[Lingtyp] Non-marking for argument roles
David Gil
dapiiiiit at gmail.com
Fri May 9 08:19:33 UTC 2025
Hi Sonja,
Thanks for your comments. Here are two examples of my Yali data
illustrating the results that I reported on:
(1) [Picture of clown drinking from a glass while reading a
book]
Puahun buku naruk
Clown book consume:real:prs.prog
84% of subjects accepted the sentence as a true
description of the picture
(2) [Picture of woman pushing a car]
Mobil heap mealtil laruk
car woman push go:real:prs.prog
79% of subjects accepted the sentence as a true
description of the picture
In (1), buku 'book', which bears a peripheral thematic role in the given
situation, is encoded in the same way as a patient would be; this is the
Yali equivalent of the English 'The clown is drinking the book', which
speakers of English consistently judge as semantically ill-formed. In (2),
the relative order of mobil 'car' and heap 'woman' is "reversed", without
any formal signaling of the reversal; this would be the equivalent of the
English 'The car is pushing the woman'. As you can see, both sentences are
accepted as felicitous descriptions of the relevant pictures.
(A couple of methodological observations. First, the experiment contains a
number of distractors, in the form of picture-sentence pairs in which the
sentence cannot, under any circumstances, be construed as a felicitous
description of the picture. Subjects who, for whatever reason, do not rule
out the sentence as a good description of the picture, are excluded from
the analysis. Secondly, the results reported above were obtained on site,
from speakers in a Yali village, Elelim. In addition, I also ran the
experiment on a group of Yali students at the university in Manokwari and
got very similar results.)
I have no explanation for the apparent discrepancy between my results and
your own impressions of Yali, which are based on years of intensive and
high-quality research. Perhaps, as you suggest, it is due, in part, to the
fact that the test sentences do indeed involve 3rd person arguments in the
present progressive.
For various dialects of Indonesian, some of my experimental results have
also been met with surprise from some quarters. My defense, in the case of
Indonesian, has been to point out that the constructions that are accepted
by experimental subjects while being rejected by experts on the language
tend to be ones that also show up in naturalistic corpora. I wonder
whether this might also be the case for Yali.
Best,
David
On Fri, May 9, 2025 at 4:27 PM Sonja Riesberg via Lingtyp <
lingtyp at listserv.linguistlist.org> wrote:
> Dear David,
>
> I am a bit puzzled with your results for Yali.
>
> Yali has obligatory subject marking on the verb, and object marking for
> animate objects. Subjects are furthermore optionally case marked in
> unmarked SOV order, and obligatorily so in marked OSV order. Object marking
> distinguishes three semantic roles: theme, beneficiary, and target. So Yali
> is actually pretty explicit in telling you who is doing what to whom (see,
> e.g. Riesberg 2018; 2021; and for other Dani languages Bromley 1981,
> Barclay 2008, and Etherington 2002)
>
> Admittedly some of these distinctions can be neutralised in certain
> contexts, i.e., third person acting on third person in the present
> (progressive), which is maybe how your experiment was designed?
>
> In any case I’d say Yali is pretty much the opposite of what Vladimir
> has been asking for.
>
> Best
> Sonja
>
>
> Barclay, Peter. 2008. *A grammar of Western Dani. *München: Lincom
> Publishers.
> Bromley, H. Myron. 1981. A grammar of Lower Grand Valley Dani. Canberra:
> Pacific Linguistics.
> Etherington, Paul A. 2002. *Nggem morphology and Syntax. *Darwin: Charles
> Darwin University PhD dissertation.
> Riesberg, Sonja. 2018. Optional ergative, agentivity, and discourse
> prominence – Evidence from Yali (Trans-New Guinea). *Linguistic Typology *22.1.
> 17–50*. *https://doi.org/10.1515/lingty-2018-0002.
> Riesberg, Sonja. 2021. Introduction to the Yali – English – German
> dictionary with a short grammatical sketch. In Sonja Riesberg, in
> collaboration with Carmen Dawuda, Lucas Haiduck, Nikolaus P. Himmelmann,
> and Kurt Malcher (eds.), *A Yali *(*Angguruk*) – *English – German
> dictionary*, 1–49. Canberra: Asia-Pacific Linguistics.
> https://shs.hal.science/halshs-03436264v1
>
>
>
>
> Am 09.05.2025 um 06:51 schrieb David Gil via Lingtyp:
>
> Dear Vladimir,
>
>
>
> You mentioned Riau Indonesian. While my early writings on Riau
> Indonesian apparently contributed to the impression that this language was
> somehow exceptional with respect to the absence of obligatory thematic role
> encoding, subsequent work suggests that it is anything but a typological
> outlier in this respect.
>
>
>
> In order to situate Riau Indonesian in typological context, and to examine
> the degree to which different languages encode thematic roles such as
> agent, patient, locative, instrumental and so forth by various
> morphosyntactic devices such as word order and flagging, I have been
> conducting a cross-linguistic psycholinguistic experiment, details and
> preliminary results of which are presented in Gil and Shen (2019:5-8) and
> references therein. So far, the experiment has been conducted on 69
> languages. The final results have yet to be written up and published,
> but here are some figures for a handful of languages from the 69-language
> sample:
>
>
>
> English 5.3%
>
> Hebrew 6.7%
>
> Standard Japanese 9.6%
>
>
>
> Standard Indonesian 22.8%
>
> Riau Indonesian 43.4%
>
> Minangkabau 65.0%
>
>
>
> Tikuna 75.8%
>
> Mursi 77.4%
>
> Yali 82.3%
>
>
>
> In the above, percentages represent the degree to which arguments
> associated with different thematic roles can be interchanged (e.g. the
> extent to which an agent can be encoded in the same way as a patient) —
> averaged over 30-plus experimental subjects responding to 16 experimental
> stimuli testing various morphosyntactic configurations. Thus, lower
> percentages represent greater rigidity and obligatoriness in the encoding
> of thematic roles, while higher percentages represent greater flexibility
> and optionality — the state of affairs that prompted Vladimir's query.
>
>
>
> As suggested by the above figures, Riau Indonesian turns out to be
> mid-range with respect to the extent of grammaticalization of thematic
> roles. The true outlier turns out to be English, which scores the
> highest of the 69 languages (albeit not statistically significantly higher
> than a handful of other mostly WEIRD languages). At the other end of the
> scale, a wide range of languages from all over the world exhibit much
> greater flexibility in the encoding of thematic roles than Riau Indonesian.
>
>
>
> The experimental results suggest that the main factor governing the degree
> of grammaticalization of thematic roles is the complexity of the polity
> associated with the language in question: sociopolitical complexity
> correlates positively with grammatical complexity as manifest in thematic
> role encoding. The above correlation also explains why the absence of
> encoding of thematic roles is massively under-represented in the linguistic
> literature, which, even in the 2020s, retains a bias towards languages
> associated with greater sociopolitical complexity.
>
>
>
> Best wishes,
>
>
>
> David
>
>
>
>
>
> Gil, David and Yeshayahu Shen (2019) "How Grammar Introduces Asymmetry
> into Cognitive Structures: Compositional Semantics, Metaphors and
> Schematological Hybrids", *Frontiers in Psychology - Language Sciences* (
> doi:10.3389/fpsyg.2019.02275)
>
>
>
>
> On Thu, May 8, 2025 at 4:50 PM Vladimir Panov via Lingtyp <
> lingtyp at listserv.linguistlist.org> wrote:
>
>> In order to specify my question a little bit: By saying NO MARKING I mean
>> exactly this: NO MARKING AT ALL. E.g. if there is marking not on noun
>> phrases but on the verb or by clitics elsewhere in the clause, then there
>> definitely is marking of arguments. So typical "polysynthetic" languages
>> don't count.
>>
>> V.
>>
>> чт, 8 мая 2025 г. в 10:28, Vladimir Panov <panovmeister at gmail.com>:
>>
>>> Dear linguists,
>>>
>>> I have the following question. Are you aware of any doculects/languages
>>> upon which there is a consensus that semantic roles like S, A, P, R are not
>>> obligatorily encoded, neither morphologically, nor through word order or
>>> adpositions? That is, languages in which the assignment of semantic roles,
>>> if any, is entirely matter of context/pragmatics. The famous Riau
>>> Indonesian comes to my mind. Any other suggestions? Maybe there are
>>> publications dedicated specifically to this problem?
>>>
>>> Best,
>>> Vladimir Panov, Vilnius University
>>>
>> _______________________________________________
>> Lingtyp mailing list
>> Lingtyp at listserv.linguistlist.org
>> https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp
>>
>
>
> --
>
> David Gil
>
> Senior Scientist (Associate)
> Department of Linguistic and Cultural Evolution
> Max Planck Institute for Evolutionary Anthropology
> Deutscher Platz 6, Leipzig, 04103, Germany
>
> Email: dapiiiiit at gmail.com
> Mobile Phone (Israel): +972-526117713
> Mobile Phone (Indonesia): +62-082113720302
>
>
>
> _______________________________________________
> Lingtyp mailing listLingtyp at listserv.linguistlist.orghttps://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp
>
>
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org
> https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp
>
--
David Gil
Senior Scientist (Associate)
Department of Linguistic and Cultural Evolution
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6, Leipzig, 04103, Germany
Email: dapiiiiit at gmail.com
Mobile Phone (Israel): +972-526117713
Mobile Phone (Indonesia): +62-082113720302
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20250509/9678f351/attachment-0001.htm>
More information about the Lingtyp
mailing list