[Lingtyp] Coexpression of source and agent

Mon Jul 23 08:59:24 UTC 2018

 Dear Daniel and other readers,

A morpheme reduces the number of possible interpretations of an utterance
on the part of the listener, but morphemes do not "express"anything. This
is a fallacy very clearly explained in Reddy 1979 "The conduit metaphor".
Thus, the morpheme <=sang> in Atong (Tibeto-Burman, Northeast India) can be
interpreted as a morpheme indicating that someone or something goes in a
certain direction (mobilitative interpretation), or that someone or
something is an instrument (instrumental interpretation). The morpheme
itself does not "express" anything; rather, the speaker uses it with a
certain purpose (to restrict the amount of inference on the part of the
listener about the purpose of the utterance), and listener has to infer
what that purpose is based on his/her knowledge of the context in which it
occurs, which includes textual context, situational context and real-world
knowledge.
Thus, without context, the Atong expression <rong=sang> (rock=sang) can
either be interpreted as 'to/from the rock' or 'with a rock (as an
instrument, no accompaniment).
I make it seem as if the categorical boundaries between the different
interpretations of this Atong morpheme are very sharp, but in fact they are
not. Other interpretations of this morpheme are sometimes possible, and
this makes it difficult to guess the speaker's communicative intentions
even within the context of an utterance in which <=sang>.

Another example is the use of the morpheme <house> in English. This
morpheme can be used to refer to a variety of different objects. For
example, when a speaker lives in a cardboard box and says: "This is my
house" the most logical interpretation is that the morpheme <house> refers
to the cardboard box.

Thus, for me, morphemes do not express anything, but have different
possible interpretations (see van Breugel, 2014. A Grammar of Atong.
Leiden, Boston: Brill).  When the possible interpretations of a morpheme
are limited, they can easily be listed and explained. However, this is not
always the case.

Best regards,

Seino van Breugel
https://independent.academia.edu/SeinovanBreugel

On Mon, Jul 23, 2018 at 7:03 AM, Daniel Ross <djross3 at gmail.com> wrote:

> Hi Eitan and thanks for the reply,
>
> The problem is the analogy itself:
>
> Express is not to lexify as "coexpress" would be to colexify. Colexify
> refers to multiple meanings within a lexeme, similar to "cohabitate"
> referring to multiple people in a household. Note that timeshare owners
> co-own a timeshare but do not co-habitate the timeshare, just like
> ambiguous expressions do not "co-express" multiple meanings simultaneously.
>
> The difference is also much clearer in another tense, which reveals that
> "(co)lexify" is pseudo-diachronic:
>
> "English has (co)lexified comitative and instrumental meanings in a single
> preposition."
>
> ?English has co-expressed X and Y as...???
>
> On par with colexify we might be able to use "coexpressionize", but that
> is of course awkward.
>
> The awkwardness of "syncretize" also seems to disappear (for me at least)
> in the past or perfect. ("Neuter nominative and accusative have
> syncretized.")
>
> My main point is that I don't see any reason why "coexpress" is a better
> candidate than, e.g., "syncretize", given that the meaning is not
> transparent or etymologically natural in either case. (Setting aside other
> reasons why "syncretism" may not fit the original question.)
>
> If for convenience we end up using "coexpress" I don't strongly object,
> but I just wanted to be on record as emphasizing how much that sounds to me
> like a description of puns! "Coexpress", literally, should refer to
> expressing two meanings at once, not the possibility for expressing two
> different meanings at different times with what happens to be the same form.
>
> Consider possible parallels to "coexpress" for how awkward they would be:
> "co-say", "co-mean", "co-indicate", "co-occur", etc. Rather than
> "together", the adverbial meaning needed is "alternatively", and I think
> that would mean "ambi-" (or possibly 'equivalently'-- "equi-"?), which
> makes sense because we're talking about ambiguity of forms, not literal
> (simultaneous) multi-expression of forms. Thus, "ambi-express" (or
> "equi-express"?).
>
> Instead of "express" we would also substitute a term parallel to "lexify",
> as in "co-grammaticalize" (or maybe "co-encode"), or the awkward
> "expressionize" above.
>
> Daniel
>
> On Sun, Jul 22, 2018 at 8:18 PM, Eitan Grossman <
> eitan.grossman at mail.huji.ac.il> wrote:
>
>> Hi all,
>>
>> I don't share Daniel's (and others') reservations about the term
>> 'coexpress(ion).'
>>
>> First of all, it's handy to have a verb that can be predicated of a noun
>> referring to a linguistic item ("Hebrew le- coexpresses goal and
>> recipient"), which isn't the case for 'polyfunctional(ity)' or the like,
>> which is unwieldy at best ("Hebrew le- is polyfunctional with respect to"
>> or "The polyfunctionality of Hebrew le- includes"). Also, when the subject
>> is singular, there isn't much room for misinterpretation, so the intended
>> reading is easily coerced even if the reader is unfamiliar with the term.
>>
>> Second, the term already exists in biology in the context of genes, where
>> coexpression means 'the simultaneous expression of two or more genes.' If
>> the analogy here is between genes and meanings/senses, and 'simultaneous'
>> is taken to mean 'associated with the same signifier,' then it works all
>> right. I'm sure that it is easy to pick apart the analogy, of course. But
>> interestingly, geneticists have been making coexpression maps of genes,
>> which may be comparable to our semantic maps.
>>
>> Finally, as others have pointed out, the terms 'colexify' and
>> 'colexification' are already pretty established in lexical typology, so
>> 'coexpression' is not itself a big innovation.
>>
>> Eitan
>>
>>
>>
>> Eitan Grossman
>> Lecturer, Department of Linguistics/School of Language Sciences
>> Hebrew University of Jerusalem
>> Tel: +972 2 588 3809
>> Fax: +972 2 588 1224
>>
>> On Mon, Jul 23, 2018 at 2:26 AM, Daniel Ross <djross3 at gmail.com> wrote:
>>
>>> Dear all,
>>>
>>> While I would agree about the importance of clear terminology, I would
>>> suggest that it is not "coexpress(ion)" in this case.
>>>
>>> Martin is correct that "co-" indicates "together", but more specifically
>>> it relates to a collective interpretation rather than a distributive
>>> interpretation. "He and I read a wrote" could be rephrased as "He and I
>>> cowrote" a book if and only iff we wrote parts of the same book, not if we
>>> each wrote separate books.
>>>
>>> I think an analogy may be in order to show why "coexpression" is
>>> confusing in that sense. Consider the extreme confusion that "biannual" and
>>> "biennial" produce, with readers uncertain about whether the event will
>>> take place once every two years or twice every year. Compare this to
>>> "coreference", where we say that two pronouns are "coreferential" if they
>>> were to the same referent, versus saying that "he" is "coreferential" when
>>> it refers to two different male humans (e.g. John and Bill):
>>>
>>> a) He[i] saw himself[i] in the mirror. = coreferential
>>> b) He[i] saw him[j] in the mirror. = non-coreferential
>>>
>>> The proposed usage of "coexpress" would be more like (b), which I find
>>> to be very confusing.
>>>
>>> To me, if a word "coexpresses" two ideas, then it would literally be
>>> doing that in actual usage, as in a pun or zeugma. Indeed, "coexpress"
>>> could only make sense in this way in habitual ('simple present') usage, as
>>> in "Malay dari coexpresses source and agent", but never referring to a
>>> specific instance as in "In this sentence, Malay dari coexpresses source
>>> and agent", which is what the prefix "co-" would suggest to me! Indeed, it
>>> never actually COexpresses multiple meanings at all!
>>>
>>> Instead of "coexpress" we would need a term that means something more
>>> like "variably express" or "alternatively express". The closest I can think
>>> of at the moment would be "ambiexpress", not that is easily rolls of the
>>> tongue, and unfortunately suggests only two alternatives, not multiple. In
>>> that case, I would second Tom's comment about "poly-functional" (or
>>> "multi-functional"), although the replies after that did suggest the need
>>> for a broader typology. Maybe "poly-express" (or "multi-express")?
>>>
>>> In the end, maybe English doesn't offer an ideal term and "coexpress"
>>> will catch on, but I am unconvinced of its etymological appropriateness.
>>>
>>> Daniel Ross
>>> Ph.D. Candidate in Linguistics
>>> University of Illinois at Urbana-Champaign
>>>
>>> On Sun, Jul 22, 2018 at 1:13 AM, Mattis List <mattis.list at lingpy.org>
>>> wrote:
>>>
>>>> What I find important, and often not very satisfying, regarding
>>>> terminology in linguistics, is that we offer descriptive -- as opposed
>>>> to explanative -- terms for the phenomena we study. Polysemy is
>>>> explanative, as it requires that we know for sure that some word gained
>>>> new meanings through semantic change processes. Homophony is explanative
>>>> as well, as it requires that we identify sound change (resulting in
>>>> merger) or borrowing as the origin of the phenomenon that two meanings
>>>> are expressed by the same word form.
>>>>
>>>> Unfortunately, these are not the only examples where we lack descriptive
>>>> terms in linguistics. I often struggle with terms like "assimilation",
>>>> "shared innovation", "pronominalization", and many others, since these
>>>> terms all do not only describe, but also try to explain the phenomena at
>>>> hand.
>>>>
>>>> We should all work harder in establishing a purely descriptive
>>>> terminology in our field. Explanative terminology should be restricted
>>>> to the situations where we really know what happened.
>>>>
>>>> Best,
>>>>
>>>> Mattis
>>>>
>>>> On 22.07.2018 07:05, Martin Haspelmath wrote:
>>>> > Thanks to Juergen and Tom for discussing my proposal!
>>>> >
>>>> > As for Tom's comment:
>>>> >
>>>> > Well, it may be that "A and B co-VERB C and D" more often means 'A
>>>> and B
>>>> > together VERB C and D', but I think that "co-" in English works just
>>>> > like "together", i.e. that it can also mean 'A and B VERB C and D
>>>> > together' (i.e. C and D are together). For example, Wiktionary gives
>>>> > "coapply" meaning 'apply C and D together'
>>>> > (https://en.wiktionary.org/wiki/coapply).
>>>> >
>>>> > As for Juergen's comment (also copied below):
>>>> >
>>>> > YES! It's really helpful that the term "coexpression" is that it is
>>>> > neutral between vagueness (=underspecification) and ambiguity, because
>>>> > in practice these are often difficult to distinguish (just as polysemy
>>>> > and homophony are difficult to distinguish).
>>>> >
>>>> > If one feels that one can distinguish them, one should of course use
>>>> the
>>>> > more specific terms, but in cross-linguistic studies, they are better
>>>> > avoided (comparativists will not want to commit themselves to claims
>>>> > about particular languages).
>>>> >
>>>> > In my 2003 paper on semantic maps, I used the term
>>>> "multifunctionality"
>>>> > to be neutral between vagueness and ambiguity/polysemy, but
>>>> > "coexpression" is better (see also the brief discussion in the first
>>>> > section of Georgakopoulos & Polis's recent survey article on semantic
>>>> > maps, https://doi.org/10.1111/lnc3.12270).
>>>> > The semantic map model: State of the art and future avenues for
>>>> > linguistic research
>>>> > Best,
>>>> > Martin
>>>> >
>>>> > On 21.07.18 19:33, Tom Payne wrote:
>>>> >>
>>>> >> While I share Martin’s objection to the use of “syncretism”, I also
>>>> >> think “co-expression” is problematic. Words like “co-author” or
>>>> >> “co-operate” imply two actors working together to accomplish one
>>>> task,
>>>> >> e.g., “I co-authored an article with Taeho Jang,” “We co-parent our
>>>> >> children”, “They co-wrote a linguistics textbook,” etc. If
>>>> “co-author”
>>>> >> meant what Martin and Juergen suggest “co-expression” could mean,
>>>> that
>>>> >> would be like one person writing two articles – “I co-authored an
>>>> >> article on tense, and one on aspect” (meaning I was the only author
>>>> of
>>>> >> those two articles).
>>>> >>
>>>> >>    I think “co-expression” normally refers to situations where more
>>>> >> than one form together (“co-operatively”) express one category, like
>>>> >> one might say “/ne/ and /pas/ co-express negation in French.”
>>>> >>
>>>> >>   What happened to “poly-functional”?
>>>> >>
>>>> >> Tom
>>>> >>
>>>> >>
>>>> >>
>>>> >> *From:*Lingtyp <lingtyp-bounces at listserv.linguistlist.org> *On
>>>> Behalf
>>>> >> Of *Martin Haspelmath
>>>> >> *Sent:* Saturday, July 21, 2018 03:20
>>>> >> *To:* lingtyp at listserv.linguistlist.org
>>>> >> *Subject:* Re: [Lingtyp] Coexpression of source and agent
>>>> >>
>>>> >>
>>>> >>
>>>> >> Dear all,
>>>> >>
>>>> >> A side comment on terminology: The term "syncretism" is not only
>>>> >> opaque, ugly and ambiguous (it originally referred to merging of case
>>>> >> distinctions in Indo-European, which was likened to religious
>>>> >> syncretism, in a strange metaphor; it can still have this purely
>>>> >> diachronic meaning referring to Indo-European cases) – it is also
>>>> >> impractical because it does not have a good corresponding verb (cf.
>>>> >> ??"Malay /dari/ syncretizes source and agent").
>>>> >>
>>>> >> Moreover, it is typically associated with inflection (cf. the Surrey
>>>> >> definition: " The term 'syncretism' refers to the phenomenon whereby
>>>> a
>>>> >> single form fulfils two or more different functions within the
>>>> >> inflectional morphology of a language":
>>>> >> http://www.smg.surrey.ac.uk/syncretism/).
>>>> >>
>>>> >> (And syncretism seems to have been construed as a relation between
>>>> >> forms: cf. the original title of this thread "syncretism between
>>>> forms
>>>> >> encoding source and agent" – a very cumbersome formulation.)
>>>> >>
>>>> >> I would like to propose replacing the term "syncretism" by
>>>> >> "coexpression" when it is not used in a context of inflectional
>>>> >> morphology (and maybe also in that context). The term "coexpression"
>>>> >> is transparent and clear – and it can be used for all kinds of
>>>> >> situations where one form corresponds to two meanings or functions.
>>>> >>
>>>> >> It has the transparent corresponding verb "coexpress": "Malay /dari/
>>>> >> coexpresses source and agent".
>>>> >>
>>>> >> This term was first used in our 2014 paper on semantic role
>>>> >> coexpression patterns (Hartmann et al. 2014), and was taken up in
>>>> >> David Gil's recent paper on DO/GIVE coexpression. It was inspired by
>>>> >> Alex François's (2008) term "colexification" (also used in
>>>> >> Johann-Mattis List's new CLLD database on colexifications:
>>>> >> http://clics.clld.org/).
>>>> >>
>>>> >> A colexification pattern is just a special kind of coexpression
>>>> >> pattern – and one might also want to coin the term "coexponence" for
>>>> >> inflectional morphology, for a situation where a single vocabulary
>>>> >> item coexpones two feature values; i.e. for what has been known as
>>>> >> "inflectional syncretism".
>>>> >>
>>>> >> Finally, a semantic map could be called a "coexpression map",
>>>> allowing
>>>> >> us to be neutral between different interpretations (cf. different
>>>> >> terms such as "conceptual map", "cognitive map", "implicational map",
>>>> >> which will confuse many students).
>>>> >>
>>>> >> Best,
>>>> >> Martin
>>>> >>
>>>> >> ***************************
>>>> >>
>>>> >> References
>>>> >>
>>>> >> François, Alexandre. 2008. Semantic maps and the typology of
>>>> >> colexification: Intertwining polysemous networks across languages. In
>>>> >> Martine Vanhove (ed.), /From polysemy to semantic change: Towards a
>>>> >> typology of lexical semantic associations/ (Studies in Language
>>>> >> Companion Series 106), 163–216. Amsterdam: Benjamins.
>>>> >>
>>>> >>
>>>> >>
>>>> >> Hartmann, Iren, Martin Haspelmath & Michael Cysouw. 2014. Identifying
>>>> >> semantic role clusters and alignment types via microrole coexpression
>>>> >> tendencies. /Studies in Language/ 38(3). 463–484.
>>>> >>
>>>> >>
>>>> >>
>>>> >
>>>> >> Dear Martin et al — There may be another reason not to apply the
>>>> term ‘syncretism’ to the phenomenon at issue in this thread: the latter
>>>> presumably involves metaphor and thus constitutes (at least initially) an
>>>> instance of polysemy. In contrast, the term ‘syncretism’, as I understand
>>>> it (I’m not a historical linguist), involves loss of formal distinction
>>>> among paradigm cells, which would thus constitute a case of homophony (with
>>>> respect to the formerly distinctly expressed cells in the now simplified
>>>> paradigm).
>>>> >>
>>>> >> (There doesn’t seem to be a theoretical reason to assume that
>>>> syncretism and polysemy are mutually exclusive, though. If syncretism is
>>>> indeed defined in terms of a morphological merger, then such mergers may
>>>> well occasionally be motivated by metaphor or metonymy. However, that
>>>> doesn’t seem to apply to any of the cases discussed in this thread so far.)
>>>> >>
>>>> >> Now, the proposed ‘coexpression’, as I understand it, would be a
>>>> hypernym of, and coverterm for, polysemy, homonymy, and semantic
>>>> underspecification (as in the case of _linguist_ not distinguishing
>>>> gender). This would lead to the following terminological network:
>>>> >>
>>>> >>                         coexpression
>>>> >>                           |                 |
>>>> >> underspecification           ambiguity
>>>> >> (a.k.a. vagueness)           |            |
>>>> >>                              polysemy       homonymy
>>>> >>
>>>> >> Syncretism would then be a particular instance of homonymy,
>>>> presumably restricted to morphological paradigms, as opposed to homonymy in
>>>> lexical items (with the understanding that sound mergers can cause both
>>>> syncretism and lexical homophony).
>>>> >>
>>>> >> Enough terminology for a Saturday morning! — Best — Juergen
>>>> >
>>>> >
>>>> >
>>>> >
>>>> >
>>>> >
>>>> >
>>>> >
>>>> > _______________________________________________
>>>> > Lingtyp mailing list
>>>> > Lingtyp at listserv.linguistlist.org
>>>> > http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>>> >
>>>>
>>>> _______________________________________________
>>>> Lingtyp mailing list
>>>> Lingtyp at listserv.linguistlist.org
>>>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>>>
>>>
>>>
>>> _______________________________________________
>>> Lingtyp mailing list
>>> Lingtyp at listserv.linguistlist.org
>>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>>
>>>
>>
>
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org
> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20180723/297cae7e/attachment.htm>