Constructions et contexte 2023

15th November, 2023

Bât. A, salle A2 201, Université Paris 8

 

Les constructions, grossièrement définies comme des associations forme-sens, constituent une notion fondamentale pour de nombreuses recherches basées sur l’usage. Cet atelier s’intéresse à la manière de décrire et de modéliser l’utilisation des constructions à l’aide de données de corpus et/ou d’élicitations.

L’atelier réunit trois conférences plénières données par d’éminents chercheurs dans le domaine et quatres présentations en session générale par de jeunes chercheurs. L’objectif est de promouvoir une discussion axée sur les préoccupations méthodologiques et leurs implications pour les modèles de description du langage, que ce soit la production ou la grammaire.

 

Programme

09h30-10h00

Coffee

 

10h00 - 11h00
Susanne Flach
University of Zurich
Association vs. frequency (again) : What usage measure best predicts reading times ?
Usage-based linguistics assumes that linguistic knowledge (as proxied through experiments) and linguistic performance (as proxied through corpora) are tightly connected. However, much of that connection is still unknown, and many methodological questions remain that we shall revisit in this talk.

 

11h00 – 11h30
Michele Cardo
Université Paris Cité
Question sequences as discursive constructions in TED talks
Speakers often use a sequence of questions when giving TED talks. The frequency of this discursive strategy brings into question the motivation behind this linguistic choice and the effect that it has in the facilitation of information exchange. This talk will provide an overview of my recent M1 thesis which posits a classification of question sequence types in TED talks and the possible effect different question sequences might have on an audience. The conventionalization of question sequences in TED talks provides evidence for the idea that they are a construction in their own right which may be parallel to various directive speech acts.

 

11h30-12h00
Cameron Morin
Université Paris Cité
Social meaning in Construction Grammar
In this talk, I will present my recently completed PhD thesis. The thesis aims to improve the explanatory adequacy and internal consistency of the theory of Construction Grammar, by considering why and how to integrate social meaning in a CxG model. To do so, it considers the case of a particularly rare, unusual, and socially salient grammatical feature in dialects of American and British English : the double modal construction. The discussion relies on an in-depth empirical description of double modals based on (i) a corpus study of naturalistic double modals in two unprecedently large databases of American and British geolocated social media data, and (ii) a fieldwork study of the acceptability and social meaning of double modals in four samples of dialect speakers from England, Scotland, and Northern Ireland.

 

12h00-13h30
Lunch

 

13h30-14h30
Eva Zehentner
University of Zurich
English prepositional verbs emerged through constructionalisation

The present paper reports on a quantitative corpus study on the development of prepositional verbs such as insist on, look at, or deal with in the history of English, specifically from Middle to Late Modern English (1150-1900). The main focus is on issues relating to constructionalisation processes and the modelling of (loss of) phrase boundaries in construction grammar approaches.

 

14h30 - 15h00
Emma Kious, Anne Abeillé & Yanis da Cunha
Université Paris Cité
Gender mismatch in nominal ellipsis : the case of French stripping”
Cross-linguistically, within elliptical constructions, mismatches in person and number are usually possible (Abeillé et al. 2014, Bîlbîie 2017). Under stripping, antecedent features may not be matched by the elided material, calling into question identity constraints. Yet, there seems to be a penalty in acceptability in the case of gender mismatches. Merchant (2014) posited an asymmetry of Greek nominal gender features, later confirmed upon by Sprouse et al. (2002) for English gendered nouns (John is an actor and Mary too. vs *Mary is an actress and John too). For Spanish, Aparicio et al. (2014) found a preference for the Nmasc-Nfem order, with the feminine remnant more susceptible to be elided. We conducted a corpus study to provide a more thorough understanding of the construction in French, with corpus-based findings supplemented with experimental data from an online acceptability judgment task, seeking to narrow down the conditions which favor gender mismatch.

 

15h00 - 15h30
Piotr Wyroslak
Université Paris 8
Checklists of joy : Modelling collocational networks of stacked participles

 

15h30 - 16h30
Dirk Pijpops
University of Antwerp
What are alternations ? Definitions and methodological implications
In an alternation study, a researcher contrasts two or more linguistic variants and aims to uncover the motivating factors that determine which one is preferred in a given context. This research set-up has developed into a major methodological paradigm, especially in usage-based linguistics, building on corpus data, experiments and computer simulations (e.g. respectively Szmrecsanyi et al. 2016, Klavan and Divjak 2016, Pijpops 2022). Still, there are several ways of defining the concept of ‘alternation’ in use throughout the literature. Each of these definitions comes with its own methodological consequences, and how you conduct an alternation study is hence crucially dependent on how you define ‘alternation’. In this talk, I will therefore discuss six possible definitions and lay out their theoretical merits and their methodological implications.

 

16h30 - 18h30
Drinks and nibblies

 

References
Bresnan, J. & Ford, M. 2010. Predicting Syntax : Processing Dative Constructions in American and Australian Varieties of English. Language 86, 186-213.
Bresnan, J. 2021. Formal grammar, usage probabilities, and Auxiliary contraction. Language 97, 108-150.

Dabrowska, E. 2013. Functional constraints, usage, and mental grammars : A study of speakers’ intuitions about questions with long-distance dependencies. Cognitive Linguistics 24, 633-665.

Dabrowska, E. 2014. Recycling utterances : A speaker’s guide to sentence processing. Cognitive Linguistics 25, 617-653.
Flach, S. Forthc. Serial verb constructions in English : Formal constraints from a usage-based perspective. Berlin : Mouton.

Flach, S.. 2020. Schemas and the frequency/acceptability mismatch : Corpus distribution predicts sentence judgements. Cognitive Linguistics 31, 609-645.
Goldberg, A. 2006. Constructions at Work. The nature of generalization in language. Oxford : OUP.
Hilpert, M. 2008. Germanic Future Constructions. A usage-based approach to language change. Amsterdam : Benjamins.
Hilpert, M. 2008. New evidence against the modularity of grammar : Constructions, collocations, and speech perception. Cognitive Linguistics 19, 391-411.
Pijpops, D., Speelman, D., Grondelaers, S., & Van de Velde, F. 2018. Comparing explanations for the Complexity Principle. Evidence from argument realization. Language and Cognition 10, 514-543.
Pijpops, D., Speelman, D., Van de Velde, F., & Grondelaers, S. 2021. Incorporating the multi-level nature of the constructicon into hypothesis testing. Cognitive Linguistics 32, 487-528.
Schmid, H.-J. 2016. Why cognitive linguistics must embrace the pragmatic and social dimensions of language and how it could do so more seriously. Cognitive Linguistics 27, 543-557.
Schmid, H.-J. 2020. The Dynamics of the Linguistic System. Usage, conventionalization, and entrenchment. Oxford : OUP.
Zehentner, E. 2021. Alternations emerge and disappear : the network of dispossession constructions in the history of English" Corpus Linguistics and Linguistic Theory 17, 525-561.
Zehentner, E. 2022. Ambiguity avoidance as a factor in the rise of the English dative alternation. Cognitive Linguistics 33, 3-33.