IE pers.pron. (dual forms)

[Jens Elmegaard Rasmussen]

On Sun, 16 May 1999, Patrick C. Ryan wrote:

>> In the case of the non-neuter dual, the Sanskrit form -a:(v)
>> corresponds fully with the evidence from the other languages, only
>> elsewhere the other stem-classes use a different morpheme that can
>> everywhere continue an IE *-e.

> That, IMHO, is a big "ONLY"; "correspond" means something a little more
> restrictive to me. I fail to see how OCS C-st. -i, and u-stem -y  (for nom.
> masc./fem.) "corresponds".

I meant to speak of the _thematic_ non-neuter dual, cf. the following.

>> So, the other languages distinguish o-stems from non-o-stems in this
>> point, Sanskrit does not. English has -s in the genitive sg. and pl., most
>> related languages only in the sg.  The two problems are quite parallel, and
>> analogy is known to be the answer in the latter case, what's wrong with
>> suspecting it in the former?

> There is nothing wrong with a suspicion that is in keeping with the data.

That's the point! Then also Slav. -i and -y are analogical forms
just like the Skt. ones. In all probability the analogy was already
completed in the protolanguage: Just as o-stems form non-ntr. dual in
*-o:, thus i- and u-stems form *-i: and *-u:.

>> I'm not saying the non-neuter was _originally_ *-e, I'm saying it was
>> (or things look as if it was) in the IE protolanguage. I am not sure
>> there were no vowel-initial roots, it is mostly very hard to prove that
>> something as hazy as *H1 was not present. On the other hand, there is
>> absolute certainty that IE did have vowel-initial affixes. You may take
>> the gen.sg. ending; whether you want to posit *-os or *-es, there
>> is no place for "-Hos" or "-Hes";

> Sorry that I did not make clear that my objection to *-e was principally
> directed towards the termination of the perfect.

Okay, but we _were_ talking about the dual.

> Beekes reconstructs most
> dual forms as containing [H{1}] and, formwise, I have no problem with that.

So, for the dual *-e, it _would_ be a problem for you if it did not have
a laryngeal, as sva'sa:rau indicates it did not.

> As for the genitive, I reconstruct *-s(V), with the final vowel falling off
> due to stress-accent of a preceding syllable: "Ce/oC + sV -> Ce/o-"Ce/os.

With accent shifting onto a vowel that was not there?

>>  I have managed to explain the nom.pl. *-es from an earlier
>> vowel-less sequence *-z-c (two different sibilants, one marking the
>> nom., the other the pl., in that order, structurally parallel with the
>> acc.pl. in orig. *-m- + sibilant),

> Why not simply *-s(V)s(v)?

In the reading *-ss, no objection, except that we do need two different
sibilants for other reasons.

>> the calculation giving at the same time phonetic explanation of all the many
>> other oddities of the nom. pl. forms, esp. the type in *-'-or-es, *-'-on-es:
>> Why is the -o- not lost? Why is it o, not e? Why is it not long? Why has the
>> e not been lost? And, of course, why is the e not accented?

> There is no doubt that these are all good questions.

>> All of this is explained by **-z-c where there was no vowel to shift
>> to: An unaccented -e- is first reduced to -o-; then the nom. sibilant
>> ("-z-") lengthens (result now *-'-o:r-zc); then short unstressed
>> vowels are lost (but this form does not have any, so the rule operates
>> vacuously here); a long vowel is shorted before word-final
>> triconsonantal clusters [containing the nom. sibilant].

> For some languages, perhaps. But for IE, a tri-consonantal cluster of this
> form is not likely to have been a realized phenomenon at *any* stage of IE.

So, the effects I say they have caused on the surface forms come from
elsewhere? From where? And why are precisely three consonants of a
structure pointed to by the morphology excluded by your ruling? Is it the
number three in itself that is excessive? Or is it the sibilant character
of two of the elements? Or is it my assumption that they were not
identical? Would it help if I said that, for the present purpose, they may
as well be identical since their difference is irrelevant here?

> I think one important factor that should be considered in these equations is
> the example Beekes gives on p. 195 of Greek <o{'}sse>, 'pair of eyes', which
> he derives from *ok{w}-iH{1} THROUGH *ok{w}-ye. We could, just as easily,
> posit a dual in -y and forget about the laryngeal.

In this particular case we could - for Greek. But not for Slavic oc^i
(would have reduced i), nor for Arm. ac^'k' (would not be a-stem, gen.
ac^'ac'). And especially it would not give Skt. -i: with length in the
ntr.du. of cons.-stems.

>> Neither Beekes nor I see the i of *-iH1 as underlyingly syllabic;

> If there is a school that does not accept /i/ as syllabic, I suggest you
> think about changing schools.

>> in phonemic terms it may just as well be given the notation *-yH1. And,
>> hurrah, it is a strong case, i.e. contains no underlying vowels.

> In view of Sanskrit o-stem <yuge:{'}>, I am not sure what the cause for
> jubilation is.

I am pleased by the fact that there is no rightward accent shift in
the ntr.du. of mobile paradigms. Don't look at the o-stems for control,
they never change the accent.

>> *H2ner-y cannot give Gk. ane'r-e;

> In view of Gk. o{'}sse, why not?

There is no such rule. The /y/ would syllabify and yield **ane'r-i. That's
what happened in the loc.sg. *p at 2-te'r-i > Gk. dat. pate'ri, Skt. loc.
pita'ri.

>> u-stem *-u-y cannot give Skt. -u:.

> Why cannot su:nu{'}- + -y yield su:nu:{'} with compensatory lengthening?

There is no such rule. In one instance, a stem amu- got segmented off by a
funny analogy in the inflection of the pronoun asau 'that one' (acc.sg.M
amu-m) and was used in the formation of a pl. with /-y/, this giving
ami:, not **amu:.

>> What _is_ the basis of your decision in favour of *-y ??

> The majority of the attested forms, there being no /i/ in IE..

For most positions, you are right: The original difference between the
morphophonemes /i/ and /y/ are neutralized almost across the board and can
therefore be represented by one phoneme. However, that is not the point
we're discussing; we're arguing about the presence or absence of a
laryngeal in the ntr.du., remember? And in this point Skt. yuge' 'two
yokes' passes judgment, for this form is sandhi resistent ("pragrhya")
and so _must_ have ended in a laryngeal.

>> It may be illogical to abbreviate wordforms when forming enclitic variants
>> of them, but many languages plainly do that.

> Not in my opinion. I believe that emphatic variants are marked by expansions
> of the underlying forms found in enclitics.

That is often the case, perhaps mostly, but not always: Is howdydo an
expansion of howdy? And how-do-you-do an even further expansion?

>> It's like numerals and greetings, you get all sorts of reduced
>> shapes in allegro speech, since people already understand the
>> message at the beginning (sometimes even before).

> Some holes do not improve with additional digging.

>  I believe it is beyond unreasonable to suggest that jam is
> a reduction of jamui! Simple always comes before complex.

So this is wrong too? Well, let me tell you that the Danish decadic
numerals 40-90 were formerly two or three syllables longer than now. A
hundred years ago, you could only write

fyrretyve, halvtredsindstyve, tresindstyve, halvfjerdsindstyve,
firsindstyve, halvfemsindstyve,

but in my lifetime they have always been as short as

fyrre, halvtreds, tres, halvfjerds, firs, halvfems;

and informed scholars know of _no_ phonetic rule to delete word-final
-tyve or -indstyve in this language. These are true abbreviations.

It seems to me that there just are some solutions you do not _like_ on
apriori grounds. Too bad, then, if the material at hand points in that
direction, I am still gonna accept them, even if I do not like them
myself. I am no authority on taste in scholarly solutions, nor do I see on
what grounds anybody else could be. Conversely, there also appear to be
some solutions you would _like_ to be true, even if the material at hands
points unambiguously away from them. I'll be doubting them very much
until, by some unlikely turn of events, the hard evidence turns out to be
illusory and is overruled by a new and better possibility. In the case of
*-e or *-H1e for the non-ntr. dual, not much is needed (but none is
present yet). For *-y or *-iH1, it takes miracles to vindicate the former
over the latter. And that wordforms can be shortened is simply beyond
dispute.

Jens