I have as some of you may know proposed an Armenian >sound law whereby inter alia PIE *w ends up ultimately as >j^ after *r(H) or *l(H) in Armenian, e.g., olj^ is then simply >*solwo-. The only counterexample that I am aware of is targal >if this is really, as many think, from *drwaHlaH or the like. >So the obvious question is can anyone think of an alternative >etymology. One (maybe even the) possibility would be some >proposal whereby targal is from *tarigal or *tarugal with the >regular Armenian syncope of the relevant vowels. Any takers? We could postulate a form like *deru'wa(:)l- or doru'wa(:)l-, with *e or *o > a and regular development of w- before the stressed syllable, but in such a derivation (or compound) I would sooner expect *druwa(:)l-, which would have given **artugal. If the the etymology is, as I suspect, *derwa(:)l-, have you considered the possibility that it's not a counterexample at all? Armenian *w behaves very differently depending on whether it's syllable initial or syllable final and on where the stress used to be. I gather it was lost after the stress in intervocalic position (erkan "millstone" < 'gwra:wen-), it remains as -w in absolute final position after a vowel (naw "ship"), and becomes g- syllable initially, at least if the (PIE) syllable was stressed. The case of , at least, fits none of these patterns, as it occurs after the stress, not intervocalically, and, if the sound law operated late enough, maybe even in syllable-final position (*[h]olw). And on seeing *[h]olw, I wonder if there might be a connection between your proposed *(r/l)w > j^ (by way of -y-?) and what we see in the o- and a:-stems in -i, where the oblique cases have -w- (hogi, hogw-oy "spirit", maybe from *hogw, *hogwosio; and aygi, aygwoy "vineyard" from *aygw, *aygwosio, and thus from plain PIE *oiwa: instead of Pokorny's dreadful *oiwiia:)? That would of course require explaining why *gw < *w was not delabialized here, but seems otherwise a neat solution: after a consonant, syllable final (?) -w > -y, and further develops into -j^ after liquids, and -i after stops. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From manaster at umich.edu Fri Nov 20 12:08:50 1998 From: manaster at umich.edu (manaster at umich.edu) Date: Fri, 20 Nov 1998 07:08:50 EST Subject: Arm. targal 'spoon' (fwd) In-Reply-To: <365dbbf8.116934866@mail.wxs.nl> Message-ID: ----------------------------Original message---------------------------- On Fri, 20 Nov 1998, Miguel Carrasquer Vidal wrote: > We could postulate a form like *deru'wa(:)l- or doru'wa(:)l-, with *e > or *o > a and regular development of w- before the stressed syllable, > but in such a derivation (or compound) I would sooner expect > *druwa(:)l-, which would have given **artugal. > This does seem a dead end. > > If the the etymology is, as I suspect, *derwa(:)l-, have you > considered the possibility that it's not a counterexample at all? > Armenian *w behaves very differently depending on whether it's > syllable initial or syllable final and on where the stress used to > be. I have, but there are too few examples to hang anything like that on. > And on seeing *[h]olw, I wonder if there might be a connection > between your proposed *(r/l)w > j^ (by way of -y-?) and what we see > in the o- and a:-stems in -i, where the oblique cases have -w- (hogi, > hogw-oy "spirit", maybe from *hogw, *hogwosio; and aygi, aygwoy > "vineyard" from *aygw, *aygwosio, and thus from plain PIE *oiwa: > instead of Pokorny's dreadful *oiwiia:)? That would of course > require explaining why *gw < *w was not delabialized here, but seems > otherwise a neat solution: after a consonant, syllable final (?) -w > > -y, and further develops into -j^ after liquids, and -i after stops. > I was just thinking about this. Wow! Alexis From Roger.Wright at liverpool.ac.uk Fri Nov 20 12:13:42 1998 From: Roger.Wright at liverpool.ac.uk (Roger Wright) Date: Fri, 20 Nov 1998 07:13:42 EST Subject: intervocalic devoicing In-Reply-To: <3.0.3.32.19981119115151.006f2a38@pop3.redestb.es> Message-ID: ----------------------------Original message---------------------------- Alan King says: >The proof of the pudding is that there is, or was, an important colony of >(originally, anyway) Welsh speakers, dating from a hundred years ago, in >the Patagonia region of Argentina. ... Life being what it is, nowadays >Spanish seems to have taken over most functions in Welsh Patagonia, >which has been largely Argentineanized by now, but there is still a >significant residue of maintenance of group identity and even some >linguistic maintenance. Predictably, the Welsh back in Wales find all >this very interesting, and some of the Patagonians are drawn to their >ancestral homeland, so some ties exist. There is, or was, a scheme whereby two such Patagonians a year had a scholarship to attend Bangor University (North Wales). Part of the point of this, for the linguists, is that Welsh is now spoken almost only by bilinguals, and perhaps the "pure" Welsh is that which can be deduced to be the highest common factor of the Welsh-Spanish and Welsh-English bilinguals. > .... is their pronunciation of the stops, since the more >recent generations, who are Spanish speaking, apparently assimilated the >Welsh system to the Castilian one. Argentinian Spanish isn't very Castilian, but, yes, it still has the voicing opposition, normally without aspiration. (Although some Castilian specialists prefer to think the contrast usually called one of voicing is essentially a tense-lax opposition). RW From tbd at mailserv.waikato.ac.nz Mon Nov 23 12:08:52 1998 From: tbd at mailserv.waikato.ac.nz (T. F. Baer-Doyle) Date: Mon, 23 Nov 1998 07:08:52 EST Subject: help dorothy Message-ID: Dear Collegues, Here is some sad news about one of our colleagues which I'm passing on to you. -Dorothy Disterheft ----------------------------Original message---------------------------- Dorothy I'm a complete stranger to linguistics but I have sad news of one of your members. Michael John Edmonds from University of Auckland, New Zealand, died 10 days ago. His death had been expected for some time, sadly, but he went out with all flags flying in Chomsky'sk fashion in a protest against the colonisation of disabled students by able bodied bureaucrats. Should any members of the list wish to send condolences to his Mum, Karen Shaw, they may send emails to my address and I shall pass them on. She would be glad to hear from any friends especially, as most of Mike's communications were via this medium. Many thenks - and sorry I can't negotiate the listserver instructions that were provided. Dr T. F. Baer-Doyle University of Waikato. Dr T F Baer-Doyle University of Waikato Private Bag 3105 Hamilton, New Zealand Ph (direct line) 64 7 838 4907 (w) Ph 64 7 824 1991 (h) Fax 64 7 838 4434 email. tbd at waikato.ac.nz From mccay at redestb.es Mon Nov 23 12:10:47 1998 From: mccay at redestb.es (Alan R. King) Date: Mon, 23 Nov 1998 07:10:47 EST Subject: intervocalic devoicing: language nomenclature In-Reply-To: Message-ID: ----------------------------Original message---------------------------- I wasn't going to bother, but as there isn't much excitement here at the moment I'll go for it, and defend myself on a "correction" of one of my last posts. Roger Wright "pulled me up" on my referring to "Spanish" as "Castilian". I said: >> .... is their pronunciation of the stops, since the more >>recent generations, who are Spanish speaking, apparently assimilated the >>Welsh system to the Castilian one. to which Roger remarks: >Argentinian Spanish isn't very Castilian, but, yes, it still has the >voicing opposition, normally without aspiration. For the record, it is perfectly commonplace, particularly in Latin America, to refer to "Spanish" as "castellano". Agreed, in English the most widespread term is "Spanish", but the two can be considered synonyms (although some other usages have sometimes been suggested). The point is that "Castilian" does *not* normally refer to a certain dialect of the Spanish language, either traditionally or in present usage (although it perfectly well could do, of course). Rather, it refers to the Romance *language* which first sprang up in Cantabria and Castile, became the tongue of the medieval kingdom of Castile, was spread from there by conquest to other parts of the peninsula and thence elsewhere around the world (Argentina included), has also been progressively but still incompletely imposed on other linguistic communities within the state of Spain itself (namely Galicia, Asturias, the Basque Country and the "Paisos Catalans"), and is now known as either the Castilian or the Spanish language. As for the preferred Politically Correct designation, there are different viewpoints (on what isn't there?). Among those with strongly aware of the existence, and indeed official recognition by the present constitutional regime, of languages other than "Spanish" in the "Spanish State" (another PC expression in such circles), there are (at least) two conflicting usages. Many prefer the designation "Castilian" to emphasise that it is not, after all, the only "Spanish" language. A politically more radical position (at least in the Basque Country) says that calling it "Spanish" is perfectly PC. Why? Well, because Galicia, the Basque Country and the "Catalan Countries" aren't really *Spain* anyway.... I offer these remarks, naturally, not to start an argument but as documentation of current alternative usages on a subject perfectly tangential and immaterial to the ongoing discussion, but pertinent for linguists and perhaps of passing interest to the present readership, if only out of curiosity. Alan Alan R. King, Ph.D. alanking at bigfoot.com - - - - - - - - - - - - - - - - - - - - - - - - SNAIL: Orkolaga plaza 3 1A, 20800 Zarautz, Basque Country, Spain. PHONE: +34-943-134125 / FAX: +34-943-130396 Alternative email addresses: mccay at redestb.es, a at eirelink.com, 70244.1674 at compuserve.com Internet: From bwald at HUMnet.UCLA.EDU Mon Nov 23 12:15:05 1998 From: bwald at HUMnet.UCLA.EDU (bwald) Date: Mon, 23 Nov 1998 07:15:05 EST Subject: intervocalic devoicing Message-ID: ----------------------------Original message---------------------------- Alan King's most recent message includes the following statement: >There is the "see-sea-say" factor: owing to "shifting undercurrents", a language can perform apparent about-turns, even in violation of historical linguistic "laws". I think the reference is to the same thing usually exemplified with "meet - meat - mate". It is indeed an interesting historical problem, but not one to be treated as casually as the above statement does. The "mystery", well-known in the history of English, is that in Early Modern English, during the Great Vowel Shift, there was a point at which orthoepists and the like reported a merger of "meat" and "mate", more generally, for earlier /E:/ and /ey/, as "meet", and more generally /e:/ rose to /i:/ position. Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A small residue of the apparent more general merger remains in the words, "great", "steak", "break" and a few others (like "drain" < "drean") and many words before "r" in closed syllables, e.g., "tear" (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, etc.) There have been many attempts to solve the problem. Halle is (in)famous for suggesting that the merger took place on the phonetic level but that the earlier *phonemes* remained distinct. He did not produce morphophonemic alternations for the "ea" words that presumably detached themselves from /e:/ (or /ey/) and rose to /i:/, leaving mysterious and mystical the support for the maintenance of an underlying (morpho)phonemic distinction in face of the phonetic merger. A more traditional attempt at solution is to hem and haw about "dialect mixture" (esp in London), such that there were "meet" vs. "meat/mate" merger dialects (earlier prestigious) in contact with "meet/meat" merger vs. "mate" dialects. Somehow the latter dialects replaced the former (in the relevant area) but the "exceptional" words I mentioned above from the former dialects survived the replacement process (as if constituting a substratum). Though complicated, this solution is not totally unreasonable, and was obviously aimed at defending the notion of the regularity of sound change at the expense of seeking any control over lexical borrowing across dialects. More recent theories, aimed at greater reconciliation of the two traditional approaches, involve lexical diffusion of sound change. This does not really require two dialects in contact but a gradual diffusion of a sound change through the lexicon. A variant could propose (I don't know if anyone has made this proposal in print) something like: a lot of "ear" words and the other exceptions were the first to rise from /E:/ to /e:/ and somehow they merged with /ey/. However, as the process of /E:/ > /e:/ continued they did NOT merge with /ey/ but /e:/ caught up with the "ee" words either at or on their way to /i:/. More radical, and, unsurprisingly, my preferred solution, is the one of "near-merger" (Labov). According to this theory, /e:/ and /ey/ did not merge as reported, but remained distinct along a dimension of "tenseness" ("frontedness") that the orthoepists did not recognise (even though they themselves may have unconsciously made the distinction). So, the fronter (tenser) /e:/ remained distinct from the front (but less front, laxer) /ey/ and continued to rise to /i:/, to merge with the "ee" words. There was phonetic conditioning holding back many of the "ea" words before same-syllable /r/, and they indeed did merge with "-are" and "-air". A similar laxing effect is still operant for most vowel before (consonantal) -r in current dialects. Similarly, pre-vocalic r (as in great, break, drean = drain) and post-vocalic k (as in steak, break) retarded further raising (although "creak", "streak" etc), a statement about "mini" phonetic conditioning effects during the variable stage of a sound change. None of the above solutions are totally satisfying or above criticism (and further research for testing), but I don't think one should throw up one's hands and cavalierly talk about "apparent about-turns, even in violation of historical linguistic "laws". Instead, I think one should recognise the challenge to investigate the matter further. It is the next frontier to be crossed in understanding the nature of sound change, and indeed the nature of linguistic "laws". (I guess "laws" has a double-meaning, first, man-made/artificial and ultimately fallable attempts to explain/account for a phenomenon, which I take to be the point of Alan's scare quotes, and, second, the actual principles governing linguistic behavior and change, which I take to be the purpose of linguistic research and the way linguists channel and DISCIPLINE their curiosity about the facts of language.) From mccay at redestb.es Tue Nov 24 13:09:59 1998 From: mccay at redestb.es (Alan R. King) Date: Tue, 24 Nov 1998 08:09:59 EST Subject: Apparent violations of historical linguistic laws In-Reply-To: Message-ID: ----------------------------Original message---------------------------- Benji Wald says: >None of the above solutions are totally satisfying or above criticism (and >further research for testing), but I don't think one should throw up one's >hands and cavalierly talk about >"apparent about-turns, even in violation of historical linguistic "laws". >Instead, I think one should >recognise the challenge to investigate the matter further. It is the next >frontier to be crossed in understanding the nature of sound change, and >indeed the nature of linguistic "laws". (I guess "laws" has a >double-meaning, first, man-made/artificial and ultimately fallable attempts >to explain/account for a phenomenon, which I take to be the point of Alan's >scare quotes, and, second, the actual principles governing linguistic >behavior and change, which I take to be the purpose of linguistic research >and the way linguists channel and DISCIPLINE their curiosity about the >facts of language.) On a first reading, I agree with much of what this paragraph seems to be saying, which could even be taken as a summary of what I was trying to say. Being a complex statement, however, there is room for many subtle changes of emphasis among the various points, and some of Benji's wording is perhaps a little unfair. I don't actually remember "throwing up my hands". And I would certainly be in favour of "recognising the challenge to investigate the matter further". But the real crux of my message was in these words, reproduced by Benji but perhaps not read quite in the way I had intended: "apparent about-turns, even in violation of historical linguistic 'laws'". The above "summary" focuses on the word *laws*, and indeed defends the concept against what is seen as my attack. But in fact, I was more interested in the word *apparent*. The *interesting* thing about problems such as the one I was recalling (which I didn't claim to have discovered, even though I seem to have made the mistake of calling it the "see-sea-say" problem rather than the "meet - meat - mate" problem - silly me!) is that they serve to exemplify the possibility that what *apparently* happens may be different from what really does. We notice that the *apparent* reversal of the merger of "meet" and "mate" CAN'T be exactly what happened BECAUSE it would violate the laws which we believe to operate in such cases. IF on the other hand the apparent event had, by chance, NOT violated our law, we would not be aware of a "problem" and in consequence would probably not question the reality or exact nature of the said event. All the same, I don't think the "laws" (hence, in the last resort, my scare quotes) ought to be seen as infallible and unassailable, since they do rest for their validation on observed events (or *apparently* observed ones, anyway!). And at the risk of repeating a clich?: we are in a privileged position when discussing the history of a language as well documented as English, which makes it easier to "catch ourselves out". Where most languages and language families are concerned, it is presumably easier to get away with our mistakes. As for the emphasised word DISCIPLINE, I am all for it, but there is more than one kind. I think that true intellectual discipline should also encompass a constant preparedness to criticise one's own theories, even looking for *possible* cracks in the foundations, rather than blindly building higher and higher. But then, I happen to think that the ultimate goal is not to finally discover "all" the laws, but in the process to arrive at a better acquaintance with the phenomena they attempt to synthesize. Fortunately, sometimes this goal is achieved. From DISTERH at UNIVSCVM.SC.EDU Tue Nov 24 14:16:27 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Tue, 24 Nov 1998 09:16:27 EST Subject: reversal of merger In-Reply-To: Message-ID: ----------------------------Original message---------------------------- This message was originally submitted by larryt at COGS.SUSX.AC.UK to the HISTLING list at VM.SC.EDU. If you simply forward it back to the list, using a mail command that generates "Resent-" fields (ask your local user support or consult the documentation of your mail program if in doubt), it will be distributed and the explanations you are now reading will be removed automatically. If on the other hand you edit the contributions you receive into a digest, you will have to remove this paragraph manually. Finally, you should be able to contact the author of this message by using the normal "reply" function of your mail program. ----------------- Message requiring your approval (74 lines) ------------------ On Mon, 23 Nov 1998, Benji Wald wrote: > I think the reference is to the same thing usually exemplified with > "meet - meat - mate". It is indeed an interesting historical > problem, but not one to be treated as casually as the above > statement does. The "mystery", well-known in the history of > English, is that in Early Modern English, during the Great Vowel > Shift, there was a point at which orthoepists and the like reported > a merger of "meat" and "mate", more generally, for earlier /E:/ and > /ey/, as "meet", and more generally /e:/ rose to /i:/ position. > Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A > small residue of the apparent more general merger remains in the > words, "great", "steak", "break" and a few others (like "drain" < > "drean") and many words before "r" in closed syllables, e.g., "tear" > (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, > etc.) > There have been many attempts to solve the problem. For the record, I've recently been collecting published proposals for accounting for apparent cases of reversal of merger. So far I have six, as follows: (1) The merger occurred at the phonetic level, but speakers retained different phonological representations in their heads, allowing later reversal (Halle). (2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had this role from those which did not (Michelena). (3) The merger occurred in the prestige variety but not in less prestigious varieties; since only prestige varieties tend to be well recorded, a shift in prestige shows up in the record as an apparent reversal of merger (Weinreich, Labov and Herzog). (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). (5) The merger never really occurred; instead there was only a near-merger, resulting in the usual inability of speakers to recognize the difference (Labov). (6) The merger occurred for most speakers, but a handful of influential conservative speakers succeeded in reversing the merger by semi-official action (Jahr). Of these, (1) is now dismissed, I think; (2) is unchallenged for the case for which it was proposed; (3) is still widely accepted; (4) is taken seriously but is sparsely documented; (5) is widely defended; (6) is documented for a particular case. I do not claim that these six exhaust the proposals, especially since I still have a small folder of relevant articles to read. My final list may be somewhat longer. Of course, only some of these proposals can account for the cases of "residue" noted by Benji. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From DISTERH at UNIVSCVM.SC.EDU Wed Nov 25 13:43:27 1998 From: DISTERH at UNIVSCVM.SC.EDU (Dorothy Disterheft) Date: Wed, 25 Nov 1998 08:43:27 EST Subject: reversal of merger In-Reply-To: Message-ID: ----------------------------Original message---------------------------- ----------------------------Original message---------------------------- This message was originally submitted by larryt at COGS.SUSX.AC.UK to the HISTLING list at VM.SC.EDU. If you simply forward it back to the list, using a mail command that generates "Resent-" fields (ask your local user support or consult the documentation of your mail program if in doubt), it will be distributed and the explanations you are now reading will be removed automatically. If on the other hand you edit the contributions you receive into a digest, you will have to remove this paragraph manually. Finally, you should be able to contact the author of this message by using the normal "reply" function of your mail program. ----------------- Message requiring your approval (74 lines) ------------------ On Mon, 23 Nov 1998, Benji Wald wrote: > I think the reference is to the same thing usually exemplified with > "meet - meat - mate". It is indeed an interesting historical > problem, but not one to be treated as casually as the above > statement does. The "mystery", well-known in the history of > English, is that in Early Modern English, during the Great Vowel > Shift, there was a point at which orthoepists and the like reported > a merger of "meat" and "mate", more generally, for earlier /E:/ and > /ey/, as "meet", and more generally /e:/ rose to /i:/ position. > Later, however, /E:/ rose to /i:/ and merged with previous /e:/. A > small residue of the apparent more general merger remains in the > words, "great", "steak", "break" and a few others (like "drain" < > "drean") and many words before "r" in closed syllables, e.g., "tear" > (the verb), "wear", "bear", etc. (contr. "spear", "tear" the noun, > etc.) > There have been many attempts to solve the problem. For the record, I've recently been collecting published proposals for accounting for apparent cases of reversal of merger. So far I have six, as follows: (1) The merger occurred at the phonetic level, but speakers retained different phonological representations in their heads, allowing later reversal (Halle). (2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had this role from those which did not (Michelena). (3) The merger occurred in the prestige variety but not in less prestigious varieties; since only prestige varieties tend to be well recorded, a shift in prestige shows up in the record as an apparent reversal of merger (Weinreich, Labov and Herzog). (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). (5) The merger never really occurred; instead there was only a near-merger, resulting in the usual inability of speakers to recognize the difference (Labov). (6) The merger occurred for most speakers, but a handful of influential conservative speakers succeeded in reversing the merger by semi-official action (Jahr). Of these, (1) is now dismissed, I think; (2) is unchallenged for the case for which it was proposed; (3) is still widely accepted; (4) is taken seriously but is sparsely documented; (5) is widely defended; (6) is documented for a particular case. I do not claim that these six exhaust the proposals, especially since I still have a small folder of relevant articles to read. My final list may be somewhat longer. Of course, only some of these proposals can account for the cases of "residue" noted by Benji. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From mcv at WXS.NL Wed Nov 25 16:20:27 1998 From: mcv at WXS.NL (Miguel Carrasquer Vidal) Date: Wed, 25 Nov 1998 16:20:27 GMT Subject: reversal of merger In-Reply-To: Message-ID: Larry Trask wrote: [on "reversal of merger":] >I do not claim that these six exhaust the proposals, especially since I >still have a small folder of relevant articles to read. My final list >may be somewhat longer. The one that's certainly missing is "the distinction was reintroduced from outside (a different dialect)". This may be seen as a "spatial" supplement to the socially motivated (3) [change in prestige dialect] and (6) [reintroductioon by conservative minority]. Sorry, I don't have a specific source. I'm fairly sure this is the explanation that Leskien or Brugmann would have given. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From bwald at HUMNET.UCLA.EDU Thu Nov 26 04:45:53 1998 From: bwald at HUMNET.UCLA.EDU (bwald) Date: Wed, 25 Nov 1998 20:45:53 -0800 Subject: reversal of merger Message-ID: I appreciated Larry Trask's enumeration of distinct cases which resemble reversal of merger. Among the cases he mentioned, the following is particularly interesting for the meet:meat:mate case: (4) The merger occurred, but only variably, and speakers retained both merged and unmerged pronunciations, though they may have reported only the merged one (Milroy). More examples come to light almost daily, and they tend to be just on the other side of an isogloss of an innovative merger. Speakers report the reference dialect, not their own. According to the Milroys (if I remember correctly -- and I think it is mentioned in Labov 1994 as well), Belfast maintains a separate "meat" class, distinct from both "meet" and "mate" in that the "meat" class varies between both the "meet" and "mate" pronunciations (stressed vowel, that is) while the "meet" and "mate" classes are invariably distinct. If I got this straight -- and I think I have the gist right if not all the details, and it exemplifies Larry's (4) above (though it shows a class which merges with two distinct other classes (or vowels) and has no unique vowel of its own except that it varies between two distinct vowels), then the "residue" in most current varieties of English may have arisen through collapse of the variable class into the two invariant classes. That is, FIRST, Class/Vowel 1 = i: as in "meet", Class/Vowel 2 = e: as in "mate" and the problematic Class/"Vowel" 3 = i: OR e: as in "meat", THEN, Class/"Vowel" 3 breaks up into Classes 1 and 2 but eventually no words remain in both. This is not neogrammarian sound change, but reassignment of the words of a distinct (variable) word class to two distinct word classes. The problem remains why most of the "ea" words were reassigned to /i:/, but much fewer of "ear$" words were. Again, there seems to be phonological conditioning of a familiar type in English here, such that the centralising tendency of a following tautosyllabic r inhibited reassignment of "ear$" words to /i:/ (rather a more mid vowel). How often in practice can we distinguish phonologically conditioned lexical partition of a variable phoneme from sound change? Of course, the Belfast situation may not reflect the situation of the earlier reported London merger of "meat" and "mate" and then unmerger and merger with "meet", but, in the absence of compelling evidence to the contrary, it remains a possibility to the extent that such a linguistic possibility as the one described for Belfast is attested. It also goes without saying that it is quite interesting to contemplate how such a variable phoneme could arise in the first place X -- collapse of two "dialects" into one? (more on this later.) I also remember that when I read Labov's (1994) account of the Belfast situation I was struck that he did not report phonetic measurements (of vowel formant structure) for the three word classes at issue, so even if my characterisation of the Milroy's remarks on the Belfast variable class is accurate, it is still not clear that there is indeed absolute merger with the two other distinct word classes, rather than a near-merger (for the mid-level "meat" and "mate"). This problem is aggravated for extrapolating to historical London by the problem of whether or not /ey/ as in "hair", "pain", etc. had merged with /e:/ as in "hare", "pane" etc., in the relevant time period -- it did later, by the end of the 18th c., and they are still distinct in many British areas. That's why I preferred the "meat"-"mate" example to the "sea" - "say", since "say" descends from a diphthong and may have never become a monophthong -- again the "reversal" problem of whether [ay] > [ey] OR [ay] > [a:] > [e:] > [ey] (the "reversal" if the diphthong ever became a monophthong and then diphthongised again). With regard to Belfast, I do not know if the Milroys ever performed acoustic measurements to determine whether Class 3 shows mergers with Classes 1 and 2, or whether they rely simply on ear and self-report (neither of which is sufficient for near-merger). I wonder if I should assume that they did do measurements, given that Labov (1994) does not discuss Class 3 as a possible case of near-merger. With regard to the whole concept of a "variable" phoneme (one that is phonemically but not phonetically distinct from one (or more?) other phonemes), I was struck in the mid 1970s by "unmerger" in second language phonology. The particular case I examined was the distinction between /ch/ and /sh/ in the fluent English of L1 Spanish (Mexican) speakers. Although /ch/ and /sh/ had the same range of allophones (namely, [ch] and [sh] and something in-between), the patterns of realisation of English /ch/ and /sh/ words was different, such that /ch/ words were more often realised as [ch] for the same speaker and vice-versa for /sh/ -- with the same phonetic conditioning for both (the sound [ch] was more favored word initially for both classes and [sh] more favored elsewhere -- but still affected in frequency by the English word class). Further work has been done on such L2 processes since then. Certainly, investigators generally speak of "developmental" processes in such cases, and L1 acquisition is somewhat similar, .but the speakers I looked at were adults, some quite old, and had spoken English for decades (since their early 20s), and their systems seemed to be stable -- as if they had distinct phonemes /ch/ and /sh/, and a single pronunciation rule which variably merged them, with different frequencies in different phonetic positions. Trask's case (2) piqued my curiosity: >(2) The merger genuinely occurred, but just one of the merged segments possessed a distinctive phonological role in the language, allowing speakers later to separate out instances of the merged phoneme which had >this role from those which did not (Michelena). I am not familiar with the Michelena example. Is the general case the one that, say, Kiparsky used to explain the general case of Yiddish final voiced stops. That is, for verbs and nouns the lexical morpheme may appear finally or intervocalically due to inflection, e.g., hob-en 'have' but 'hop' '(I) have. The phonological rule is: devoice all final stops. Intervocalic position preserves the word class distinct from the word class of final voiceless stops, which remain voiceless in intervocalic position. Generally, Yiddish has revoiced the final stops when they are voiced in intervocalic position, i.e., Yiddish has withdrawn the rule: devoice final stops. Kiparsky proposed that the former rule can be detected where there was no alternation, e.g., with the adverb 'avek' ("away", historically from 'a-veg'). This kind of "unmerger" is apparently what Halle was hoping he could establish for the shift in "ea" words, but it turned out to be an inappropriate example. But, now, what is the Michelena example? Last but, most emphatically, not least, I want to apologise to Alan King if I misrepresented the thrust of his remarks on the complexity of linguistic, even phonological, even phonetic, change. As I've said before, I am grateful to Alan for his frequent insightful examples and comments. The Welsh problem deserves to become a classic, and I do not doubt that lasting fame will come to the linguist (or group of linguists) who manage to make the linguistic community at large aware of the general theoretical significance and challenge posed by the problem (and to who/mever "solves" the Welsh problem). From larryt at COGS.SUSX.AC.UK Thu Nov 26 15:51:35 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Thu, 26 Nov 1998 15:51:35 +0000 Subject: reversal of merger, proposal (2) Message-ID: People are asking me about this, so I'll spell it out a bit. The case is made for a development in the Gipuzkoan variety of Basque, by Luis Michelena, in his magnum opus Fonetica Historica Vasca and elsewhere. Note the following orthographic conventions in Basque: = laminal [s] = English = the affricate [ts] = English represents the glide [j] (= US [y]) historically, but today represents a variety of consonants, according to dialect, derived from this glide by strengthening. Basque possesses a set of palatal and palato-alveolar consonants which, historically, never occur in ordinary lexical items, but only in expressive variants of these items and in expressive formations generally. For example, `bull' has an expressive variant `little bull', and `corner' has the expressive variant or `nook, cosy little place'. Such expressive palatalization was formerly pervasive in Basque, and it survives today in some varieties. Originally, therefore, the consonants and (among others) occurred *only* in such expressive forms, and nowhere else at all. But then the language acquired new instances of in ordinary lexical items, by borrowing, as in `soap' and `care, attention, concern', both borrowed from Romance. These were not expressive variants or forms, since they had no other form. Meanwhile, the historical glide /j/ was undergoing fortition in most dialects to some kind of consonant. The result differed according to region, but, in much of the country, it became a voiced palato-alveolar fricative [ezh], similar to French . This is still the state of affairs in the eastern extremity of the country. In the center of the country, however, and particularly in the Gipuzkoan dialect, this fricative underwent devoicing to [esh} (English ), thus merging with the historical . This is still the state of affairs today in some southern parts of Gipuzkoa: complete merger. In most of Gipuzkoa, however, there was a later change: (that is, [esh]) was backed to velar [x] or to uvular [X], a change which was apparently borrowed from Castilian Spanish, which underwent the same development. Now, we might have expected *all* instances of to undergo this backing, but that's not what happened. Instances of derived from earlier /j/ underwent backing. Hence native `lord', `eat' and `owner', which had historically contained the glide, underwent backing, and their modern Gipuzkoan forms are <[x]aun>, <[x]an> and <[x]abe>, with velars or uvulars. Instances of borrowed also underwent backing, and so borrowed and became <[x]aboi> and , also with velars or uvulars. But instances of *original* (that is, expressive) did *not* undergo backing, and so expressive , for example, has remained , with the sound of English and no backing. Hence Gipuzkoan has apparently first merged historical /j/ with and then reversed the merger, so that only non-historical instances of underwent backing. Now, one might suppose that the merger never really took place, that the two sounds somehow remained distinct until after the backing of one of them had occurred. But Michelena argues against this as inadequate. For one thing, the backing observed in loan words shows that these words must have contained the same sound that developed from original /j/. For another, those varieties that failed to undergo the backing invariably exhibit a complete merger. But there's more. In one or two cases, the expressive value of had apparently been lost -- and *these* instances of *did* undergo backing. Consider the word `man'. This has the regular combining form , and this in turn has the regular expressive variant . >From this , Basque formed a derivative `poor fellow', with the rare diminutive suffix <-xo> (diminutive suffixes, being intrinsically expressive, always contain the special consonants). And this has developed in Gipuzkoan to the unexpected , in which the first instance of has failed to back while the second instance has undergone backing to the velar/uvular [x]/[X], notated . Michelena's explanation is therefore the following. Instances of which were expressive in nature clearly retained their expressive function, and hence they remained palato-alveolar and did not undergo backing -- because they were part of a wider system of palatal and palato-alveolar consonants confined to expressive forms. All other instances of -- that is, those derived from /j/ and those in loan words -- lacked this expressive value and hence could and did undergo backing. In the case of , the expressive value of the first remained obvious, and hence this did not back. But, perhaps because of the rarity of diminutive <-xo>, the expressive value of the second was lost, and so this one underwent backing. This account seems to explain the facts admirably, and, to my knowledge, it has never been challenged. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Thu Nov 26 14:31:23 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Thu, 26 Nov 1998 14:31:23 +0000 Subject: reversal of merger In-Reply-To: <365f2e22.604986038@mail.wxs.nl> Message-ID: On Wed, 25 Nov 1998, Miguel Carrasquer Vidal wrote: > Larry Trask wrote: > > [on "reversal of merger":] > >I do not claim that these six exhaust the proposals, especially since I > >still have a small folder of relevant articles to read. My final list > >may be somewhat longer. > > The one that's certainly missing is "the distinction was reintroduced > from outside (a different dialect)". This may be seen as a "spatial" > supplement to the socially motivated (3) [change in prestige dialect] > and (6) [reintroductioon by conservative minority]. Sorry, I don't > have a specific source. I'm fairly sure this is the explanation that > Leskien or Brugmann would have given. Yes. I've been looking for a good explicit source for a proposal along these lines, but so far the only mentions I've dug up have been too inexplicit to constitute a formal proposal. Indeed, the ones I've seen have amounted to little more than mouthing mantras. I would be grateful for a reference to a source that offers a fully explicit proposal of this type. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Fri Nov 27 15:50:40 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Fri, 27 Nov 1998 15:50:40 +0000 Subject: reversal of merger, proposal (2) In-Reply-To: Message-ID: On Fri, 27 Nov 1998 Alexis Manaster Ramer wrote: > I think that there is at least one alternative. Truly > expressive forms are often pronounced somewhat differently > from non-expressive forms with the same phonemes. As Labov > points out, for example, the f-word in English often has > the /f/ made with the lower lip tucked in. Sapir's > paper on "abormal speech" in Nootka contains several > examples. > What would fit > all the facts cited is the theory that there were two > somewhat different articulations of what is written > as , which I will write as and , and that > only one of these, (x2), went to . Note that the strikingly > anomalous development of to , *, > can be explained if we assume that, once the force of the > dim. suffix - . I agree that this interpretation of the Basque case is perfectly possible. However, it seems less economical than Michelena's proposal, in requiring speakers of the Gipuzkoan dialect to construct and maintain a contrast between two different kinds of [esh], a contrast which is not observed or recorded in any variety of Basque. Moreover, such a contrast would have produced an immensely crowded sibilant system. Quite apart from the further contrast proposed here, Basque already has a contrast among three voiceless sibilants: a laminal (notated ), an apical (notated ), and a palato-alveolar (notated ), and it also has the three corresponding affricates, notated , and . This is already a surprisingly crowded system for a language which has only a very modest number of phonemes to start with. And note further that this proposal would have required speakers to change the articulation of to in words in which the expressive value of had been lost: hence it requires that speakers should always have carefully distinguished between expressive and non-expressive occurrences of , which is precisely what Michelena was proposing in the first place. > Another possibility is that the expressive 's that do > not change to are due to interference from other dialects. > There is in general little doubt that many expressive > features do come from external sources. E.g., in Polish > a few words which normally have get in expressive > forms, at least for some speakers, e.g., , an > obsolescent word for 'young female servant' vs. > roughly 'whore', 'poverty' vs. , used to > express sympathy or the like when talking about some > particular people's poverty, etc. In this case, we > happen to know that and are different dialect > realizations of older /e:/, which earlier in the century > were in competition for standard status. In general, > won this fight, thus making eminently suitable to > an expressive function. Yes, but we are not here dealing with a handful of exceptional words. With just one or two exceptions, like the second in , *every* instance of original (expressive) remained unchanged, while *every* instance of non-original underwent backing. The process was completely systematic, not merely sporadic. And no known dialect of Basque exhibits a contrast between two types of , nor does any dialect have a reflex of expressive which is different from , except that word-initial develops regularly to the affricate in some varieties. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk From larryt at COGS.SUSX.AC.UK Fri Nov 27 15:26:03 1998 From: larryt at COGS.SUSX.AC.UK (Larry Trask) Date: Fri, 27 Nov 1998 15:26:03 +0000 Subject: reversal of merger, proposal (2) Message-ID: Alexis Manaster Ramer is temporarily cut off from this list because of a fault, and so he's asked me to forward the posting below. If you want to reply, please reply to him and not to me. Larry Trask COGS University of Sussex Brighton BN1 9QH UK larryt at cogs.susx.ac.uk ---------- Forwarded message ---------- Date: Fri, 27 Nov 1998 08:30:21 -0500 (EST) From: manaster at umich.edu To: Larry Trask Subject: Re: reversal of merger, proposal (2) I think that there is at least one alternative. Truly expressive forms are often pronounced somewhat differently from non-expressive forms with the same phonemes. As Labov points out, for example, the f-word in English often has the /f/ made with the lower lip tucked in. Sapir's paper on "abormal speech" in Nootka contains several examples. What would fit all the facts cited is the theory that there were two somewhat different articulations of what is written as , which I will write as and , and that only one of these, (x2), went to . Note that the strikingly anomalous development of to , *, can be explained if we assume that, once the force of the dim. suffix -. Another possibility is that the expressive 's that do not change to are due to interference from other dialects. There is in general little doubt that many expressive features do come from external sources. E.g., in Polish a few words which normally have get in expressive forms, at least for some speakers, e.g., , an obsolescent word for 'young female servant' vs. roughly 'whore', 'poverty' vs. , used to express sympathy or the like when talking about some particular people's poverty, etc. In this case, we happen to know that and are different dialect realizations of older /e:/, which earlier in the century were in competition for standard status. In general, won this fight, thus making eminently suitable to an expressive function. On Thu, 26 Nov 1998, Larry Trask wrote: > The case is made for a development in the Gipuzkoan variety of Basque, > by Luis Michelena, in his magnum opus Fonetica Historica Vasca and > elsewhere. [snip] > Hence Gipuzkoan has apparently first merged historical /j/ with and > then reversed the merger, so that only non-historical instances of > underwent backing. [snip] > Consider the word `man'. This has the regular combining form > , and this in turn has the regular expressive variant . > >From this , Basque formed a derivative `poor fellow', > with the rare diminutive suffix <-xo> (diminutive suffixes, being > intrinsically expressive, always contain the special consonants). And > this has developed in Gipuzkoan to the unexpected Michelena's explanation is therefore the following. Instances of > which were expressive in nature clearly retained their expressive > function, and hence they remained palato-alveolar and did not undergo > backing -- because they were part of a wider system of palatal and > palato-alveolar consonants confined to expressive forms. All other > instances of -- that is, those derived from /j/ and those in loan > words -- lacked this expressive value and hence could and did undergo > backing. In the case of , the expressive value of the first > remained obvious, and hence this did not back. But, perhaps because of > the rarity of diminutive <-xo>, the expressive value of the second > was lost, and so this one underwent backing. From henryh at LING.UPENN.EDU Fri Nov 27 14:28:28 1998 From: henryh at LING.UPENN.EDU (Henry M. Hoenigswald) Date: Fri, 27 Nov 1998 09:28:28 -0500 Subject: Merger reversed Message-ID: Reversed merger is not a matter for empirical confirmation or disconfirmation. In their haphazard but productive fashion, the nineteenth-century scholars devised a framework (still uncritically used, and, incidentally, uncritically used by them in the first place) in which there is no room for reversal of merger by 'sound-change' (='a replacement, along a line of descent, such that it can be stated without naming particular morphs'). By their lights, there was either no merger, or there is more than one line of descent. There are no doubt other possible frameworks but they have not been explored; maybe for good reasons maybe not. HMH From bwald at HUMNET.UCLA.EDU Sat Nov 28 02:05:36 1998 From: bwald at HUMNET.UCLA.EDU (bwald) Date: Fri, 27 Nov 1998 18:05:36 -0800 Subject: reversal of merger, proposal (2) Message-ID: Larry Trask's example of Basque "unmerger" is quite interesting, and not unexpected for anyone who recognises the reality of synchronic phonological(= morphophonemic) processes. He explains the gist as: > For example, `bull' has an expressive variant > `little bull', and `corner' has the expressive variant > or `nook, cosy little place'. Such expressive >palatalization was formerly pervasive in Basque, and it survives today >in some varieties. Similar (and probably related) expressive palatalisation is used in Spanish, e.g., in talking to babies and forming nicknames, e.g., Jo*s*e > *ch*e, vamo*n*os > vamo*ny*o etc. The relevant (part of the) Basque expressive rule seems to be: s > sh/ch (by /s/ I mean the phoneme written "z") If I understand right, the rule remains in the relevant area as an alternation between s and sh/ch in affected lexical items, keeping it transparent and alive. This souce of (t)sh, on a synchronic level, is completely distinct from the historical phonemes sh and ch, which do NOT alternate with s (and NEVER did). [Larry implied above that it is not STILL a productive affective rule in some Basque areas but was not clear on whether it is in the relevant area -- and, in any case, the issue would only be relevant at the time that sh-backing arose; he seems to take it for granted, so I will too.] I did not quite get the following out of Larry's account of Michelena, who seemed not to have any tool of analysis that goes beyond surface phonetic changes, but it seems clear from the data that the sound change of sh > x is easily expressed as arising as an ordered rule which precedes the affective rule. The affective rule is not affected because its INPUT is /s/ not /sh/. If you cling to the phonetic surface for sound change (the traditional concept), well, the affective rule operates on the surface as a MORPHOPHONEMIC RULE FOLLOWING THE SOUND CHANGE. IE, both the affective rule and the sound change operate on the "surface", and the sound change operates FIRST. Larry notes: >Consider the word `man'. This has the regular combining form >, and this in turn has the regular expressive variant . As we expect. Then he notes: >>From this , Basque formed a derivative `poor fellow', >with the rare diminutive suffix <-xo> (diminutive suffixes, being >intrinsically expressive, always contain the special consonants). And >this has developed in Gipuzkoan to the unexpected , in >which the first instance of has failed to back while the second >instance has undergone backing to the velar/uvular [x]/[X], notated . OK. So we get historical confusion for this particular word, so that the expressive origin of the second /sh/ became obscure for some reason. If I understand the above account right, -xo is RARE, and it does NOT alternate with -zo. So, the origin of its /sh/ in an *affective rule* is not obvious. It's a crap-shoot whether the affect of the rest of the word will save this /sh/ from backing -- and the lexical rarity doesn't help build up a case one way or the other. (Still, if the -xo morpheme generally tended to back, then that would support the idea that it is the morphophonemic alternation per se that protected /sh/ from backing, and NOT the "affect" of the /sh/.) According to Larry: >Michelena's explanation is therefore the following. Instances of >which were expressive in nature clearly retained their expressive >function, and hence they remained palato-alveolar and did not undergo >backing -- because they were part of a wider system of palatal and >palato-alveolar consonants confined to expressive forms. As I said, it's worth explicitly mentioning that such forms ALTERNATE with non-palatals. The palatals subject to backing do NOT have such alternations. That's ALL we need to know to make this case similar to other cases of "unmerger" in which a morphophonemic alternation, captured in a phonological rule, distinguishes one phoneme from another but allows them to overlap phonetically in pronunciation. The important fact is that the rule is MORPHOPHONEMIC. The fact that the rule is "affective" is basically IRRELEVANT, as far as I can see, except for the point I make below. With regard to affective rules, something I discounted as relevant above in favor of the alternation, there is something of further interest about it. If in the relevant dialects, there is NO palatal except by the affective rule, then it is NOT like Spanish affect. Spanish affective palatalisation rule does NOT create segments that do not exist independently as lexical phonemes, i.e., /ch/ and /ny/ (also f > p and other irrelevancies which are part of the same affect). Presumably, the Basque rule does. All I can suggest about that is that languages vary as to whether affective (and onomatopoetic) words are constrained by canonical phonological rules. It is notable that Spanish onomatopoetic words for sounds (as represented, for example, in cartoons) can end in stops, while canonical words cannot. At this point we come to the boundary between words and conventionalised sounds. Forms like "bop!" and whatever in Spanish certainly seem to be composed of phonemes, but violate word canonicity. Forms like English "whew!" and "tsk tsk", when pronounced, are NOT composed of phonemes, and there is not much gained by calling them "words", though you can call them what you like, as long as you recognise what's going on. P.S. Strange things happen in Spanish. Speakers from Mexico City who cannot recognise the difference between "ch" and "sh" in English, nevertheless have at least one word with [sh], "Xola", the name of a street in DF, that they pronounce quite distinctly from "chola" (with [ch]). In classical phonological theory this is a phonemic minimal pair, but is quite clear that the pronunciation "Xola" is exceptional for the speakers, as if some kind of "affect" on [ch] (actually assumed to be an Aztec word by the speakers and thus treated with respect -- affect?) From mcv at WXS.NL Sat Nov 28 03:48:35 1998 From: mcv at WXS.NL (Miguel Carrasquer Vidal) Date: Sat, 28 Nov 1998 03:48:35 GMT Subject: reversal of merger, proposal (2) In-Reply-To: Message-ID: bwald wrote: >Similar (and probably related) expressive palatalisation is used in >Spanish, e.g., in talking to babies and forming nicknames, e.g., Jo*s*e > >*ch*e, vamo*n*os > vamo*ny*o etc. The relevant (part of the) Basque >expressive rule seems to be: > > s > sh/ch (by /s/ I mean the phoneme written "z") The phoneme written (apical sibilant) behaves in exactly the same manner (sagu "mouse" ~ xagu "little mouse"). >If I understand right, the rule remains in the relevant area as an >alternation between s and sh/ch in affected lexical items, keeping it >transparent and alive. This souce of (t)sh, on a synchronic level, is >completely distinct from the historical phonemes sh and ch, which do NOT >alternate with s (and NEVER did). Such historical phonemes *x and *tx (/S/, /tS/) do not seem to exist. There are a couple of other facts that need to be taken into account. Basque does have a non-expressive palatalization rule, to the effect that in a sequence -(V)iCV- the consonant C tends to palatalize, at least if it's one of: /t/ /dit,ut/ "I have them" /s/ /giSon/ "man" /s_/ /iSil/ "quiet" /ts/ /itSal/ "shadow" /ts_/ /itSu/ "blind" /n/ /ban,o/ "but" /l/ /ol,o/ "chicken" (/d/, /b/, /p/, /g/, /k/, /r/ and /rr/ do not palatalize). The palatalization of /t/, /s/ and /ts/ seems to be optional, or less general (Larry?). This may of course also be relevant in the case of the word . Another aspect is the rule (in Gipuzkoan) that initial x- /S/ > tx- /tS/ (and /tS/ does not back, as didn't in Castillian). The fact that "soap" (< OCast. xabon(e)) becomes [xaBoi] would suggest that the change /S/- > /tS/- occurred after the change /S/ > /x/. On the other hand, I notice that this rule too seems to be subject to the effects of the affective/non-affective sense of the word. Words like "dog" (originally "little dog") and "calf" (from "calf", in disuse) are affected, but if my Gipuzkoan doesn't fail me, "torito" (as opposed to "toro") and "mouslet" (as opposed to "mouse") are not. I'm not sure what to make of that. ======================= Miguel Carrasquer Vidal mcv at wxs.nl Amsterdam From manaster at UMICH.EDU Sat Nov 28 18:40:29 1998 From: manaster at UMICH.EDU (Alexis Manaster-Ramer) Date: Sat, 28 Nov 1998 13:40:29 -0500 Subject: Undoing a merger? Message-ID: On Fri, 27 Nov 1998, Larry Trask wrote: > I agree that [AMR's] interpretation of the Basque case [as involving > two kinds of [esh] is perfectly possible. However, it seems less economical than Michelena's proposal, > in requiring speakers of the Gipuzkoan dialect to construct and maintain > a contrast between two different kinds of [esh], a contrast which is not > observed or recorded in any variety of Basque. > > Moreover, such a contrast would have produced an immensely crowded > sibilant system. I have to agree, barring new evidence. > > Yes [re AMR's proposal that we are dealing with dialect interference], , but we are not here dealing with a handful of exceptional words. > With just one or two exceptions, like the second in , > *every* instance of original (expressive) remained unchanged, while > *every* instance of non-original underwent backing. Here, I think I did not make myself clear. What I am saying is that exceptional-looking expressive phonology is often easily explained as involving dialect interference. The Polish example I cited involves only a few examples in my speech, to be sure, but there are many other examples that are more productive, and if fact even in my speech a much more general e -> i rule exists, along with several other rules, as an expressive device for mocking "peasant" speech. There are plenty of expressive phonological devices in lgs and many must surely come from dialect or even foreign-lg contact. One last point: the Basque example is NOT an example of a merger that was undone, because by assumption (Michelena's and Larry's, not mine) the /s^/ which changed to /x/ and the one did not were the same to begin with. Rather, if valid, this would be an example of something else, namely, the resistance of expressive vocabulary to otherwise regular sound changes, of which many examples have been proposed over the years. AMR