[Lingtyp] Summary numeral in the world languages

Mon Sep 11 14:38:25 UTC 2023

Dear Michael and all,

Thank you for all the engagement in this thread—I truly appreciate it.

IMO, whether the summarizing construction should include an array of strategies is essentially a question of whether the counting items (numeral, number and dual pronouns) are coordinators.

A list of nominals is essentially a coordinating construction, e.g., ’Stephen Joe Virginia’. If the counting items are coordinators, as suggested by Haspelmath (2007), then a summarizing construction should only include examples in which the coordinands and counting item form an NP constituent (e.g., [Stephen Joe Virginia three]np) and exclude examples that do not (e.g., [Stephen Joe Virginia]np [go-three]vp). But I am a bit skeptical about them being coordinators, consider the following examples:

(1) Stephen Joe and Virginia 
(2) three people

(1) and (2) refer to the same group of people. (1) with a coordinator focuses on the differences of referents whereas (2) with a numeral focus on the similarity of referents, i.e., being [+human]. Hence, there is seemingly a functional difference between a general coordinator and a numeral. Ofc, one could claim that numerals in natural languages can actually focus on differences as well (e.g., Stephen Joe Virginia three), so it is a matter of perspectives.

If the counting items are not coordinators, then we can include a large array of strategies, notably those that are not part of the list NP, e.g., dual indexing and appositive construction with NUM-CL. 

(Yes, I have no idea if there is an umbrella term for numeral and number— perhaps a “summarizer” (Croft, 2022) )would be the closest candidate).

Warmest,
Joe Pun Ho Lui

> Michael Daniel <misha.daniel at gmail.com> 於 2023年9月11日 下午9:19 寫道：
> 
> Dear Pun Ho Lui, dear colleagues.
> 
> I wanted to come back to what David wrote in one of the first answers. 
> 
> In this thread, there were mentions of constructions that are usually considered separately - summarizing constructions (as the example from Mongsen Ao, probably also Japhug and Adamawa) and inclusory constructions (as the explicative Mwotlap construction), to which Alexandre François added what looks to me as an unmistakably associative plural construction (his recognitional construction in Mwotlap). I have no knowledge about summarizing constructions, but both inclusory constructions and associative plurals are of course very common cross-linguistically. To the reference to Linchtenber on inclusory constructions in Austronesian, I would add at least Ruth Singer's 2001 survey of them in Australia or discussions of dedicated inclusory pronouns in Mande languages - but of course many other languages also have these. 
> 
> ***
> 
> Now, on the one hand, we see, e.g. from Alexandre's examples, that the three types of constructions can be morphosyntactically similar, at least to some extent. Language-internally, the boundaries between them may be fuzzy, and certainly not all typologists would agree to draw them in the same way (if they would want to draw them, in the first place). So - probably unlike David? - I think that it may make sense to survey them together, as has in fact been going on in this thread, to my view. On the other hand, they may also be completely unrelated, or only some of them may be present in an individual language. 
> 
> In other words - now very much following David (I think) - I believe that, while being prepared to see the three constructions to be formally related in some or many languages, or even considering them as one empirical domain of investigation, one would want to remain very sensitive to the potential differences between these constructions, based on signature properties proposed in earlier of each of these constructions separately. 
> 
> To give just one example, as David pointed out, a very special property of inclusory construction, something that is often considered to be definitional for them, is that there is an unexpected referential property of "strict inclusion" - the reference of the inclusory element includes, and is wider, than the reference of the "added" phrase ("we with you.sg <http://you.sg/>" meaning 'you.sg <http://you.sg/> and I'). Another special property, not mentioned by David (probably because it is too obvious), is that in the vast majority of cases, the inclusory element is either a plural pronoun or plural index on the verb. Associative plurals have their own signature properties (some of which are briefly referred to in Alexandre François' detailed reply), and probably summarizing constructions also have some of their own, which I am unaware of because of my limited knowledge of them. Fuzzy as the boundaries may be in individual languages, to me, these properties are not just part of arbitrary "etic" categories but reflect a certain internal logic of the functional domain in question. If one does not try to keep an eye on these properties, one runs a risk of confusion. 
> 
> As an example of what seems to me to be a confusion, your own initial examples would not, to me, constitute exemplary cases of summarizing constructions - none of them contains a lexical numeral; moreover, arguably, they all can be interpreted as extension of inclusory constructions. I am aware of the fact that they contradict one or both of the properties that I considered signature for inclusory pronouns; but I think many would agree they are not prototypical summarizing constructions, either. As a result, some people, based on your examples, provided examples of inclusory constructions; while David, based on your definition, discarded contributions as not directly related to your query.
> 
> (Personally, I think this is a result of your equating number marking with numerals, and I do not know whether the studies you refer to for "summarizing" numerals explicitly license this. Again, I do agree that it might be a meaningful enterprise to consider all ways of expressing cardinal numerosity ('two' and dual, 'three' and trial) in your study, but one should be prepared to face a very heterogeneous range of morphosyntactic constructions.) 
> 
> Michael            
> 
> 
> вс, 10 сент. 2023 г. в 22:14, Lora Litvinova <loravlitvinova at gmail.com <mailto:loravlitvinova at gmail.com>>:
> Dear all,
> 
> Kugama, an Adamawa language of Nigeria (ISO 639-3: kow; Glottocode: kuga1239), has noun phrases that also summarize the number of elements involved in nominal coordination, see examples below. Structurally, these phrases are possessive constructions with a numeral as a head and a possessive pronominal as a modifier. More information about these constructions can be found in my PhD dissertation (Litvinova 2023) that will be available online after the defense in December 2023.
> 
> (1)   Ɔ̀zūrúwā   ɗéè      Lúkpɛ̂wà    à        Bɛ̀nā     tì          sɛ́ɛ́=kī
> 
>        Ozuruwa  CONJ  Luikpewa   COM   Bena    PROG   fight=NMLZ
> 
>        níLsá̰ā̰=L=rī=L=rē
> 
>        three=GEN=POSS=GEN=3PL.HUM.POSS.ALIEN
> 
>       ‘Ozuruwa, Luikpewa and Bena are fighting. All three of them.’
> 
>  
> (2)  ā       kā=tí         dìŋ=L           kísā          ɗéè       dìŋ=L           zɛ᷇m
> 
>       1SG  see=PFV   tree=GEN baobab     CONJ    tree=GEN  tamarind
> 
>       à        dìŋ=L             nɔ̄rī        à           dìŋ=L            gbɛ̀lɛ́L                   tì          gɛ̀pí=ɛ̄
> 
>       COM   tree=GEN  shea      COM      tree=GEN  gbele_tree        PROG  grow=NMLZ
> 
>       ɓáā=L        háákī      níLhɛ̰̄=L=rɛ̂
> 
>       in=GEN   farm        four=GEN=3PL.NHUM.POSS.ALIEN
> 
>      ‘I saw a baobab tree, a tamarind tree, a shea tree and a gbele tree (the species of this tree is unknown) growing in the farm. All four of them.’
> 
>  
> Reference: Litvinova, Lora. 2023. A grammatical analysis and documentation of Kugama (Wam), an Adamawa language of Nigeria. Paris: INALCO PhD thesis
> 
>  
> Best wishes,
> 
> Lora
> 
> 
> On Wed, Sep 6, 2023 at 8:50 AM Pun Ho Lui <luiph001 at gmail.com <mailto:luiph001 at gmail.com>> wrote:
> Dear All,
> 
> Haspelmath (2007) and Croft (2022) discussed a coordinating construction in which a numeral “summarizes” the number of referents in a list. There are different strategies in which the numeral behaves this way, e.g., a free numeral (1); a dual affix on a coordinand (2); a dual pronoun in apposition with the list+verb with dual marker (3). These numeral may be mono-syndetic or bi-syndatic.
> 
> 
> (1) Zaozou
> ŋu55-mu55 na53 phiɛ33
> 
> 1-PL[EXCL] two father_and_child 
> 
> “we two (exclusive), my daughter and I.”  (Li, 2020)
> 
> 
> (2) Kham 
> syar sono:h pusum-ni 
> 
> louse and flea-DL
> 
> ‘the louse and the flea’ (Watters, 2004)
> 
> 
> 
> (3)  Mapudungu
> (iñché) eymi    inchiu   i-y-u
> 
>  I          you:SG we:DU eat-IND-1NONSG-DU
> 
>   ‘You and I ate.’ 
> 
> 
> Languages with this construction I know are Zaozou, Kham, Mapudungu, Alto Perené, Bangla, Cantonese, Mandarin, Papuan Malay, Yakut, 
> Inari Saami, Mongolian, Classical Tibetan, Huallaga Quechua, Wardaman, Khanty, Vedic Sanskrit, Mparntwe Arrernte, Daga, Mapudungu, Enets, Kham and Hualapai
> 
> 
> I am wondering if there are other languages sharing similar constructions.
> 
> Thank you.
> 
> Warmest,
> Joe Pun Ho Lui
> 
> 
> 
> 
> 
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org <mailto:Lingtyp at listserv.linguistlist.org>
> https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp <https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp>
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org <mailto:Lingtyp at listserv.linguistlist.org>
> https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp <https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20230911/41bd137f/attachment.htm>