Session 1F

L'un des concepts fondamentaux du langage concerne l'expression de l'espace dynamique. Dans le cadre de la linguistique cognitive, Talmy (2000) propose une typologie des langues reposant sur la saillance de l'expression de la Trajectoire dans les événements de mouvement. Il distingue ainsi au moins deux grands types de langues : les langues à cadrage verbal, comme le français et le tamoul, où la Trajectoire s'exprime dans le verbe principal, et les langues à cadrage satellitaire, tel l'anglais, où la Trajectoire se situe en dehors du verbe. De plus, la langue tamoule est agglutinante à morphologie nominale riche, à ordre SOV (cf. Naidu, 2018, pour le telugu), alors que le français et l'anglais sont des langues fusionnelles, à ordre SVO. Des recherches ont montré l'influence, parfois partielle et selon le cadrage, de la langue maternelle (L1) lors de l'acquisition des concepts spatiaux dans une autre langue.
Ainsi, en étudiant des tamoulophones-L1 acquérant le français-L3, mes questions de recherche sont les suivantes : 1) Les propriétés de la L1 facilitent-elles l'encodage des procès spatiaux en L3 ? 2) La vitesse d'acquisition est-elle accrue par le fait que la L1 et la L3 sont de même cadrage typologique ? 3) Quel est l'impact de l'anglais-L2 sur l'acquisition L3 ? Mes hypothèses principales portent sur la Trajectoire qui devrait être encodée particulièrement dans les deux langues et sur une facilité d'expression de cette composante par les apprenants.
Pour effectuer cette étude, j'ai utilisé une tâche expérimentale de production orale (Hendriks & Hickmann, 2010) montrant des agents se déplaçant selon différentes manières et trajectoires. Les participants sont tous des adultes : les deux groupes de contrôles comprennent 15 tamoulophones L1 (TAD-Pondichéry-Inde), 24 francophones L1 (FAD-Paris) et 16 tamoulophones L1, anglais L2 et apprenants du français (BTF) à trois niveaux : A1, B1, B2 selon le CEFR (Alliance Française de Madras, Inde). 
Les résultats montrent des similitudes dans l'expression de la Trajectoire qui semble dominante chez les TAD et les BTF, alors que les FAD expriment aussi la Manière et la Trajectoire. En début d'apprentissage, les apprenants se focalisent sur l'expression de la Trajectoire, ce qui est conforme à la L1 et à la L3. Mais ils expriment la Manière dans le verbe pour une trajectoire bornée, interrogeant sur un transfert de la L1 et/ou de la L2. Les types de Trajectoire semblent fortement influer sur les productions de tous les locuteurs. Mais si les apprenants semblent d'abord influencés par leur L1, dont ils retrouvent des traits dans la L3, l'anglais L2 paraît un recours, lorsque la Manière est exprimée. 
Ainsi, les apprenants montrent une souplesse dans l'usage des différents encodages, indiquant que les conceptualisations distinctes en L1 et L2 facilitent et/ou entravent l'AL3. 
Bibliographie 
Hendriks, H. & Hickmann, M. (2010). Expressing voluntary motion in a second language: English learners of French. https://doi.org/10.4324/9780203836859
Naidu (2018) Holistic spatial semantics and post-Talmian motion event typology: A case study of Thai and Telugu. https://doi.org/10.1515/cogsem-2018-2002
Talmy (2000). Toward a cognitive semantics. MIT Press.

Learners of French differ widely in their ability to correctly produce gender agreement (e.g., mon petit problème 'myMASC.SG smallMASC.SG problemMASC.SG', ma grande idée 'myFEM.SG bigFEM.SG ideaFEM.SG; e.g., Bartning, 2000; Ayoun, 2007). Many studies have focused on the role of crosslinguistic influence (CLI), including the challenges faced by learners whose L1 lacks grammatical gender (e.g., Hawkins & Franceschina, 2004). While CLI can explain between-group differences, less is understood of within-group variability. Given the role of cognitive individual differences in non-native grammatical learning (e.g., Granena, Jackson & Yilmaz, 2016; Côté, 2021), we investigate the potential contributions of working memory and metasyntactic ability to the production of French grammatical gender agreement after controlling for general morphosyntactic knowledge.
Twenty-seven Mandarin-speaking learners of French of low-intermediate (B1.1) or higher proficiency were tested on (i) their general French morphosyntactic knowledge via a closed Little Red Ridinghood passage, (ii) their working memory via a digit span task, and (iii) their metasyntactic ability using an error replication task adapted from Simard, Foucambert & Labelle (2014). French grammatical gender agreement production was elicited via a spoken picture identification task targeting determiner+noun+adjective phrases (12 masculine, 12 feminine nouns).
A two-step hierarchical multiple regression revealed a significant contribution of French grammatical knowledge (Step 1), explaining 54% of the variance in gender agreement accuracy (F(1, 25) = 29.61, p < .001). The addition of working memory and metasyntactic awareness at Step 2 made no significant new contribution. These findings are in keeping with the central role of linguistic knowledge in the acquisition of French gender (Lee, 2018) as well as studies having found a lack of effect for individual differences when task complexity is relatively low (Côté, 2021; Gaffney, 2021).
References
Ayoun, D. (2007). The second language acquisition of grammatical gender and agreement. In D. Ayoun (Ed.), French applied linguistics (pp. 130–170). John Benjamins.
Bartning, I. (2000). Gender agreement in L2 French: Pre-advanced and advanced learners. Studia Linguistica, 54(2), 225–237.
Côté, S. (2021). The independent and combined effects of cognitive and affective variables on the L2 acquisition of French grammatical gender agreement [Unpublished doctoral dissertation]. University of Toronto.
Gaffney, C. (2021). The Effects of Intelligence, Personality Traits, L1 Fluency, and L2 proficiency on L2 Spoken Fluency. [Unpublished doctoral dissertation]. University of Toronto.
Granena, G., Jackson, D. O., & Yilmaz, Y. (2016). Cognitive individual differences in second language processing and acquisition. John Benjamins.

Hawkins, R., & Franceschina. F. (2004). Explaining the acquisition and non-acquisition of determiner-noun gender concord in French and Spanish. In P. Prévost & J. Paradis (Eds.), The acquisition of French in different contexts: Focus on functional categories (pp. 175–205). John Benjamins.
Lee, A. H. (2018). The effects of different instructional and cognitive variables on the acquisition of grammatical gender by second language learners of French. [Unpublished doctoral dissertation]. McGill University.
Simard, D., Foucambert, D., & Labelle, M. (2014). Examining the contribution of metasyntactic ability to reading comprehension among native and non-native speakers of French. International Journal of Bilingualism, 18(6), 586–604.

Recent studies indicate that there is microvariation in anaphora resolution in null subject Romance languages (NSRL) (Filiaci etal2014; Torregrossa etal2020). For instance, Madeira et al (2021) found that, in Italian, Spanish and European Portuguese (EP), anaphora resolution preferences in intrasentential Main-Subordinate contexts are similar only in the case of overt pronominal subjects and when antecedents are animate: here overt subjects retrieve the object antecedent. When the object is inanimate, the overt subject recovers the subject antecedent in Italian and Spanish, whereas in EP the preference for the object antecedent is maintained. These languages also differ in the interpretation of null subjects: EP has a strong bias towards the subject antecedent, regardless of animacy, while Italian and Spanish show no clear preference. In L2 acquisition, learners' sensitivity to microvariation is still understudied, as most studies assume that NSRL behave alike in anaphora resolution. The research conducted to date, which has mostly considered contexts with animate antecedents, has generally found that learners display persistent optionality regarding overt but not null subjects (Sorace2016). 
To examine whether learners are sensitive to microvariation in anaphora resolution, this study investigates the interpretation of subject pronouns in L2 EP by Italian and Spanish native speakers. The following questions are addressed: are Italian and Spanish learners of L2 EP sensitive to L1-L2 differences regarding the role of animacy in overt subject resolution?; are they sensitive to L1-L2 differences regarding the strength of the subject antecedent bias in null subject resolution?
Participants were 25 adult EP native speakers, 69 Italian native speakers (25 upper-intermediate, 25 advanced and 19 near-native learners), and 50 Spanish native speakers (20 upper-intermediate, 20 advanced and 10 near-native learners). Two multiple-choice tasks (speeded/untimed) were used to elicit interpretation preferences in intrasentential Main – Subordinate contexts. The tasks had a 2x2 design crossing animacy of the matrix object (animate vs. inanimate) and type of embedded pronominal subject (overt vs.null). 
Regarding overt subjects, all learner groups exhibit a target-like preference for the object antecedent in the animate condition, in both tasks and across levels. However, in the inanimate condition, only the near-native groups and the advanced L1 Spanish – L2 EP group have a target-like preference for the object antecedent, which is observed in the untimed task, but not in the speeded task, where these groups exhibit optionality. Regarding null subjects, intermediate learners display optionality in both tasks. Advanced and near-native learners display target-like preference for the subject antecedent in the untimed task, but optionality in the speeded task. 
Our results indicate that learners become sensitive to microvariation in anaphora resolution at highly advanced proficiency levels (cf. untimed task). However, their performance in the speeded task remains permanently unstable in the areas where the L1 and the L2 differ, suggesting that they may be caused by processing inefficiencies and not representational deficits. Our findings challenge the ideas that only overt subjects are persistently problematic and that the L1 plays a minor role in anaphora resolution.