Anaphora resolution in L2 English: The L1 advantage 

Lydia White¹, Heather Goad¹, Guilherme Garcia², Natália Guzzo², Liz Smeets³, Jiajia Su⁴ 

McGill University¹, Université Laval², York University³, Beijing Foreign Studies University⁴ 

lydia.white@mcgill.ca 

In Italian, null pronouns prefer prominent antecedents, usually subjects. The antecedent for an overt pronoun, in contrast, is typically another NP (henceforth object) (Carminati 2002). Interpretation of pronouns in non null subject languages like English is more open to ambiguity: the antecedent can be either subject or object. Nevertheless, native speakers are reported to prefer subject antecedents (Contemori & Dussias 2020; Cunnings et al. 2017); see (1). 

            However, prosody can impact pronoun interpretation, leading to a shift in preferences. Grimshaw and Rosen (1990) report that unstressed pronouns in English prefer a prominent antecedent, while stressed pronouns prefer non-subject antecedents; compare (1) and (2). Similarly, Gargiulo et al. (2019) find that, in production of Swedish (another non null subject language), speakers signal a shift in antecedent preferences by means of stress on a pronoun or a pause between clauses.  

The role of prosody in L2ers’ interpretations of pronouns in non null subject languages has not previously been considered. We examine interpretation of English pronouns by L1 Italian speakers and English native speakers (NSs). We hypothesize that NSs and L2ers will show a preference for subject antecedents in the case of unstressed pronouns, with a shift away from this preference in cases involving stressed pronouns or a pause between clauses.  

In order to manipulate prosody, it is necessary to use auditory stimuli. With the exception of recent eye-tracking studies (Contemori & Dussias 2020; Cunnings et al. 2017), most previous research on pronoun interpretation has relied on written stimuli (e.g., Sorace & Filiaci 2006) (but see White et al. (2024) for the use of auditory stimuli with learners of L2 Italian).  

We report on an online experiment, involving 21 L2ers (L1 Italian) and 21 English NSs. Stimuli were 24 biclausal sentences (like (1)/(2)), recorded by a native speaker and presented auditorily. Stimuli manipulated presence/absence of stress and presence/absence of pause between clauses. Stressed pronouns were produced with higher pitch peaks, increased duration and greater intensity relative to unstressed pronouns; pauses averaged 400ms in length. Each sentence was preceded by a written context introducing potential referents (subject/object/ external). After listening to a sentence, participants indicated their preferred referent for the pronoun. See example test item in (3). 

The L2ers showed sensitivity to prosody, as predicted (see Figure 1): subjects were preferred for unstressed pronouns while stress led to a statistically credible decrease in subject choices, regardless of pause ([Equation] = −0.99, 95% HDI = [−1.89, −0.14] (no pause); [Equation] = −1.60, 95% HDI = [−2.58, −0.69] (pause)). NS results, however, were unexpected. Although there was an overall preference for subject antecedents over object antecedents ([Equation] =0.56, 95% HDI = [−0.08, 1.18]), neither stress nor pause led to a shift in antecedent choices. 

The L2ers’ sensitivity to stress may be indirectly attributable to the L1: the two different pronoun types in Italian (null versus overt) differ in their antecedent preferences; this contrast may make L2ers extra-sensitive to differences in the L2 (unstressed versus stressed). The lack of sensitivity shown by the NS group, on the other hand, suggests that prosody alone may be insufficient to lead to changes in their antecedent choices and that richer contextual information may be required as well.

Figure 1. Antecedent choices in English by L2ers and native speakers (in %) 

(Note: external referents were rarely chosen, ~5% of all responses.) 

References:  

1. Alchemer. https://www.alchemer.com/ 
2. Carminati, M. 2002. The processing of Italian subject pronouns. PhD dissertation, UMass Amherst.  
3. Contemori, C. & P. Dussias. 2020. The processing of subject pronouns in highly proficient L2 speakers of English. Glossa 5(1): 38. 
4. Cunnings, I., G. Fotiadu & I. Tsimpli. 2017. Anaphora resolution and reanalysis during L2 sentence processing. Studies in Second Language Acquisition 39: 621-652. 
5. Gargiulo, C., M. Tronnier & P. Bernardini. (2019). The role of prosody in overt pronoun resolution in a null subject language and in a non-null subject language: A production study. Glossa 4(1): 135. 
6. Grimshaw, J. & S. Rosen. 1990. Knowledge and obedience: the developmental status of the Binding Theory. Linguistic Inquiry 21: 187-222. 
7. Sorace, A. & F. Filiaci. 2006. Anaphora resolution in near-native speakers of Italian. Second Language Research 22: 339-368. 
8. White, L., H. Goad, G. Garcia, N. Guzzo, L. Smeets & J. Su. 2024. Pronoun interpretation in Italian: exploring the effects of prosody. Linguistic Approaches to Bilingualism (online first). 

Are textually enhanced subtitles overrated?

Hanneke Loerts, Vincent Fan, Laura Fiche, Anastasia Pattemore

Center for Language and Cognition Groningen, University of Groningen

h.loerts@rug.nl

Subtitled audiovisual input that exposes learners to images combined with sound and on-screen text is increasingly being used for foreign language learning. For learners to notice the target language and successfully uptake it, textually enhanced (TE) subtitles have been proposed as a powerful tool, as they cause increased reading time and longer fixations (as measured by eye-tracking, e.g. Puimège et al., 2023). However, the studies that found a positive effect of TE subtitles on learning implemented short video treatments and/or one-off data collections. Conversely, studies that incorporated longitudinal viewing treatments (Pattemore & Muñoz, 2022) found no evidence of TE subtitle benefits, presumably due to the loss of attention towards TE subtitles over time, as was observed in Indrarathne et al. (2018). To date, no studies have examined processing of TE subtitles through eye-tracking over time. This study aims to fill this gap by implementing a prolonged viewing intervention where participants watch the entire season of a TV show with TE subtitles. The participants’ eye-movements were recorded during the first and the last episode to measure a potential shift in attention to TE target words.  

            Twenty-five international students with little to no knowledge of Dutch were asked to watch the first season (13 episodes of 21 minutes each) of the TV series ‘The Good Place’ (Schur, 2016) in the original version with L2 English audio and L3 Dutch subtitles. The participants completed vocabulary size tests as a proficiency proxy for English and Dutch. Before the first episode and after the final (13^th) episode, participants were asked to perform a vocabulary test containing (1) 16 regularly occurring target words that were enhanced in the subtitles throughout the season; (2) 16 unenhanced target words that appeared equally often in the subtitles as the enhanced ones; and (3) 16 filler words that did not appear in the subtitles at all. The first and last episodes were viewed with an eye-tracker (EyeLink Portable Duo), while the episodes in between were viewed at home over two weeks. Using this set-up, we aimed to determine the extent to which TE increases fixations towards words, as well as the acquisition of these words. Over time, we also aimed to observe whether enhanced words receive the same amount of attention or whether the words are increasingly being skipped. 

            Preliminary analyses of the eye-tracking data suggest that TE words receive more attention and are less often skipped than unenhanced words. Skipping rate decreased overall when comparing the first and last viewing session, but the amount of time participants fixated on target words only decreased for the unenhanced target words and not for TE words. The L3 Dutch vocabulary tests additionally revealed significantly larger vocabulary gains for TE words as compared to both filler and unenhanced words. These results extend research into the benefits of subtitled television, and TE subtitles in particular, to vocabulary acquisition of L3 Dutch through L2 English and stress the importance of examining learning and viewing behaviour over a prolonged period of time. 

References: 

1. Indrarathne, B., Ratajczak, M., Kormos, J. (2018). Modelling Changes in the Cognitive Processing of Grammar in Implicit and Explicit Learning Conditions: Insights From an Eye-Tracking Study. Language Learning, 68(3), 669–708. https://doi.org/10.1111/lang.12290  
2. Pattemore, A., & Muñoz, C. (2022). Captions and learnability factors in learning grammar from audio-visual input. The JALT CALL Journal, 18(1), 83–109. https://doi.org/10.29140/jaltcall.v18n1.564  
3. Puimège, E., Montero Perez, M., & Peters, E. (2023). Promoting L2 acquisition of multiword units through textually enhanced audiovisual input: An eye-tracking study. Second Language Research, 39(2), 471-492. https://doi.org/10.1177/02676583211049741   
4. Schur, M. (Creator). (2016). The Good Place. [TV series]. United States: Fremulon. 

Bilingualism and Mental health in Middle Childhood – A longitudinal investigation

Paulina Salgado-Garcia, Rory T. Devine, Andrea Krott

University of Birmingham

pds208@student.bham.ac.uk

Extensive evidence suggests that poor mental health during childhood predicts academic difficulties, later mental health problems, and work instability (Constantini et al., 2023; Bornstein et al., 2010). It remains unclear if bilingual and monolingual children exhibit similar mental health trajectories. Some findings indicate that speaking the parents’ language reduces mental health risks for bilingual children (Rumbaut, 1994), but other studies suggest that bilingual children have higher levels of anxiety and depression (Tilley et al., 2021). This project aimed to investigate whether bilingualism influenced the trajectories of children’s internalising and externalising problems regardless of language proficiency, family background and school characteristics. Utilizing a large dataset, the Millenium Cohort Study (MCS), this study provides robust insights by adjusting for a numerous covariates, highlighting the specific contributions of bilingualism to mental health.

            Methods.The MCS is an on-going longitudinal study conducted across England, Scotland, Wales, and Northern Ireland. The present study involved the subset of 15,435 children (49% girls), including 2,146 (49.4% girls) bilinguals, who were assessed at ages three, five, seven and eleven. Mental health was measured using the Strengths and Difficulties Questionnaire, capturing internalising (emotional and peer problems) and externalising problems (conduct problems and hyperactivity symptoms). Results were analysed by means of latent growth curve models. The study was pre-registered on the Open Science Framework at https://osf.io/ghsw6/.

            Results.We first analysed the data without controlling for covariates (sex, non-verbal IQ, parents’ mental health, ethnicity, socio-economic status, language proficiency, bilingual peers, and teacher in class). For internalising problems, Unadjusted models indicated there was a significant effect of bilingualism on the intercept (.676, p < .001), indicating that bilingual children had higher levels of internalizing problems than monolingual children at age 3. There was also a significant effect of bilingualism on the slope for internalising, indicating that bilingual children showed greater declines in internalising problems compared to monolinguals. But they still scored higher than monolinguals at ages 5 and 11. For externalising problems, there was no effect of bilingualism on the intercept (.216, p.09), indicating that bilinguals had similar levels of externalizing problems than monolinguals.

Importantly, when models were adjusted to account for the influence of covariates, bilinguals exhibited lower levels of both internalising (-.309, p < .05) and externalising problems (-.671, p < .05) at all ages. Socioeconomic status moderated the effect on externalising problems, with benefits observed only in low (unstandardized estimate -1.36, p < .00) and mid-level

(-.54, p < .00) socioeconomic groups, but not in the high socioeconomic group (.11, p .61).

            Conclusion. If not accounting for confounding variables, such as ethnicity, socio-economic status or language proficiency, bilingualism is associated with higher mental health issues in terms of internalising problems. But after controlling for covariates, bilingualism might play a protective role in children’s mental health. For internalising problems, this is the case for children in all socio-economic groups. For externalising problems bilingualism is beneficial for lower socioeconomic groups. These findings have the potential to inform public policies and benefit bilingual communities.

Figure 1 Bilingualism shows benefits for both, internalising and externalising problems, after controlling for covariates.

References:

1. Bornstein, M. H., Hahn, C. S., & Haynes, O. M. (2010). Social competence, externalizing, and internalizing behavioral adjustment from early childhood through early adolescence: Developmental cascades. Development and psychopathology, 22(4), 717-735.
2. Costantini, I., Sallis, H., Tilling, K., Major‐Smith, D., Pearson, R. M., & Kounali, D. Z. (2023). Childhood trajectories of internalising and externalising problems associated with a polygenic risk score for neuroticism in a UK birth cohort study. JCPP advances, 3(1), e12141.McNeish, D. (2020). Relaxing the proportionality assumption in latent basis models for nonlinear growth. Structural Equation Modeling: A Multidisciplinary Journal, 27(5), 817-824. https://doi.org/10.1080/10705511.2019.1696201
3. Rumbaut, R. G. (1994). The crucible within: Ethnic identity, self-esteem, and segmented assimilation among children of immigrants. International Migration Review, 28(4), 748-794.
4. Tilley, J. L., Huey Jr, S. J., Farver, J. M., Lai, M. H., & Wang, C. X. (2021). The immigrant paradox in the problem behaviors of youth in the United States: A meta‐analysis. Child Development, 92(2), 502-516.
5. University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2023). Millennium Cohort Study. [data series]. 15th Release. UK Data Service. SN: 2000031, DOI: http://doi.org/10.5255/UKDA-Series-2000031​

Comparing bilingual and monolingual children’s interpretation of novel-verb sentences: is their development related to vocabulary or executive function?

Noorin Rodenhurst (University of Warwick), Katherine Messenger (Lancaster University)

n.rodenhurst@warwick.ac.uk

k.messenger@lancaster.ac.uk

Previous research into children’s comprehension of syntactic structures has investigated awareness of transitive and intransitive structures amongst monolingual children (Noble et al., 2011). This research has found that children as young as two associate novel-verb transitive sentences (e.g. the rabbit is meeking the duck) with a causal scene and do not associate intransitive sentences (e.g. the rabbit is meeking)with a causal action; though they do not necessarily choose a non-causal scene as a match for intransitive sentences either (Arunachalam et al., 2016). These findings suggest that monolingual children acquire verb-general knowledge of syntax early on, but less attention has been paid to bilingual children. Bilingual children may acquire verb-knowledge more slowly than monolingual children due to decreased input for each language, but research into bilingual children’s understanding of the same sentence structures is lacking. Moreover, their bilingual experience may affect their language learning in other ways: because bilingual children need to attend to different languages in different contexts, it is argued they have a “bilingual advantage” in executive function (Bialystok, 2015). One possibility is that this could support children’s interpretation of sentences.

            This study compared 46 bilingual children, aged 3 to 5 years, who spoke English and any other language, with 64 monolingual, English-speaking 3- to 5-year-olds, in their ability to associate transitive sentences with a causal scene, and intransitive sentences with a non-causal scene. Participants saw pairs of pictures depicting causal events and non-causal events and heard transitive or intransitive descriptions (see Figure 1). They were instructed to point to the image they thought matched the sentence. The study was conducted online using Pavlovia; parents were asked to complete the tasks with their child and record the responses they gave. They completed eight trials per structure, increasing the number of trials (two) typically used in such studies for greater reliability. Children’s executive function skills were tested via the Bivalent Shape Task (Mueller & Esposito, 2014) and vocabulary was measured with the British English Crosslinguistic Lexical Tasks, verb comprehension test (Haman et al., 2012).

            Logistic mixed effects model found that children pointed more at the causal scene on transitive than intransitive trials in each group, but overall, monolingual children were more likely to do so than bilingual children. In both groups, children matched causal scenes to transitive sentences greater than chance; older children’s points to the scene were at chance on intransitive trials but the youngest children pointed above chance to the causal scene on intransitive trials. Executive function performance did not differ between bilingual and monolingual children and was also found not to relate to performance in the sentence task, however those with higher vocabulary more frequently pointed to the causal image for transitive than for intransitive sentences.

            Therefore, the number of languages spoken by a child does not affect their ability to comprehend transitive and intransitive sentences: bilingual preschoolers showed the same verb-general knowledge of syntax as monolingual preschoolers. Verb knowledge was related to this performance but executive function, as measured by the BST, was not.

References:

1. Arunachalam, S., Syrett, K., & Chen, Y. (2016). Lexical disambiguation in verb learning: Evidence from the conjoined-subject intransitive frame in English and Mandarin Chinese. Frontiers in Psychology, 7, 138.
2. Bialystok, E. (2015). Bilingualism and the development of executive function: The role of attention. Child Development Perspectives, 9(2), 117-121.
3. Haman, E., Łuniewska, M., Polišenská, K., Mieszkowska, K., & Chiat, S. (2012). Cross-linguistic Lexical Tasks: British English version (CLT-EN) [Unpublished material]. University of Warsaw
4. Mueller, S. T., & Esposito, A. G. (2014). Computerized testing software for assessing interference suppression in children and adults: The Bivalent Shape Task (BST). Journal of Open Research Software, 2(1), e3.
5. Noble, C. H., Rowland, C. F., & Pine, J. M. (2011). Comprehension of argument structure and semantic roles: Evidence from English‐learning children and the forced‐choice pointing paradigm. Cognitive Science, 35(5), 963-982.

Conceptual Restructuring in L2 Acquisition: the Influence of Reading Direction on Actants’ Selection in Causal Events

Mireille Copin¹, Inès Saddour ¹, Cyrille Granget¹

¹Laboratory of NeuroPsychoLinguistics, University of Toulouse

mireille.copin@univ-tlse2.fr  

cyrille.granget@univ-tlse2.fr

 ines.saddour@univ-tlse2.fr

Recent studies on second language acquisition have used on-line measures to investigate the influence of L1 structures on cognition, but research on conceptual restructuring through L2 acquisition have yielded mixed results (Flecken et al. 2015, Wang & Wei 2021). However, research has mostly focused on Indo-European or East Asian languages, and expression of aspect or motion. We propose to study reference to entities of causal events in French and Syrian Arabic.

            Previous research have investigated the “Agent-first” preference, postulating a universal bias to choose Agents as subjects and focus attention on them when describing causal events – even if this can be modulated by syntactic priming and reading direction (Sauppe & Flecken 2021, Esaulova et al. 2020). Moreover, previous works have also shown that preferences in visual event representations are derived from reading direction and that first fixated characters are usually chosen as subjects (Gleitman et al 2007).

            Our study aims to investigate the influence of reading direction in the perception and representation of causal events’ actants by L1 and L2 French speakers. We wish to study the “Agent first” preference in French L1 and Syrian Arabic L1, and whether it can be modulated by the Agent’s position on the screen. It is part of a longitudinal project on conceptual restructuring in Syrian Arabic learners of French, to see if exposure to a different reading direction can affect dynamic scene perception.

            We collected data from 29 Syrian Arabic speakers and 22 French L1 speakers using verbalization tasks in French, and then again for the learners group in Syrian Arabic L1 after distracting tasks. Eye-tracking data were collected for both tasks. The same group of participants did the task in French L2 and Syrian Arabic L1. In each experiment, participants had to watch and narrate short videos, in which one character (Agent) hits the other (Patient) with an object. We counterbalanced the Agent’s left-right position across the videos and presented mirror versions of the stimuli to half the participants. Our two areas of interest (AOI) were the two protagonists’ faces and shoulders. In the verbalizations, we analyzed the first introduced character in the retelling and the subject of the collision predicate.

            Analyses show mixed results. While the analysis of collision verbs’ subjects confirms the linguistic Agent-first preference for both groups, independently of the Agent’s position, total dwell times reveal that more time is spent looking at Patients. Moreover, discourse analyses suggest no Agent-first preference, as order of introduction “Patient > Agent” is more frequent in French L1. However, this preferential order is modulated by reading direction: Agents on the left are more likely to be mentioned first in French L1. Conversely, native Arabic speakers introduce right-positioned Agents first more often in Syrian Arabic L1 but not French L2, as there is no preferential order of mention. These results suggest some conceptual restructuring that lessens the influence of a right-to-left reading bias.

References:

1. Esaulova, Y., Dolscheid, S., Reuters, S., & Penke, M. (2021). The alignment of agent-first preferences with visual event representations : Contrasting German and Arabic. Journal of Psycholinguistic Research, 50, 843‑861.
2. Flecken, M., Carroll, M., Weimar, K., & Von Stutterheim, C. (2015). Driving Along the Road or Heading for the Village? Conceptual Differences Underlying Motion Event Encoding in French, German, and French–German L2 Users. The Modern Language Journal, 99(S1), 100‑122. https://doi.org/10.1111/j.1540-4781.2015.12181.x
3. Gleitman, L. R., January, D., Nappa, R., & Trueswell, J. C. (2007). On the give and take between event apprehension and utterance formulation. Journal of memory and language, 57(4), 544‑569.
4. Sauppe, S., & Flecken, M. (2021). Speaking for seeing : Sentence structure guides visual event apprehension. Cognition, 206, 104516. https://doi.org/10.1016/j.cognition.2020.104516
5. Wang, Y., & Wei, L. (2021). Cognitive restructuring in the multilingual mind: Language-specific effects on processing efficiency of caused motion events in Cantonese–English–Japanese speakers. Bilingualism: Language and Cognition, 24(4), 730-745

Crosslinguistic priming of syntactic and thematic roles: Evidence from Polish-English bilingual children

Katherine Messenger¹, Marta Wesierska¹, Vanessa Cieplinska², Ludovica Serratrice²

¹University of Warwick, ²University of Reading

K.Messenger@lancaster.ac.uk

A key issue in the study of bilingual language development concerns the nature of syntactic representations in children’s two languages – specifically, the extent to which they are shared cross-linguistically. Evidence from structural priming in adults suggests that proficient second language learners can rely on shared syntactic structures [1] and that thematic role ordering can be primed across languages [2, 3]. The evidence for shared syntax in bilingual children is limited to a small number of studies showing significant syntactic priming effects in one [4] or both directions [5]. Other studies have found word order, but not syntactic, priming in the same group of children [6]. Whether bilingual children are sensitive to the priming of thematic roles instead of or alongside the priming of syntactic structure is not yet known.  

            In this study we investigated the priming of syntactic structure and thematic role order in a group of Polish-English bilingual children. In Polish, like in English, transitive events can be linguistically encoded by active (A) and passive (B) constructions. Unlike in English, Polish has a third construction (OVS) with the same thematic role order as the passive – the object/theme appears before the subject/agent – but with the syntactic structure of an active (C). This allowed us to test whether children could be primed to produce a passive in English by a passive in Polish (syntactic priming), and/or by an OVS prime in Polish (thematic role order priming). Equally, whether a passive prime in English would exclusively prime a passive in Polish (syntactic priming), and/or whether it would also prime an OVS construction (thematic role order priming).  

            In a bidirectional priming experiment, we tested 48 Polish-English bilinguals (26 girls; Mage: 9;5, range: 7;11–11;4). Children completed two sessions of a Snap! game with 48 prime-target items (Figure 1) a week apart (order of prime language counterbalanced across children and sessions). In a pre-priming baseline task, children produced no passives in English or Polish; after priming they produced more passives in English than in Polish (Table 1); they produced no OVS constructions in Polish. 

            After hearing Polish primes, children produced more passives in English than in the baseline phase (Zs>3, ps≤.001) and after hearing passive and OVS primes than after hearing active primes (Z=2.75, p=.006); the effect of passive and OVS primes did not differ (Z=0.88, p=.39). After hearing English primes, children produced more passives in Polish than in the baseline phase (Zs=1.9, ps≤.05) but there was no difference in the effect of active compared to passive primes (i.e. no effect of priming; Z=0.06, p=.95).   

            We observed small (~2%) crosslinguistic priming effects, both syntactic priming and thematic role order priming, from Polish to English. However, priming was not bidirectional: English passive primes did not elicit more Polish passive responses than active primes; this may reflect greater competence in the societal language, English. Using another sample of the same population we are testing whether priming for these structures is stronger within-languages than across. We are also examining what crosslinguistic priming effects adult bilinguals show. 

Examples of prime sentences: 

1. Policjant czyści pistolet. (Active) 

            “The policeman [agent] cleans the gun [theme].” 

• Pistolet jest czyszczony przez policjanta. (Passive) 

            “The gun [theme] is cleaned by the policeman [agent].” 

• Pistolet czyści policjant. (OVS) 

            “The policeman [agent] cleans the gun [theme].” 

Figure 1. Snap game pictures: Polish prime, English target  

Prime condition	English passives	Polish passives
Baseline	0	0
Active	18 (2.2%)	8 (1.0%)
Passive	36 (4.3%)	8 (1.0%)
OVS	30 (3.6%)	–

 Table 1. Total number (%) of Passive target responses by target language and priming condition. 

References: 

1. Van Gompel, R. P., & Arai, M. (2018). Structural priming in bilinguals. Bilingualism: Language and Cognition, 21(3), 448-455. 
2. Fleischer, Z., Pickering, M. J., & McLean, J. F. (2012). Shared information structure: Evidence from cross-linguistic priming. Bilingualism: Language and Cognition, 15(3), 568-579. 
3. Vernice, M., Pickering, M. J., & Hartsuiker, R. J. (2012). Thematic emphasis in language production. Language and Cognitive Processes, 27(5), 631-664. 
4. Wolleb, A., Sorace, A., & Westergaard, M. (2018). Exploring the role of cognitive control in syntactic processing: Evidence from cross-language priming in bilingual children. Linguistic Approaches to Bilingualism, 8(5), 606-636. 
5. Gámez, P. B., & Vasilyeva, M. (2020). Shared syntactic representations in balanced bilinguals: Cross-linguistic priming with and without verb overlap. Language Learning and Development, 16(1), 89-106. 
6. Wesierska, M., Serratrice, L., Cieplinska, V., & Messenger, K. (in press). Investigating crosslinguistic representations in Polish-English bilingual children: Evidence from structural priming. Bilingualism: Language and Cognition.  

Cross-linguistic ungrammatical priming in Canadian French-English bilinguals:

Evidence from a self-paced reading task

Foteini Karkaletsou^1*, Gunnar Jacob², Shanley E.M. Allen¹

¹University of Kaiserslautern-Landau, ²University of Mannheim

*karkalet@rptu.de

Language contact in bilingual contexts often leads to the emergence of ungrammatical structures and hence to language change[2]. However, so far relatively little is known about the psycholinguistic mechanisms driving this process. Relevant work has suggested that cross-linguistic structural priming might be a potential driver of contact-induced language change[6]. To date, most priming research has examined structures that are grammatical in both languages[7], with only some studies testing structures that are ungrammatical in the target language when copying the syntax of the source language[2,3]. While these mostly focus on production, very little is known about how bilinguals process such structures after encountering grammatical equivalents in the other language[4]. 

The present study investigates priming in the processing of French ungrammatical ditransitive sentences that copy the syntax of English grammatical ditransitive sentences. While English allows both Prepositional Object – PO (1a) and Double Object – DO structures (1b), French only allows POs with canonical word order (1c) or with scrambled word order (1d). We study this asymmetry in the contact context of Canada where evidence on production and/or comprehension of DOs in French (1e), i.e., upon omission of the dative marker à (to), is almost non-existent. Importantly, a recent study has shown that French-English bilinguals in Canada tend to rate them as relatively acceptable, but still significantly lower than POs[5].  

On this basis, a cross-linguistic priming task with self-paced reading (SPR) is implemented on Gorilla[1] to examine how Canadian French-English bilinguals are influenced by English ditransitive prime sentences (2a-c) in processing ungrammatical French DO target sentences (2d). English primes are either DOs (2a), i.e., the grammatical equivalent structure to the target, or POs with canonical word order (2b), i.e., the grammatical alternative. Unlike previous studies, ungrammatical English primes are also included, namely scrambled POs (2c) to test whether word order overlap with the French DO target (indirect object before direct object) facilitates processing. English primes are presented as whole sentences and French target are presented region-by-region in a moving window display. Reading times (RTs) in milliseconds will be measured on the critical region of the target (see Table 2) and compared across the different types of primes. Statistical analysis will be carried out via linear mixed effects models. 

If cross-linguistic priming is indeed a psycholinguistic mechanism behind language change, then we expect ungrammatical French DO targets to be read faster when primed by a grammatical English DO sentence than when primed by a grammatical English PO sentence. Moreover, we expect that ungrammatical English PO scrambled primes will facilitate processing on the target more than grammatical English PO canonical primes due to word order overlap.  

Data collection is ongoing. Data will be available to present by the time of the conference.  

Structure	English	French
Prepositional Object – PO	(1a) The man gives a book  to the woman.	(1c) L’homme donne un livre  à la femme.    (1d) L’homme donne à la femme un livre.
Double Object – DO	(1b) The man gives the woman a book.	(1e) * L’homme donne   la femme un livre.

Table 1. Cross-linguistic differences between English and French in the formation of ditransitives with full noun phrases (NPs) in the argument structure. 

Sentence	Condition
Prime (English)	(2a) The man gives the woman a book   for her birthday.    (DO)	(2b) The man gives   a book to the woman for her birthday.    (PO canonical)	* (2c) The man gives   to the woman a book   for her birthday.    (PO scrambled)
Target (French)	* (2d) Le garçon / donne / le fermier un chapeau / dans la rue.               The boy       gives       the farmer  a hat             in the street.    (DO)

Table 2. Set of conditions for primes and targets. Different regions are separated by / on the target (2d), where the critical region is marked in bold.  

References: 

1. Anwyl-Irvine, A. L., Massonnié, J., Flitton, A., Kirkham, N., & Evershed, J. K. (2020). Gorilla in our midst: An online behavioral experiment builder. Behavior research methods, 52, 388-407. 
2. Fernández, E. M., De Souza, R. A., & Carando, A. (2017). Bilingual innovations: Experimental evidence offers clues regarding the psycholinguistics of language change. Bilingualism: Language and Cognition, 20(2), 251-268. 
3. Hopp, H., & Jackson, C. N. (2023). Asymmetrical effects of cross-linguistic structural priming on cross-linguistic influence in L2 learners. Applied Psycholinguistics, 44(2), 205-236. 
4. Ilen, H., Jacob, G., & Engemann, H. (2023, October 12). Cross-Linguistic Ungrammatical Priming in Turkish-German Bilinguals. Presentation at UniKon International PhD Conference on Multilingual Acquisition. Constance, Germany.  
5. Karkaletsou, F., Kholodova, A., & Allen, S.E.M. (2023, August 14). Bilingual adaptation to structural innovations: Evidence from Canadian French. Poster at SSoL23 (Summer School of Linguistics). Ceske Budejovice, Czech Republic. 
6. Kootstra, G. J., & Şahin, H. (2018). Crosslinguistic structural priming as a mechanism of contact-induced language change: Evidence from Papiamento-Dutch bilinguals in Aruba and the Netherlands. Language, 94(4), 902-930. 
7. Van Gompel, R. P., & Arai, M. (2018). Structural priming in bilinguals. Bilingualism: Language and Cognition, 21(3), 448-455. 

Does Explicit Instruction Lead to Implicit Processing of L2 English Generic NPs?

Neal Snape^1,2, Menghan Wang³, Helen Zhao³

¹Gunma Prefectural Women’s University, ²Chuo University, ³University of Melbourne

nealsnape@fic.gpwu.ac.jp

menghan.wang@student.unimelb.edu.au

helen.zhao@unimelb.edu.au

Our study provided instruction on English generic NPs to upper-intermediate and advanced L1-Japanese L2-English learners. The instruction was designed and conducted as an experimentalized online tutor (Zhao & MacWhinney, 2018) that implements random assignments, practice tasks, and immediate feedback. Participants were randomly divided into a treatment group (n = 20) and a control group (n = 17). A pre-test, training sessions and two post-tests were administered that measured knowledge of English generic NPs based on Krifka et al.’s (1995) classification of kind NP-level and sentence-level generic sentences. Only the definite singular and bare plural can be used for NP-level generics such as a natural kind with a kind predicate like the dodo bird is extinct / dodo birds are extinct. The indefinite singular and bare plural are acceptable for sentence-level generics, e.g., an orange has lots of vitamin C / oranges have lots of vitamin C. The treatment group was trained for generic usage through an online acceptability judgement task with many practice trials and metalinguistic feedback (in Japanese). The control group received online English preposition training. The study was programmed in PsyToolKit by one of the authors.

            A follow-up session was created using a timed grammaticality judgement task (GJT) one year after instruction to investigate whether the treatment group (n = 13 of the original 20 participants) has internalized explicit knowledge. The implicit GJT aims to discover whether the learners process the generic NPs in contexts and correctly select the appropriate article within five seconds since there was no way of measuring this during the instruction period. A new control group (n = 20) was included where they received the same GJT to complete. Statistical analyses revealed that the treatment group were faster than the control group at rejecting ungrammatical sentences NP-level natural kind generics (treatment group mean RTs = 1644ms, control group mean RTs = 2503ms, p = .05, d = 4.43). For comprehension question accuracy responses, all three subtypes at the NP-level were accepted and rejected more often by the treatment group.

            We contend that not all instruction leads to a successful outcome regarding explicit knowledge becoming implicit (see Umeda et al. 2019 as an example). However, our findings are promising for L2 learners and teachers, confirming that learners can pick up even the more difficult and less frequent article cues from computer-based instruction, given the right type and amount of instruction.

References:

1. Krifka, M. Pelletier, F., Carlson, G., ter Meulen, A., Link, G. & Chierchia, G. (1995). Genericity: an introduction. In G. Carlson & F. Pelletier (Eds.), The generic book (pp.1-125). Chicago: University of Chicago Press.
2. Umeda, M., Snape, N, Yusa, N & Wiltshier, J. (2019). The long-term effect of explicit instruction on learners’ knowledge on English articles. Language Teaching Research, Special Issue: Grammatical Meaning and the Second Language Classroom, 23, 179-199.
3. Zhao, H., & MacWhinney, B. (2018). The instructed learning of form–function mappings in the English article system. The Modern Language Journal, 102, 99-119.

Does L1 attrition affect predictive processing? Evidence from Japanese expats in the U.S.

Theres Grüter, Sachiko Roos

University of Hawai‘i at Mānoa

theres@hawaii.edu

Bilinguals’ engagement in predictive processing in their weaker language is often reduced and/or delayed when compared to their stronger language and/or to functionally monolingual comparison groups (Kaan & Grüter, 2021). The extent of these differences appears to be modulated by the nature of the predictive cue, with semantic cues supporting prediction more robustly across populations than morphosyntactic ones (Hopp, 2015), and early bilinguals/heritage speakers more likely to pattern with monolinguals than post-childhood L2 learners (Fuchs, 2022). No study to date has investigated whether long-term immersive exposure to an L2 affects predictive processing in the L1. Sentence processing has shown to be remarkably robust to L1 attrition, yet some studies have reported subtle effects of L2-to-L1 influence (Dussias & Sagarra, 2007; Kasparian & Steinhauer, 2017). Here we investigate whether potential effects of L1 attrition extend to predictive processing.

            Methods. 56 L1 Japanese speakers raised in Japan – 26 long-term U.S. residents (>2yrs, median: 10yrs; “Expats”) and 30 shorter-term visitors (<2yrs, median: 2.5 mths; “L1ers”) – participated in a visual-world eye-tracking study in Japanese testing use of a semantic and a morphosyntactic predictive cue not present in English: prenominal classifiers (Fig.1; replicating Mitsugi, 2018) and case-marking (Fig.2; replicating Kamide et al., 2003).

            Results. For both cues, we used mixed logit models to calculate the likelihood of a look to the target prior to its being named in the condition where it was predictable (different-classifier, ditransitive; respectively) vs the condition where it was not (same-classifier, accusative).

Classifiers: Both groups were more likely to look at the target in the different- vs same-classifier condition (Fig.3). The effect of Condition was significant in both groups (L1: b=.86, p<.001; Expats: b=1.17, p<.001), and did not interact with Group (b=0.3, p=.38). This replicates Mitsugi’s (2018) findings from L1 (and L2) Japanese speakers, indicating semantically informative classifiers serve as robust predictive cues for Japanese listeners, with no apparent attenuation by long-term residence abroad.

Case-marking:  Both groups were more likely to look at a potential theme (N2 in Fig.4) following dative- vs accusative-marked objects (L1: b=.71, p=.04; Expats: b=.73, p=.04; no interaction with group), indicating anticipation of a ditransitive construction based on a dative-marked object. This replicates Kamide et al.’s (2003) and Mitsugi and MacWhinney’s (2016) findings from L1 (but not L2) speakers. Notably, the effect is subtle in both groups, and visually appears attenuated in the Expat compared to the L1 group (Fig.4). Further exploratory analyses showed no modulation by length-of-residence-abroad as a continuous predictor, yet when expats were median-split (>10yrs abroad, N=14, vs 2-9yrs abroad, N=12), the effect remained robust only in the subgroup with less time abroad.

            Taken together, these findings demonstrate that use of both semantic and morphosyntactic cues for predictive processing in L1 remains robust even after long-term residence in a non-L1 environment, with only tentative indication of potential attenuation of predictive use of the morphosyntactic cue among those with the longest time abroad. Further analyses are under way to examine this trend, as well as the roles of proficiency and frequency of language use.

Selected references:

1. Kamide, Altmann, & Haywood (2003). The time-course of prediction in incremental sentence processing: Evidence from anticipatory eye movements. JML.
2. Mitsugi (2018). Generating predictions based on semantic categories in a second language: A case of numeral classifiers in Japanese . IRAL.
3. Mitsugi & MacWhinney (2016). The use of case marking for predictive processing in second language Japanese. BLC.

Does reduced engagement in prediction lead to fewer false memories of predictable words in L2?

Katja Haeuser¹, Theres Grüter²

¹Saarland University, ²University of Hawai‘i at Mānoa

khaeuser@coli.uni-saarland.de

Prediction, i.e., the pre-activation of language, constitutes a hallmark of language comprehension in native (L1) speakers [1, 2]. Second language (L2) learners on the other hand, are less likely to engage in prediction [3, 4]. Prediction may also have effects on memory: When sentences disconfirm predictions (e.g., “Be careful, the top of the stove is very dirty.”), predictable words (e.g., “hot”) elicit false remembering in tests of recognition memory [5], an effect that remains to be substantiated for L2 speakers [6]. However, it is currently unclear if such false memories are a consequence of prediction during sentence processing, or if they emerge during recognition when participants use the semantic features of the predictable words to reinstate the previously encoded sentences. Here we tested these two accounts with L1 and L2 speakers of German: If false remembering is a consequence of prediction, we expect to find a reduced false memory effect in L2 speakers, in light of their limited engagement in prediction. If, however, false remembering is a consequence of semantic association, L1 and L2 speakers should show similar rates of false remembering.

            We used a self-paced sentence reading (SPR, non-cumulative) and subsequent recognition memory test. During SPR, German L1 and L2 speakers (see Table 1) read 32 constraining German sentences (e.g., “At the hospital the nurse patches up quickly the …”) which continued either in a predictable (e.g., “… wound in the room next door”) or an unpredictable fashion (e.g., “pants in the room next door”). In the recognition memory task, participants saw German nouns appear on a screen and were asked to indicate whether the words were “old” (previously seen) or “new”. There were four conditions: Previously seen predictable and unpredictable nouns (e.g., “wound”, “pants”), genuinely new nouns (e.g., “paper”) and lures (i.e., previously predictable but not encountered nouns; e.g., “wound” after reading “pants”). Participants additionally indicated their confidence about their recognition judgments; here, we report data from the high-confidence judgments [5, 7].

            During SPR, unpredictable nouns elicited integration difficulty in both L1 and L2 speakers, with no differences in the size of the predictability effect between the two groups (Figure 1A). Results for the recognition memory task (Figure 1B) showed a false memory effect in both groups (i.e., more false alarms to lures vs new nouns), and, in line with the “prediction” view of false remembering, this effect was reduced in L2 compared to L1 speakers. However, more in line with the “semantic association” view, absolute rates of false remembering for lures did not differ between L1 and L2 speakers.

            These findings could indicate that two distinct processes elicit false remembering: Predictions that become generated during initial sentence reading, and backward semantic associations that participants use in the recognition memory test to “jog” their memory of the previously read sentences. Even though L2 readers likely engage less in predictive processing during sentence encoding than L1 readers, they likely employ the same associative search strategies to reinstate their memory in the recognition task.

Figure 1A: Partial effects plot showing LMER-model estimated effects of condition (i.e., predictable and unpredictable, sum-coded) and group (i.e., L1 and L2, sum coded) on log-transformed reading times of words in the critical region. We ran separate models per single word; every model specified control predictors for trial number, word length, word frequency, RTs of the preceding word, and word position in the sentence. B: Partial effects plot showing GLMER-model estimated effects of condition (three contrasts: old vs new, lure vs new and old-unpredictable vs old-predictable, sum-coded) and group (sum-coded) on trial-by-trial “old” judgments in the recognition task. Lure: Words previously predicted but not seen. New: Genuinely new words, not previously seen during SPR. Old-Predictable/Old-Unpredictable: Old words previously seen during SPR. For old words, the bars show hits; for new and lure words, the bars show false alarms.

References:

1. Kuperberg, G. R., & Jaeger, T. F. (2016). What do we mean by prediction in language comprehension?. Language, Cognition and Neuroscience, 31(1), 32-59.
2. Huettig, F. (2015). Four central questions about prediction in language processing. Brain Research, 1626, 118-135.
3. Martin, C. D., Thierry, G., Kuipers, J. R., Boutonnet, B., Foucart, A., & Costa, A. (2013). Bilinguals reading in their second language do not predict upcoming words as native readers do. Journal of Memory and Language, 69(4), 574-588.
4. Schlenter, J. (2023). Prediction in bilingual sentence processing: How prediction differs in a later learned language from a first language. Bilingualism: Language and Cognition, 26(2), 253-267.
5. Hubbard, R. J., Rommers, J., Jacobs, C. L., & Federmeier, K. D. (2019). Downstream behavioral and electrophysiological consequences of word prediction on recognition memory. Frontiers in Human Neuroscience, 13, 291.
6. Foucart, A., Ruiz-Tada, E., & Costa, A. (2016). Anticipation processes in L2 speech comprehension: Evidence from ERPs and lexical recognition task. Bilingualism: Language and Cognition, 19(1), 213-219.
7. Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46(3), 441-517.

Does structural priming differ between heritage speakers and late bilinguals? Evidence from reaction times for the Mandarin and English transitive alternation

Vera Yunxiao Xia, Johanne Paradis, Juhani Järvikivi

University of Alberta

yunxiao4@ualberta.ca

Structural priming studies show that syntactic repetition aids syntactic processing. In bilinguals, this can occur both within and between languages (Mahowald et al., 2016). The presence and magnitude of priming is impacted by the language of the prime/target (Loebell & Bock, 2003; Schoonbaert et al., 2007) as well as L2 proficiency: L2–L2 and between-language priming might be limited to higher-proficiency bilinguals (Hartsuiker & Bernolet, 2017). However, most of these studies assume bilinguals are dominant in the L1, so it is unclear whether L1 proficiency impacts priming similarly to L2 proficiency. Our priming study compares two groups of bilinguals who share an L1 and L2, but with opposite proficiency patterns: namely, heritage speakers (HSs), who are L2-dominant, and late bilinguals (LBs), who are L1-dominant. 

            We recruited 37 L1-Mandarin, L2-English adult HSs and 53 L1-Mandarin, L2-English adult LBs for a timed aural grammaticality judgment task. They judged 72 targets and 72 fillers, in a mixture of Mandarin and English. The targets were three transitive alternations: active (1a–b), passive with preverbal agent (2a–b), and passive with postverbal agent (3a–b). Each passive was ungrammatical in one language. Each sentence was a potential prime for the next sentence. Prime–target pairs included L1–L1, L2–L2, L1–L2, and L2–L1. For this study, reaction times (RTs) were analyzed for priming. L1 and L2 proficiency, and dominance, were measured with sentence repetition tasks. Research questions were: 

1. How does group membership moderate structural priming? Each group was expected to show stronger priming for their more proficient language, i.e., English for the HSs and Mandarin for the LBs. 
2. Does structural priming occur both within and between languages? Every speaker was expected to demonstrate priming within at least one language. Between-language priming might only occur for speakers proficient in both languages. 

The data were split into Mandarin-primed targets and English-primed targets and analyzed separately with linear mixed-effects models. For both models, we found a significant interaction between prime (repetition of the previous structure) and language match with the previous stimulus; pairwise comparisons indicated that for both models, the primed, language-matched condition had faster RTs than the primed, non-matched condition as well as the non-primed conditions (Figure 1). That is, only L1–L1 and L2–L2 priming occurred; L1–L2 and L2–L1 priming did not occur. There was no three-way interaction with group for either model, indicating the HSs and LBs did not differ. We also found a significant interaction of language match and group for both the Mandarin-primed and English-primed targets: each group slowed down when switching from their respective dominant to non-dominant language (i.e., L2–L1 for the HSs and L1–L2 for the LBs). 

            The results indicate the following: First, structural priming might not always differ between HSs and LBs, since both groups demonstrated L1–L1 and L2–L2 priming. Second, structural priming occurs within languages, but not always between languages. The lack of between-language priming in our study may be explained by the relatively small proportion of bilinguals with high proficiency in both languages, or by the cost of language switching, resulting in slowdown.  

References:

1. Hartsuiker, R. J., & Bernolet, S. (2017). The development of shared syntax in second language learning. Bilingualism: Language and Cognition, 20(2), 219–234. https://doi.org/10.1017/S1366728915000164
2. Loebell, H., & Bock, K. (2003). Structural priming across languages. Linguistics, 41(5), 791–824. https://doi.org/10.1515/ling.2003.026
3. Mahowald, K., James, A., Futrell, R., & Gibson, E. (2016). A meta-analysis of syntactic priming in language production. Journal of Memory and Language, 91, 5–27. https://doi.org/10.1016/j.jml.2016.03.009
4. Schoonbaert, S., Hartsuiker, R. J., & Pickering, M. J. (2007). The representation of lexical and syntactic information in bilinguals: Evidence from syntactic priming. Journal of Memory and Language, 56(2), 153–171. https://doi.org/10.1016/j.jml.2006.10.002

Do multilingual children outperform monolingual children in a visual perspective taking task? A replication of Fan et al. (2015)

Josje Verhagen, Kimberley Mulder, & Elise van Wonderen

University of Amsterdam

j.verhagen@uva.nl

Recent studies have shown that exposure to multiple languages may enhance children’s perspective taking skills (Fan et al, 2015; Liberman et al., 2017; Yow & Markman, 2015). Specifically, an oft-cited study by Fan et al. (2015) found that multilingual 4- to 6-year-olds outperformed monolingual peers in a task that required them to consider the speaker’s visual perspective. In this task, a ‘director’ instructed children to move target objects in a grid in which distractor objects were present that were occluded from the director’s view (see Figure 1). The results showed that multilingual children (both passive and active multilinguals (N = 24 per group)) outperformed monolingual peers (N = 24): they moved the target objects more often and showed fewer first looks towards the distractor objects. The authors concluded that multilingual children more often take another person’s limited visual perspective into account.

In our study, we replicated Fan et al.’s study with a larger sample of monolingual and multilingual children in the Netherlands. The setup was the same as in the original study, except that children’s looking behaviors were not only assessed through video cameras, but also through eye tracking glasses. The Dutch Peabody Picture Vocabulary Test was used to measure receptive vocabulary. Information about children’s language exposure and use was obtained through a caregiver questionnaire. We addressed the same main question as in Fan et al. (2015): Do multilingual children outperform monolingual children on the perspective taking task?

At the moment of writing this abstract, we have collected data from 39 monolinguals (target N = 40, M_age = 5;2) and 28 multilinguals (target N = 80, M_age = 5;6). The monolingual children were only exposed to Dutch; the multilingual children to Dutch and one to three other languages. Preliminary results show that the multilinguals moved the target in 68% of the trials and the monolinguals in 61% of the trials. These accuracy rates are rather similar across the two groups, unlike in Fan et al. who obtained rates of 77% and 50%, respectively. Also, in our study, the number of children moving the target on all four critical trials was comparable across the multilinguals (29%) and monolinguals (28%), which again contrasts with the group difference in Fan et al. (60% vs. 21%).

Based on these preliminary results, we expect to not replicate Fan et al.’s original findings. However, we have not yet analyzed children’s looking behaviors. Moreover, we intend to conduct additional analyses to explore whether perspective taking abilities are enhanced in specific types of multilinguals only, for example, children with relatively low vocabulary knowledge or with caregivers who do not understand some of the languages the child speaks. In so doing, we hope to shed light on the conditions under which any potential effects of multilingualism on perspective taking may or may not occur.

Figure 1. Description and picture of a sample trial

References

1. Fan, S. P., Liberman, Z., Keysar, B., & Kinzler, K. D. (2015). The exposure advantage: Early exposure to a multilingual environment promotes effective communication. Psychological Science, 26(7), 1090–1097. https://doi.org/10.1177/0956797615574699.
2. Liberman, Z., Woodward, A. L., Keysar, B., & Kinzler, K. D. (2017). Exposure to multiple languages enhances communication skills in infancy. Developmental Science, 20(1), Article e12420. https://doi.org/10.1111/desc.12420.
3. Yow, W. Q., & Markman, E. M. (2015). A bilingual advantage in how children integrate multiple cues to understand a speaker’s referential intent. Bilingualism: Language and Cognition, 18(3), 391–399. https://doi.org/10.1017/S1366728914000133.

Educating “the next generation” of SLA researchers: Results of the UPSKILLS project

Tihana Kraš¹ and Maja Miličević Petrović²

¹University of Rijeka, ²University of Bologna

¹tk302@cantab.net

²maja.milicevic2@unibo.it

For over ten years, the field of second language acquisition (SLA) has been characterised by growing methodological awareness, or a “methodological turn” (Byrnes, 2013). A heavy focus has been placed on research quality, especially in terms of study design, quantitative analysis and results reporting, and parallel attention has been devoted to researcher training, through workshops, summer schools and publications (e.g. Plonsky, 2015, 2020).

            Following these trends, this paper presents the results of the Erasmus+ project UPgrading the SKIlls of Linguistics and Language Students (UPSKILLS), a three-year strategic partnership for higher education between institutions from Austria, Croatia, Italy, Malta, the Netherlands, Serbia and Switzerland, which was completed in August 2023. The aim of the project was to identify and tackle the gaps in the skills of linguistics and language students in order to make their university education more compatible with the requirements of the contemporary job market, as well as the trends in academic research. The project grew out of combined insights of academia and industry about the changes that big data and artificial intelligence are causing in the study of language(s).

            Findings will be presented of a needs analysis, which comprised a survey of existing language and linguistics curricula across European universities, a corpus-driven study of job advertisements as well as questionnaires for and interviews with representatives of the language industry. These findings point to the need for developing, through language-related curricula, not only the disciplinary, intercultural and transversal knowledge and skills, but also the research-oriented, data-oriented, technical and organisational ones, such as those related to analytical thinking and problem solving, statistical analysis, text processing, corpus linguistics, programming, machine learning and project management. These knowledge and skills are included in the profile of language data and project specialist developed in the project (Miličević Petrović et al., 2021). 

            A set of open access teaching/learning materials will also be presented, which have been developed for areas identified as in need of improvement, and which are intended as supporting resources to be embedded in existing programmes of study, or as new curriculum components, at the BA and MA level. Special attention will be given to materials intended to develop student skills for designing experimental tasks grounded in linguistic theory (Kraš et al., 2023) and analysing the data statistically (Miličević Petrović et al., 2023) in the area of SLA research, which might be of particular interest to university teachers of SLA and related courses, but also practicing SLA researchers. The results of the piloting of some of these materials in the context of the UPSKILLS summer school held in July 2023 will also be presented.

References:

1. Byrnes, H. (2013). Notes from the Editor. The Modern Language Journal 97(4). 825-827.
2. Kraš, T., Podboj, M., & Simonović, M. (2023). Collecting language data from human participants [Online course]. https://upskills.fil.bg.ac.rs/login/index.php
3. Miličević Petrović, M., Bernardini, S., Ferraresi, A., Aragrande, G., & Barrón-Cedeño, A. (2021). Language data and project specialist: A new modular profile for graduates in language-related disciplines. Zenodo. https://doi.org/10.5281/zenodo.5030929
4. Miličević Petrović, M., Ferraresi, A., & Marty, P. (2023). A glimpse into language data science [Online course]. https://upskills.fil.bg.ac.rs/course/view.php?id=8
5. Plonsky, L. (Ed.) (2015). Advancing quantitative methods in second language research. New York and London: Routledge.
6. Plonsky, L. (Ed.) (2020). Professional development in applied linguistics: A guide to success for graduate students and early career faculty. Amsterdam: John Benjamins.

Effects of culture relatedness on bilingual emotional responses to words: Insights from word norms and event-related potentials (ERPs)

Yanxi Lu, Francesca Citron, Kate Cain, and Bo Yao

Department of Psychology, Lancaster University, Lancaster, United Kingdom,

yanxi.lu@lancaster.ac.uk

k.cain@lancaster.ac.uk

b.yao1@lancaster.ac.uk

f.citron@lancaster.ac.uk

This study introduces culture relatedness of words as a novel variable and explores its impact on emotional responses of English-Mandarin bilinguals living in the UK, where their second language (L2) is dominant. First, we conducted a norming study to identify emotive words related to participants’ native (e.g., bamboo) and residential (e.g., scones) cultures. We then used event-related potentials (ERPs) to examine whether culture relatedness affects emotional responses to words presented in L1 and L2. We were particularly interested in investigating whether the well-established emotional distance from L2 may be due to cultural distance, and whether concepts related to one’s native culture in L2 may enhance affective responses. Initial evidence from ongoing data analyses seems to suggest an interaction of culture relatedness and emotional valance on affective responses. This research offers new insights into the interplay of language, culture, and emotion in bilingual contexts, examining how cultural salience modulates emotional responses.

Explicitness of referring expressions in heritage speakers’ majority English

Tatiana Pashkova, Shanley E. M. Allen

University of Kaiserslautern-Landau

pashkova@rptu.de

allen@rptu.de

Previous research suggests that bilinguals are more explicit than monolinguals in the domain of reference. For example, bilingual children lexicalized more concepts in their narratives than monolinguals [1]. Late L2 speakers used full NPs instead of pronouns, compared to their L1 [2,3] and to L1 speakers of their L2 [3,4]. Concerning majority language of heritage speakers (HSs), the results have been mixed: on the one hand, adult Turkish HSs used more explicit referring expressions in their majority Dutch compared to Dutch monolingual speakers (MSs) [5]. On the other hand, adult Spanish HSs tended to produce less explicit referring expressions in their majority English compared to English MSs [6,7]. Given these conflicting findings, it is still unclear whether higher explicitness is characteristic of HSs’ majority language.

Filling this gap, we compared referring expressions in elicited narratives produced by adolescent and adult German, Greek, Russian, and Turkish HSs and English MSs, all being US English dominant. We hypothesized that HSs would use more explicit referring expressions than English MSs, either due to their communication with L2 English speakers, who might benefit from extra detail [8] or due to more frequent input from L2 English speakers, who themselves might be more explicit [5].

Each speaker produced four narratives in two formalities and two modes (formal spoken and written, informal spoken and written). We annotated the narratives for 19 referents (e.g., WOMAN1, WOMAN2) and three types of referring expressions in order of explicitness – noun-headed NPs (the woman had some groceries), pronouns (she got scared) or null anaphora (and ∅ dropped her groceries). Following previous research [5], we conducted two comparisons: (1) noun-headed NPs vs. pronouns (with 64 English MSs, 34 German, 65 Greek, 65 Russian, and 59 Turkish HSs), and (2) pronouns vs. null anaphora (with 40 English MSs, 42 Turkish and 40 Russian HSs). We expected HSs to use more noun-headed NPs than MSs in the first comparison and more pronouns in the second.

Analyses with linear mixed effects models did not indicate higher explicitness of HSs in the pronoun vs. null anaphor comparison. However, a significant difference was detected in the noun-headed NP vs. pronoun comparison (Figure 1) – Turkish HSs had significantly more noun-headed NPs than English MSs in informal written narratives (both in maintenance and reintroduction) and in informal spoken narratives (in reintroduction). In addition, Russian HSs had more noun-headed NPs in the informal written narratives (in reintroduction).

In other words, HSs were not more explicit than English MSs across all contexts. However, Turkish and Russian HSs did show signs of higher explicitness in informal narratives, although not to the same extent in maintenance and reintroduction. This finding aligns well with the reasoning that HSs’ higher explicitness can stem from their frequent communication with L2 speakers of the majority language – their family and other community members. These L2 speakers are likely to speak to HSs in informal contexts, and thus HSs exhibited the results of this communication in our informal elicitation setting. Not all HS groups used significantly more noun-headed NPs in the informal setting; however, all of them trended towards higher explicitness – the predicted probabilities of a noun-headed NP in the informal narratives are almost always higher for HSs than for English MSs (except for German HSs in informal spoken maintenance, Figure 1). Overall, our findings confirm HSs’ higher explicitness in some majority language areas compared to MSs, although the effect appears limited to certain phenomena and speaker groups.

Figure 1. Predicted Probabilities, Raw Group Proportions and Raw By-Speaker Proportions of Noun-headed NPs (Error bars represent 95% CIs; blue triangles show predicted probabilities, black circles raw group proportions, colored circles raw by-speaker proportions; observation numbers are in parentheses).

Referent tracking type	Previous clause	Current clause	Explanation
Maintenance	Subject The dog ran.	Any syntactic role The driver saw it.	Referent dog in current clause is maintained
	Non-subject The driver saw the dog.	Non-subject He almost hit it.
Reintroduction	Non-subject The driver saw the dog.	Subject The dog ran.	Referent dog in current clause is reintroduced
	Absent The driver turned right.	Any syntactic role He saw the dog.

Table 1. Definitions and Examples of Maintenance and Reintroduction for Given Referents (based on 9, 10, 11).

References:

1. Barbosa, P., Nicoladis, E., & Keith, M. (2017). Bilingual children’s lexical strategies in a narrative task. Journal of Child Language, 44(4), 829-849.
2. Gullberg, M. (2006). Handling discourse: Gestures, reference tracking, and communication strategies in early L2. Language Learning, 56(1), 155–196.
3. Hendriks, H. (2003). Using nouns for reference maintenance: A seeming contradiction in L2 discourse. In A. G. Ramat (Ed.), Typology and Second Language Acquisition (pp. 291–326). De Gruyter Mouton.
4. Yoshioka, K. (2008). Gesture and information structure in L2 discourse. Gesture, 8(2), 236–255.
5. Azar, Z., Özyürek, A., & Backus, A. (2020). Turkish-Dutch bilinguals maintain language-specific reference tracking strategies in elicited narratives. International Journal of Bilingualism, 24(2), 376-409.
6. Contemori, C., Tsuboi, N., & Armendariz Galaviz, A. L. (2023). Referential choice in two languages: The role of language dominance. International Journal of Bilingualism, 24(2).
7. Contemori, C., & Ivanova, I. (2021). Bilingual referential choice in cognitively demanding situations. Bilingualism: Language and Cognition, 24(1), 83-95.
8. Polinsky, M. (2018). Heritage languages and their speakers. Cambridge University Press.
9. Hickmann, M., & Hendriks, H. (1999). Cohesion and anaphora in children’s narratives: A comparison of English, French, German, and Mandarin Chinese. Journal of Child Language, 26(2), 419-452.
10. Perniss, P., & Özyürek, A. (2015). Visible cohesion: A comparison of reference tracking in sign, speech, and co-speech gesture. Topics in Cognitive Science, 7(1), 36-60.
11. Serratrice, L. (2007). Referential cohesion in the narratives of bilingual English-Italian children and monolingual peers. Journal of Pragmatics, 39(6), 1058-1087.

Exploring the link between multilingual experiences, genetic risk factors for Alzheimer’s Disease and cognitive performance in middle-aged adults

Janine Rook^1*, Greg Poarch¹, Vincent DeLuca², Merel Keijzer¹

¹ University of Groningen, the Netherlands

² UiT the Arctic University of Norway, Norway

*j.rook@rug.nl

Aging is associated with a decline in cognitive functioning (Ferguson et al., 2021). Different lifestyle factors have been found to improve the brain’s resilience against (age-related) cognitive decline (i.e., cognitive reserve, CR). Strikingly, multilingualism has been identified  as one of these factors, and has been associated with CR and an attenuation of Alzheimer’s  Disease (AD) symptoms by approximately 4-5 years (e.g., Anderson et al., 2020). 

Recently, research has been mindful of individual differences in multilingual experiences. Such multilingual experiences have, for example, been related to individual differences in neurocognition (e.g. DeLuca et al., 2019, 2020). Research has so far focused predominantly on the younger and older lifespan, but to fully understand how CR builds up as a function of multilingual life experiences, we need a broader lifespan perspective and thus need to include middle-aged individuals, which are crucially underrepresented in the present literature. In addition to modifiable lifestyle factors, such as multilingualism, genetic predisposition may make someone more or less susceptible to cognitive impairment in later life, for instance in AD manifestation (Alzheimer’s Association, 2023). To understand the mechanism underlying the contributions of multilingual experiences to CR and to understand its implication for AD symptom attenuation, we also need to explore the role of genetic risk factors for AD in the age group that just precedes later life. 

The present project examines how individual multilingual experiences of middle-aged adults (45 to 65 years-old) affect neurocognitive outcomes, including behavioral performance on cognitive tasks and functional brain activity (measured via functional Near-Infrared Spectroscopy (fNIRS) and electroencephalography (EEG)). Crucially, we explore how familial risk for AD modulates the effect of multilingual experiences on neurocognition. In this presentation, we outline the method that forms the basis of the project and present preliminary behavioral results. Herein, we focus on performance on a Stroop Arrows task (adapted from Giezen et al., 2015) and Colour-Shape Switching task (adapted from Prior & MacWhinney, 2009). Task performance of our cohort will also be regressed against their individual multilingual experiences, captured using the Language History Questionnaire (Li et al., 2019). 

In line with previous research, we predict that greater degrees of multilingual engagement (i.e., more frequent language use in different contexts, less language dominance) will induce smaller Stroop, facilitation, and inhibition effects, and a smaller switching cost (Freeman et al., 2022; Prior & MacWhinney, 2009). To the best of our knowledge, this is one of the first multilingualism studies that actively investigates the middle-age lifespan and includes genetic risk factors for AD. In this way, we are not only able to uniquely relate individual multilingual experiences to cognitive performance in middle adulthood, but we can also explore how familial genetic risk factors modulate this association. 

References:

1. Alzheimer’s Association (2023). Causes and risk factors for Alzheimer’s Disease.  

https://www.alz.org/alzheimers-dementia/what-is-alzheimers/causes-and-risk-factors

• Anderson, J.A.E., Hawrylewicz, K., & Grundy, J.G. (2020). Does bilingualism protect against dementia? A meta-analysis. Psychonomic bulletin & review, 27(5), 952-965. https://doi.org/10.3758/s13423-020-01736-5 
• DeLuca, V.F., Rothman, J., Bialystok, E., & Pliatsikas, C. (2019). Redefining bilingualism as a spectrum of experiences that differentially affects brain structure and function. Proceedings of the National Academy of Sciences, 116(15), 7565–7574.  https://doi.org/10.1073/pnas.1811513116
• DeLuca, V., Rothman, J., Bialystok, E., & Pliatsikas, C. (2020). Duration and extent of bilingual experience modulate neurocognitive outcomes. NeuroImage, 204, 1-11.  https://doi.org/10.1016/j.neuroimage.2019.116222
• Ferguson, H.J., Brunsdon, V.E.A. & Bradford, E.E.F. (2021). The developmental trajectories of executive function from adolescence to old age. Scientific Reports, 11. https://doi.org/10.1038/s41598-020-80866-1
• Freeman, M.R., Robinson Anthony, J.J.D., Marian, V. & Blumenfeld, H.K. (2022). Individual and sociolinguistic differences in language background predict Stroop performance. Frontiers in Communication, 7, 1-18. https://doi.org/10.3389/fcomm.2022.865965  
• Giezen, M.R., Blumenfeld, H.K., Shook, A., Marian, V. & Emmorey, K. (2015). Parallel language activation and inhibitory control in bimodal bilinguals. Cognition, 141, 9-25. https://doi.org/10.1016/j.cognition.2015.04.009 
• Li, P., Zhang, F., Yu, A., & Zao, X. (2019). Language History Questionnaire (LHQ3): An enhanced tool for assessing multilingual experience. Bilingualism: Language and Cognition, 23(5), 938-944. https://doi.org/10.1017/S1366728918001153 
• Prior, A. & MacWhinney, B. (2009). A bilingual advantage in task switching. Bilingualism: Language and Cognition, 13(2), 253-262. https://doi.org/10.1017/S1366728909990526 

Gray matter volume correlates with language proficiency in the right angular gyrus of early cultural bilinguals

Mia R. Coutinho , Edith Brignoni-Pérez, Alison K. Schug, Guinevere F. Eden

Center for the Study of Learning, Department of Pediatrics, and Interdisciplinary Program in Neuroscience, Georgetown University Medical Center, Washington, DC, USA

mc2618@georgetown.edu

Introduction. Two decades ago, Mechelli et al., investigated the relationship between neuroanatomy and second-language proficiency, thereby offering insights into the anatomical consequences associated with learning a second language (2004). They found a positive correlation between gray matter and second-language proficiency in adult bilinguals in left inferior parietal lobule, a region associated with language. Participants were mostly late bilinguals (having learned their second language in later childhood, adolescence, or adulthood), raising the question whether such a relationship exists in cultural early bilinguals who learn their languages obligatorily as a result of their heritage/environment at a young age. This sets them apart from late second-language learners, whose dual language acquisition may be (innately) driven by a propensity for language learning. Generally, cultural early bilinguals attain high levels of proficiency in both languages yet individually, there is variability in proficiency for both their “weaker” and “stronger” languages.

            Methods.Here we measured by both gray matter volume (GMV) and cortical thickness (CT) in early cultural bilinguals and correlated each with language proficiency in the “weaker” and the “stronger” language in forty-six early Spanish-English bilingual children and adults (average age 16.7 ± 6.65 years; range = 7.7-28.6 years; 31 females). All participants were exposed to both languages prior to or at six years of age, with about one-third being exposed to both languages since birth. All reported using both languages daily. Language proficiency was gauged using analogous standardized test of oral reading ability in Spanish (Identificación de Letras, Muñoz-Sandoval et al., 2005) and in English (Letter Word Identification, Woodcock et al., 2001). Reading ability was in the average or above average range for all participants in both languages. The language with the lower score was assigned as “weaker” and the higher score as “stronger”. Structural magnetic resonance images were acquired on a 3T system (voxel size 1mm³). For GMV analysis, we used standard voxel-based morphometry preprocessing in SPM12 (Ashburner et al., 2000), and entered total GMV and age as covariates of no interest; and for CT analysis, we used CAT12 (Dahnke et al., 2013). Statistical thresholds were voxel-wise height threshold p < 0.005 uncorrected, cluster-level extent threshold p < 0.05 FDR.

            Results. Correlations for GMV revealed more GMV in the right inferior parietal lobule (angular gyrus; BA 39) the more proficient a bilingual was in their weaker language, but there were no relationships with the stronger language. There were no relationships between CT and either the weaker or stronger language.

            Conclusion. In cultural early Spanish-English bilinguals, the only relationship between neuroanatomy and language proficiency was in the right inferior parietal (IPL) lobule  where more GMV correlated with greater proficiency in the “weaker” language. This suggests that bilinguals who acquired their languages in early childhood as part of their culture (rather than due to a propensity for learning languages), higher proficiency in the weaker language is not related to language regions (e.g. left IPL), but to the right IPL, which is associated with attentional control. In other words, when dual-language learning is mandatory, its success may depend on domain-general processes.

References:

1. Mechelli, A., Crinion, J.T., Noppeney,U., O’doherty,J., Ashburner,J., Frackowiak, R. S., & Price, C. J. (2004), Nature, 431, no. 7010, pp. 757-757.
2. Woodcock, R. W., McGrew, K. S., & Mather, N. (2001)
3. Munoz-Sandoval, A. F., Woodcock, R. W., McGrew, K. S., & Mather, N. (2005), The Riverside Publishing Company.
4. Ashburner,J., & Friston,K.J., (2000), NeuroImage, 11, no. 6, pp. 805–821.
5. Dahnke R., Yotter R.A., Gaser C., (2013), Neuroimage. Jan 15; 65:336-48.

Higher level of biliteracy is associated with better executive function in Greek-English bilingual children 

Froso Argyri^1a, Artemis Stefani^1b,c, Theodora Papastefanou^d, Sezgi Goksan ^a,b, Li Wei ^a, Jonathan D Clayden ^b, Frederique J Liegeois ^b 

1. Joint first authors  

UCL Institute of Education, London, UK ^a 

UCL Great Ormond Street Institute of child Health, London, UK ^b 

Vrije Universiteit Amsterdam, Netherlands ^c 

Cyprus University of Technology ^d 

f.argyri@ucl.ac.uk 

Background. Recent research demonstrates that higher engagement in bilingual language environments boosts some features of cognition, including executive function, among bilinguals. Most studies however have explored cognitive effects of bilingualism by examining mainly the role of language use and proficiency without taking into account bilinguals’ literacy experience in both languages. Biliteracy is a crucial aspect of bilingual development that requires acquiring but segregating distinct orthographic, phonological and grammatical rules. Previous research suggests a strong relationship between level of dual language proficiency/use and cognition (e.g., Smith & Briggs Baffoe-Djan, 2019). Extending this premise to the context of biliteracy, the current study examines whether bilingual children with a higher level of biliteracy have better executive function (EF) abilities than bilinguals with lower level of biliteracy. In contrast to previous studies (e.g., Andreou, 2015), we examine biliteracy effects using a range of EF tasks and measure children’s reading abilities directly.   

            Method. Twenty-eight English-Greek bilingual children, exposed to both languages within the first 5 years of life (age M±SD = 10.1±1.2 years 19 females) were tested on Greek and English measures of literacy and online executive function measures. The literacy measures comprised decoding, reading comprehension and reading rate (DADA reading test for Greek and YARC reading test for English). Expressive English and Greek vocabulary (Crichton vocabulary scale) were also assessed. Based on those measures, we constructed three biliteracy indices (mean, cumulative percentage, and difference) for the three literacy measures (decoding, reading rate and reading comprehension) of each language, resulting in nine indices. The executive function measures include the Hearts & Flowers (H&F), Flanker, Corsi span and the CELF-4 Digit Span. Children additionally completed the Raven’s Progressive Matrices (nonverbal reasoning test). Bayesian model comparisons were performed with uninformative priors to test the associations between the biliteracy indices and cognitive performance.  

            Results. Higher biliteracy decoding mean was associated with faster performance in the incongruent condition of the H&F (BF=1.83). Better biliteracy reading comprehension mean was associated with higher Corsi forward span (BF=1.55). The model with the strongest evidential strength comprised the biliteracy decoding in predicting reaction time (Flowers – Hearts) in the H&F task (BF=16.4, strong). This model remained the most supported when tested against others which included age, gender, reasoning ability, Greek and English vocabulary.   

            Conclusions. Our findings suggest that, in school-aged bilingual children, better biliteracy decoding ability is associated with decreased slowing related to increased inhibitory and switching task demands, indicating enhanced efficiency. Secondly, better biliteracy reading comprehension mean is associated with higher nonverbal memory. These results support earlier research studies with monolingual young readers in showing an association between executive function and bilingual reading abilities (e.g., Spencer & Cutting, 2020; Ober, Brooks, Homer & Rindskopf, 2020). Overall, these preliminary findings indicate a mutually beneficial effect of biliteracy and EF development in childhood.  

  

How does language proficiency mediate prediction skills of early bilingual children and adults?

Figen Karaca¹, Susanne Brouwer¹, Sharon Unsworth¹, Falk Huettig^1,2

¹Radboud University, Centre for Language Studies

²Max Planck Institute for Psycholinguistics

figen.karaca@ru.nl

Prediction skills of monolingual children and adult second language (L2) learners have been found to be modulated by their language proficiency as measured by, for instance, their vocabulary knowledge (e.g., Mani & Huettig, 2012) and their knowledge of the target structure (e.g., Brouwer et al., 2017; Hopp, 2013). However, less is known about prediction skills of early bilinguals who acquired their two languages simultaneously, which might provide new insight into the role of language proficiency in predictive processing that neither monolinguals nor adult L2 speakers can (e.g., Karaca et al., 2021). Therefore, this study investigated to what extent prediction skills of early bilingual speakers as well as monolingual speakers are modulated by their language proficiency. 

Thirty Turkish-Dutch bilingual children (M_age=90 months, SD=12), 25 Turkish-Dutch bilingual adults (M_age=26.72 years, SD=5.22), 49 monolingual Turkish-speaking children (M_age=83 months, SD=6), and 24 monolingual Turkish-speaking adults (M_age=27.04 years, SD=6.14) participated in this study. To measure their prediction skills, we used a visual world eye-tracking experiment in which participants listened to simple Turkish sentences in which case-marking on the first noun (NP1) (accusative, nominative) and the verb position (sentence-medial, sentence-final) were manipulated (Table 1). The sentences were coupled with a visual display of three images: NP1, an agent and a patient (e.g., rabbit, fox, carrot) (e.g., Özge et al., 2019). Based on their eye fixations in the predictive time window, participants’ prediction skills were calculated. To measure their proficiency skills in Turkish, we asked participants to narrate a wordless picture book (Frog, where are you?, Mayer, 1969). Based on their narrations, three proficiency measures were calculated: overall accuracy, clausal complexity, and lexical diversity. Furthermore, we administered a sentence repetition task and a vocabulary task in Turkish only with monolingual and bilingual children to measure their proficiency more comprehensively. 

The results of multiple linear regression models showed that overall accuracy, clausal complexity, and lexical diversity did not significantly modulate prediction skills in any groups. Bilingual children’s performance in the sentence repetition task, however, significantly and positively modulated their prediction skills in the verb-final sentences (β=0.171, SE=0.048, z=3.557, p=0.002, adjusted  p= 0.008) while their performance in the vocabulary task did not. 
            These results revealed that certain aspects of language proficiency may not play a role in monolingual and bilingual speakers’ prediction skills as measured by simple sentence structures, while some others might still do (e.g., performance in the sentence repetition task). These findings highlight the complex relationship between language proficiency and prediction in a bilingual mind. 

#	Case Marking	Verb Position	Word Order	Sentence
1	Accusative	Final	OSV	Hızlı  tavşanı   birazdan  şuradaki  tilki∅   yiyecek Speedy rabbitACC   soon    there      foxNOM   eat “The fox over there will soon eat the speedy rabbit”
2	Nominative	Final	SOV	Hızlı∅  tavşan   birazdan  şuradaki  havucu  yiyecek Speedy  rabbitNOM    soon      there  carrotACC     eat “Speedy rabbit will soon eat the carrot over there”
3	Accusative	Medial	OVS	Hızlı  tavşanı     birazdan yiyecek şuradaki  tilki∅ Speedy rabbitACC  soon     eat         there      foxNOM “The fox over there will soon eat the speedy rabbit”
4	Nominative	Medial	SVO	Hızlı∅  tavşan     birazdan yiyecek şuradaki  havucu Speedy rabbitNOM  soon     eat          there   carrotACC “Speedy rabbit will soon eat the carrot over there”

 Table 1. Overview of the manipulations of the experimental sentences

References: 

1. Brouwer, S., Sprenger, S., & Unsworth, S. (2017). Processing grammatical gender in Dutch: Evidence from eye movements. Journal of Experimental Child Psychology, 159, 50-65. 
2. Hopp, H. (2013). Grammatical gender in adult L2 acquisition: Relations between lexical and syntactic variability. Second Language Research, 29, 33–56. 
3. Karaca, F., Brouwer, S., Unsworth, S., & Huettig, F. (2021) Prediction in bilingual children: The missing piece of the puzzle. In E. Kaan & T. Grüter (Eds.), Prediction in second language processing and learning (pp. 116–137).  
4. Mani, N., & Huettig, F. (2012). Prediction during language processing is a piece of cake – but only for skilled producers. Journal of Experimental Psychology: Human Perception and Performance, 38(4), 843–847. 
5. Özge, D., Küntay, A., & Snedeker, J. (2019). Why wait for the verb? Turkish speaking children use case markers for incremental language comprehension. Cognition, 183, 152–180. 

How L2 Accented Speech Influences Grammatical and Natural Gender Prediction

Carly Levy (Penn State University, USA), Felicity Sarnoff (Penn State University), Carrie N. Jackson (Penn State University), Holger Hopp (Technical University of Braunschweig, Germany)

cel5504@psu.edu

fus38@psu.edu

 cnj1@psu.edu

 h.hopp@tu-braunschweig.de  

Research on predictive processing shows that language users modulate their predictions by top-down assumptions and bottom-up evidence about their utility [1]. In a bilingual context, we examine how native listeners modulate their gender predictions according to the perceived reliability of an L2 speaker. L1 Dutch listeners exhibit no ERP response to gender agreement errors when made by an L2 speaker [2], presumably because L1 speakers have top-down knowledge that L2 speakers struggle with grammatical gender. Similarly, when hearing native input with grammatical gender errors, L1 speakers stop predicting nouns based on gender encoded on the article [3]. However, no studies have investigated whether L1 speakers attenuate predictions with L2 speech that does not contain grammatical gender errors or when L2 speakers use natural gender (2), for which L2ers are less likely to make errors [4]. Using the visual-world paradigm, we investigate whether L1 German speakers show differences in predicting upcoming nouns based on natural or grammatical gender cues when listening to error-free L1- and L2-accented speech.

            L1 German listeners (N=58) listened to L1 German-accented German and L2 American-accented German, while looking at a display of four images (Fig. 1). In “different” condition trials, the gender marking on the article uniquely identified the target object, while in “same” condition trials, the article had the same gender for several objects in the display. We tested whether looks to the target noun occur earlier for “different” than “same” trials, indicating that participants used the gender-marked article to predict the target noun.

(1)        Wo       ist         der_MASC gelbe                 Käse_MASC?

             Where is          the        yellow             cheese?

(2)        Wo       ist         die_FEM    blonde              Tänzerin_FEM?

             Where is          the        blonde             dancer?

“Where is the yellow cheese/blonde dancer?”

Divergence point analyses [5] revealed more predictive looks in “different” versus “same” conditions for natural gender sentences, with similar divergence points for both L1- and L2-accented speech, based on overlapping bootstrapped CIs. For grammatical gender sentences, there was no difference in looks to the target prior to noun onset for “same” versus “different” conditions with either L1- or L2-accented speech. Yet, a difference after target noun onset with L2-accented speech indicates participants used grammatical gender predictively, albeit delayed (Fig. 2). Thus, listeners engaged in predictive processing based on natural gender with both L1- and L2-accented speech, yet they engaged in delayed predictive processing for grammatical gender.

In sum, listeners relied on natural gender as a predictive cue in L2-accented speech, presumably since they expect it to be reliable. For grammatical gender, listeners may anticipate gender errors in L2 speech. We hypothesize that when faced with errorless L2 speech, L1 speakers’ surprisal at the speaker’s accuracy may lead to greater use of grammatical gender for prediction. Listeners thus modulate their predictions both according to top-down knowledge and bottom-up evidence in bilingual contexts.

Figures

References:

1. Kuperberg, G. R., & Jaeger, T. F. (2016). What do we mean by prediction in language comprehension? Language, Cognition and Neuroscience, 31(1), 32–59.
2. Hanulíková, A., van Alphen, P. M., van Goch, M. M., & Weber, A. (2012). When one person’s mistake is another’s standard usage: The effect of foreign accent on syntactic processing. Journal of Cognitive Neuroscience, 24(4), 878–887.
3. Hopp, H. (2016). Learning (not) to predict: Grammatical gender processing in second language acquisition. Second Language Research, 32(2), 277–307.
4. Lew-Williams, C., & Fernald, A. (2009). Fluency in using morphosyntactic cues to establish reference: How do native and non-native speakers differ. In Proceedings of the 33rd Annual Boston University Conference on Language Development (Vol. 1, pp. 290-301). Somerville, MA: Cascadilla.
5. Stone, K., Lago, S., & Schad, D. J. (2021). Divergence point analyses of visual world data: Applications to bilingual research. Bilingualism: Language and Cognition, 24(5), 833-841.
6. Grey, S., & van Hell, J. G. (2017). Foreign-accented speaker identity affects neural correlates of language comprehension. Journal of Neurolinguistics, 42, 93–108.

Integration of Multiple Speech Cues for Native and L2 Listeners during Online Speech Processing with Visual-World Eye-Tracking Paradigm

Xiaomu Ren, Clara Cohen

Glasgow University Laboratory of Phonetics, University of Glasgow

x.ren.1@research.gla.ac.uk

Speech processing and comprehension involve the integration of both acoustic-phonetic elements, such as segmental and suprasegmental cues, and semantic and discourse pragmatic information. Listeners must attend to both sources of information, incorporating them in real-time. This study employed a visual-world eye-tracking paradigm to examine the how native and non-native listeners adopt different perceptual strategies during online speech processing, specifi- cally how they allocate their attention to multiple speech cues and integrate prosodic cues and verb semantic information. The study particularly focused on prosodic cues, verb semantics, and whether information is given or new in a broader discourse context. The participants’ eye movements were recorded while they listened to sentences within a brief discourse context. They were then asked to click on the object named in the second sentence, such as the underlined “cabbage” in Here is a [cabbage/captain]. Susan is going to [shred/drink] the [cabbage/CABBAGE] after school. These sentences varied in three binary factors: prosodic accent location (Neutral accent on cabbage vs. Contrastive pitch accent on cabbage), target object information status (Cabbage as new information in …captain..cabbage vs. Cabbage as old information in …cabbage…cabbage), and the semantic alignment between the verb and the target object (Appropriate verb in shred cabbage vs. Inappropriate verb drink cabbage). English and Mandarin stimuli comprised two parts of experiment. Listener groups included native English listeners and native Mandarin listeners learning English as L2 for English stimuli, and native Mandarin listeners and native English listeners learning Mandarin as L2 for Mandarin stimuli. Results showed a complex three-way interaction among L1, prosody, and verb semantics. With both English and Mandarin stimuli, when processing sentences containing old information, non-native listeners tended to adopt perceptual strategies similar to those of native listeners. Specifically, both native and non-native listeners in both languages showed an interaction between prosody and semantics and also exhibited an effect of verb appropriateness, which was exaggerated under the target accent. However, when processing sentences containing new information, non-native and native listeners exhibited divergence. Native listeners can interactively combine both high- and low-level information across a variety of contexts. Non-native listeners can also do this, but only with familiar repeated information. In contexts when they must handle new information, their ability to utilize and integrate different speech cues is reduced.

Keywords: Eye movements; Prosodic cues; Pitch accent; Verb semantics; Information status; Native speech perception; L2 speech perception; Speech Processing

Investigating language effects on cognition using three-gendered languages: the case of Ukrainian simultaneous bilinguals

Oleksandra Osypenko¹ (Lancaster University), Silke Brandt² (Lancaster University), Aina Casaponsa³ (Lancaster University), Panos Athanasopoulos⁴ (Lund University)  

o.osypenko@lancaster.ac.uk

s.brandt@lancaster.ac.uk

a.casaponsa@lancaster.ac.uk

panos.athanasopoulos@englund.lu.se

Does language modulate our cognition? Latest studies argue the need to move beyond this binary question and instead focus on the specific circumstances under which language-related effects emerge (Athanasopoulos & Casaponsa, 2020), while doing so in the context of multilingual speakers (Bassetti & Filipović, 2022). The current study attempts to do both. We attempt to analyse the effects, if any, that two contrasting three-gendered grammatical systems (e.g., Ukrainian and Russian) have on cognition, as well as introduce simultaneous bilinguals into the linguistic relativity research. It extends the previously dominating focus on sequential bilinguals with a distinct L1 and a later acquired L2, as well as explores whether the effects of language vary between the two types of bilingualism. To do so, we adapted a similarity judgment paradigm, presented in a seminal experiment from Phillips and Boroditsky (2003). The current paradigm was employed due to its explicit association with grammatical gender, before further investigating more implicit gender effects of two three-gendered grammatical systems on cognition.

            37 English monolinguals and 63 Ukrainian-Russian bilinguals were recruited for the current study. Participants were asked to provide similarity ratings for 100 pairs of stimuli, comprising depicted conceptually neutral nouns, presented alongside a picture of a male or female character (e.g., a ballerina and a notebook) on a 9-point Likert scale (1 – “Not similar”, 9 – “Very similar”). Our stimulus set included objects with opposite grammatical genders in Ukrainian and Russian (masculine in Ukrainian and feminine in Russian, and vice versa), and nouns with matching genders in both languages (i.e., masculine, feminine, and neuter in both Ukrainian and Russian). Two key predictions were tested using linear mixed models. Firstly, for nouns with matching grammatical gender in both Ukrainian and Russian, we expected to find a more pronounced language effect compared to English monolinguals, with Ukrainian participants providing higher ratings for pairs where biological sex of the character and grammatical gender of the object were congruent in both Ukrainian and Russian. Secondly, for nouns with mismatching genders in Ukrainian and Russian, we anticipated observing an effect driven by the most proficient language within the Ukrainian participant group (i.e., higher ratings to pairs where biological sex of the character and object’s grammatical gender were congruent in their most proficient language).

            Analysis revealed that gender congruence of object-character pairs had no statistically significant impact on similarity ratings. Moreover, Ukrainian participants consistently rated objects as less similar than their English monolingual counterparts across all conditions. As for the second prediction, we found no difference between the ratings of pairs that are congruent and incongruent in participants’ most proficient language. Therefore, we propose a further investigation of both explicit and implicit behavioural measures. Additionally, we suggest extending the study and compare the responses when neuter gender is omitted.

References:

1. Athanasopoulos, P., & Casaponsa, A. (2020). The Whorfian brain: Neuroscientific approaches to linguistic relativity. Cognitive Neuropsychology, 37(5–6), 393–412. https://doi.org/10.1080/02643294.2020.1769050
2. Bassetti, B., & Filipović, L. (2022). Researching language and cognition in bilinguals. International Journal of Bilingualism, 26(1), 3–12. https://doi.org/10.1177/13670069211022860
3. Phillips, W., & Boroditsky, L. (2003). Can Quirks of Grammar Affect the Way You Think? Grammatical Gender and Object Concepts. Proceedings of the Annual Meeting of the Cognitive Science Society, 25(25). https://escholarship.org/uc/item/31t455gf

Language choice and naming difficulty: Evidence from bilingual degraded picture naming

Nora Kennis¹*, Martin J. Pickering¹, Holly P. Branigan¹

¹ Department of Psychology, University of Edinburgh, Edinburgh, UK

nora.kennis@ed.ac.uk

When bilingual interlocutors share the same languages, speakers can use and switch between them as they please. What determines their language choices in contexts like these? Studies on voluntary language switching have shown there are several language-internal factors that affect language choice. For example, speakers may select their non-dominant language for easily accessible words (e.g., more frequent or shorter words; Gollan & Ferreira, 2009) and their dominant language for harder-to-access words. Additionally, since even voluntarily switching between languages can incur a cognitive cost (e.g., de Bruin et al., 2018), bilinguals may switch between languages less when naming is more difficult.  

This study investigated how anticipated difficulty in naming may affect voluntary language choice and switching behaviour. We manipulated the early, non-linguistic stages of the word production process to be easy or difficult using image degradation, which is known to slow down the conceptual stages of word production (Meyer et al., 2007).  

For the first experiment, 86 Dutch-English highly proficient bilinguals performed an online picture naming task (116 items) in which they were free to choose the language for each trial. Half of all pictures were degraded using a superimposed mask of white diagonal lines (see fig. 1). We hypothesised participants would use English (their non-dominant language) and switch languages less on degraded- versus intact-image trials. Response times indicate the manipulation was successful: we found significantly slower naming for degraded than intact images, as predicted, as well as a reverse-dominance effect (e.g., Verhoef et al., 2010) and a voluntary switch cost, both of which are commonly found in bilingual picture naming. Crucially, though, logistic multi-level models refuted our other predictions: there was no effect of image degradation on either language choice or switch choice. As a result, we can be reasonably confident that difficulty in naming alone does not affect voluntary language selection.  

If this is a true effect, these data support modular theories of language production, meaning language decisions at the lexical stage occur independently from the visual processing or conceptual stages (e.g., Levelt et al., 1999). We ran a follow-up experiment on a larger scale to further test this, in which we formalised the variable of word frequency. This enabled us to investigate the potential interaction between image degradation and a linguistic factor with a previously attested effect on language choice. Data from 105 Dutch-English bilinguals show largely the same effects as found in study 1: there was no effect of image degradation on language choice or switching, but bilinguals were more likely to name highly frequent words in their L2 and stick to their L1 for lower-frequency words. Just like in study 1, there were effects of degradation and reverse dominance on naming latencies. Speakers were also faster to name higher-frequency words, but there were no interactions between word frequency and image degradation in the binomial measures or response times. These results further support the hypothesis that conceptual effects at the pre-linguistic stages of production do not “leak” into the lexical and language decision stages.      

References:  

1. de Bruin, A., Samuel, A. G., & Duñabeitia, J. A. (2018). Voluntary language switching: When and why do bilinguals switch between their languages? Journal of Memory and Language, 103, 28–43. https://doi.org/10.1016/J.JML.2018.07.005 
2. Gollan, T. H., & Ferreira, V. S. (2009). Should I stay or should I switch? A cost-benefit analysis of voluntary language switching in young and aging bilinguals. Journal of Experimental Psychology: Learning Memory and Cognition, 35(3), 640–665. https://doi.org/10.1037/A0014981  
3. Meyer, A. S., Belke, E., Häcker, C., & Mortensen, L. (2007). Use of word length information in utterance planning. Journal of Memory and Language, 57(2), 210–231. https://doi.org/10.1016/J.JML.2006.10.005 
4. Levelt, W. J. M., Roelofs, A., & Meyer, A. S. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22(1), 1–38. https://doi.org/10.1017/S0140525X99001776 
5. Snodgrass, J. G., & Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology: Human Learning and Memory, 6(2), 174–215. https://doi.org/10.1037/0278-7393.6.2.174 
6. Verhoef, K., Roelofs, A., & Chwilla, D. J. (2010). Electrophysiological evidence for endogenous control of attention in switching between languages in overt picture naming. Journal of Cognitive Neuroscience, 22(8), 1832–1843. https://doi.org/10.1162/JOCN.2009.21291 

Language learning as a non-pharmacological intervention in older adults with (past) depression

Jelle Brouwer

Radboud University Nijmegen (the Netherlands)

jelle.brouwer@ru.nl

Lifelong bilingualism has been found to delay the onset of Alzheimer’s disease (Woumans et al., 2015), but bilingual experiences later in life have also (i.e., later-life language learning) has also been proposed as a non-pharmacological intervention to promote healthy aging (Antoniou et al., 2013). In this paper we present a study into the effects of language learning specifically for older adults with a history of depression. Cognitive dysfunction is prevalent in this population, even after remission (Bhalla et al., 2006), and those with past depression are also at higher risk of developing Alzheimer’s disease (Butters et al., 2007). We studied the effects of a three-month English language course on psychosocial well-being, cognitive functioning, and English skills in Dutch older adults with (past) depression and a control group.
            To this end, an intervention study was conducted in the Netherlands with 34 older adults (19 with (past) depression; 15 in the control group). Participants were functionally monolingual Dutch speaking older adults (aged ≥65; M = 70). Participants followed English group lessons through video conferencing (due to COVID-19) every other week, and were asked to study independently for 45 mins/day, for 5 days/week. Data were collected before and after the intervention, and at four-month follow-up. The test battery included indices of psychosocial well-being (e.g., depression symptoms, loneliness), cognitive functioning (e.g., cognitive flexibility, working memory), and language outcomes (e.g,. IELTS, PPVT, can-do scales). Multivariate multiple mixed effects regression models were used to assess changes over time.

Participation in the intervention was associated with significant decreases in apathy, social loneliness, and cognitive failures, and improvements in associative memory and global cognition in those with (past) depression. The control group improved on episodic memory and letter-number sequencing. Both groups improved in linguistic self-confidence and lexical access to English, while the group with (past) depression also improved on IELTS measures.

Our results suggest that a short language course has some immediate benefits for cognition and psychosocial well-being. The reductions in apathy and loneliness are especially salient, as this could mean that group-based learning interventions (like later-life language learning) may build up social and motivational reserves (cf. Forstmeijer et al., 2021) over a longer period of time, potentially protecting against psychosocial risk factors of developing dementia such as social isolation.

References:

1. Antoniou, M., Gunasekera, G. M., & Wong, P. C. (2013). Foreign language training as cognitive therapy for age-related cognitive decline: A hypothesis for future research. Neuroscience & Biobehavioral Reviews, 37(10), 2689-2698.
2. Bhalla, R. K., Butters, M. A., Mulsant, B. H., Begley, A. E., Zmuda, M. D., Schoderbek, B., … & Becker, J. T. (2006). Persistence of neuropsychologic deficits in the remitted state of late-life depression. The American Journal of Geriatric Psychiatry, 14(5), 419-427.
3. Butters, M. A., Young, J. B., Lopez, O., Aizenstein, H. J., Mulsant, B. H., Reynolds III, C. F., … & Becker, J. T. (2008). Pathways linking late-life depression to persistent cognitive impairment and dementia. Dialogues in Clinical Neuroscience, 10(3), 345-357.
4. Forstmeier, S., Verena Thoma, M., & Maercker, A. (2021). Motivational reserve: The role of motivational processes in cognitive impairment and Alzheimer’s disease. In G. Sedek, T. Hess, & D. Touron (Eds.), Multiple pathways of cognitive aging: Motivational and contextual influences (1st ed.). Oxford University Press.
5. Woumans, E., Santens, P., Sieben, A., Versijpt, J. A. N., Stevens, M., & Duyck, W. (2015). Bilingualism delays clinical manifestation of Alzheimer’s disease. Bilingualism: Language and Cognition, 18(3), 568-574.

L1 attrition in instructed L1 Spanish-L2 English bilinguals: Evidence from relative clause attachment preferences

Elena García-Guerrero¹,Cristóbal Lozano¹

¹Universidad de Granada

egarciaguerrero@ugr.es

 cristoballozano@ugr.es

Traditional second language acquisition (SLA) research has focused on bilinguals’ second language (L2) as modulated by their first language (L1). However, more recent studies have challenged this unidirectional influence, suggesting that the L1 may also undergo adjustments due to extensive L2 use and exposure (Chamorro et al., 2016; Dussias & Sagarra, 2007), a phenomenon termed L1 attrition. This evidence mostly comes from studies on long-immersed bilinguals in naturalistic settings, but there is scarce data on the potential influence of other L2 exposure types like instructed settings (i.e., classroom settings). This study aims to bridge this gap by investigating potential attrition in bilinguals living in an L1 environment but extensively exposed to formal L2 instruction.

            The present investigation explores L1 attrition related to relative clause attachment (RCA) ambiguities, given the crosslinguistic variation observed in monolingual speakers of different L1s. For instance, the relative pronoun who in (1) may refer to either el alumno/the student, i.e. high attachment (HA), or la científica/the scientist, i.e., low attachment (LA). Native Spanish shows a tendency for HA, whereas native English speakers seem to favour LA (Dussias, 2003; Fernández, 2002; Jegerski et al., 2016).

Thus, this study focuses on: (1) the RCA disambiguation preferences of L1 Spanish-L2 English bilinguals in their L1 as modulated by extensive L2 instructed exposure and (2) whether attrition occurs solely in processing or also in offline comprehension.

            To do so, three groups were tested: advanced L1 Spanish-L2 English bilingual students of English Studies in Spain (N=44), Spanish functional monolinguals (N=50), and English functional monolinguals (N=49). Data were collected from a picture selection task, an auditory sentence-picture verification task, and a visual-world eye-tracking experiment to address both offline comprehension and online processing preferences. All participants were tested in their L1, and target sentences were manipulated to create three conditions (i.e., HA-bias, LA-bias and ambiguous) by manipulating the relative pronoun gender as in (2).

When comparing the bilingual group vs. the Spanish monolingual group, results from the picture selection task and their reaction times show that L2 instructed exposure seems to influence bilinguals’ comprehension and processing of RCA in their L1 Spanish. While a clear HA preference is observed for Spanish monolinguals, bilinguals show more optionality (cf. Error! Reference source not found.), suggesting higher acceptance of the L1 dispreferred strategy (i.e., LA). Additionally, with ambiguous sentences, bilinguals are significantly slower than Spanish monolinguals in the final interpretation of the sentence (cf. Error! Reference source not found.).

References:

1. Chamorro, G., Sorace, A., & Sturt, P. (2016). What is the source of L1 attrition? The effect of recent L1 re-exposure on Spanish speakers under L1 attrition. Bilingualism: Language and Cognition, 19(3), 520–532.
2. Dussias, P. E. (2003). Syntactic ambiguity resolution in L2 learners: Some effects of bilinguality on L1 and L2 processing strategies. Studies in Second Language Acquisition, 25(4), 529–557.
3. Dussias, P. E., & Sagarra, N. (2007). The effect of exposure on syntactic parsing in Spanish–English bilinguals. Bilingualism: Language and Cognition, 10(01), 101.
4. Fernández, E. M. (2002). Relative clause attachment in bilinguals and monolinguals. In Advances in Psychology (Vol. 134, pp. 187–215). Elsevier.
5. Jegerski, J., Keating, G. D., & VanPatten, B. (2016). On-line relative clause attachment strategy in heritage speakers of Spanish. International Journal of Bilingualism, 20(3), 254–268.

L1 variation and L2 acquisition: L1 German /eː/-/ɛː/ overlap and its effect on the acquisition of L2 English /ɛ/-/æ/^[1]

Marcel Schlechtweg¹, Jörg Peters¹, Marina Frank²

¹Carl von Ossietzky Universität Oldenburg, ²Philipps-Universität Marburg

marcelschlechtweg@gmail.com

Although standard German distinguishes between /ɛː/ (Dänen, /ˈdɛːnən/, ‘(the) Danish’) and /eː/ (dehnen, /ˈdeːnən/, ‘(to) stretch’), speakers realize the distinction to different degrees, some maintain it (= /eː/-/ɛː/ maintainers), others merge the two vowels to /eː/ (= /eː/-/ɛː/ mergers). Hence, both groups have (short) /ɛ/ (Kette, /ˈkɛtə/, ‘chain’) in their vowel inventory, but only /eː/-/ɛː/ maintainers have (long) /ɛː/. Standard varieties of English, in turn, distinguish between /ɛ/ (pen,/pɛn/) and /æ/ (pan,/pæn/), whose formants (F1, F2) and duration differ (/æ/ is longer than /ɛ/). Consequently, English /ɛ/ should not pose difficulties to L1 German learners of English (since it exists in German). English /æ/, however, does not occur in German and therefore typically represents a burden for learners of English. Here, we aim at investigating whether the degree of distinctiveness of /eː/-/ɛː/ in German (L1) has an impact on how individuals perceive the English vowel contrast /ɛ/-/æ/ (L2). Could keeping /eː/ and /ɛː/apart (in production) in the L1 German facilitate the perception of the L2 English contrast /ɛ/-/æ/? This question is related to how L1 German speakers might deal with /æ/: Although neither /eː/-/ɛː/ maintainers nor /eː/-/ɛː/ mergers are familiar with /æ/ from their L1, /eː/-/ɛː/ maintainers might benefit from their capacity to differentiate between two vowels ((short) /ɛ/ versus (long) /ɛː/) in a similar area of the vowel space (/eː/-/ɛː/ mergers only have (short) /ɛ/).

            We tested 56 L1 German speakers and relied on two /ɛ/-/æ/ minimal pairs (pen/pan and pedal/paddle). For each pair, a spectral continuum of eleven steps was created (Step 1 = extreme /æ/; Step 11 = extreme /ɛ/). Each spectral step was crossed with the vowel durations short (representing the duration of /ɛ/), middle, and long (representing the duration of /æ/). In an identification task, participants saw two pictures, heard a sound file, and pointed via button press to the picture they associated with the sound (2 pairs, 3 duration types, 11 steps). In a discrimination task, participants heard two sound files and indicated whether the two were the same or different (2 pairs, 3 duration types, distance between the two files was 0, 1, 2, or 3 steps of the continuum).

            The major independent variable was the German Pillai score, which relied on F1/F2 values and indicated the degree of overlap of /eː/ and /ɛː/ for a specific person. These values were obtained in a production experiment on German, where words like Dänen and dehnen were examined. Pillai values range from 0 (absolute overlap of two vowels) to 1 (clear separation). While the results of the discrimination task remained inconclusive, the identification task revealed, among other things, significant interactions GermanPillai*Step and GermanPillai*Duration (Accuracy), indicating that maintaining the vowel contrast in the L1 German led to a more native-like identification of the English /æ/. We will discuss our findings against the background of the role variation in the L1 plays during L2 acquisition.

Reference:

1. Schlechtweg, Marcel, Peters, Jörg, and Frank, Marina. 2023. L1 variation and L2 acquisition: L1 German /eː/-/ɛː/ overlap and its effect on the acquisition of L2 English /ɛ/-/æ/. Frontiers in Psychology 14. 1–13. 

L3 Processing of Verbal Aspect in English by Russian-German Bilingual Children: Evidence from Eye-Tracking

Natalia Mitrofanova¹, Serge Minor¹, Nadine Kolb², Marit Westergaard^1,3

¹UiT The Arctic University of Norway, ²University of Stavanger, ³NTNU

natalia.mitrofanova@uit.no

A growing body of experimental studies has suggested that grammatical representations from both previously acquired languages are likely to influence the development of a third language (see [1] for a review). Two recent studies, [2] and [3], investigated the acquisition of L3 English by bilingual Norwegian-Russian and German-Russian adolescents in comparison to age- and proficiency-matched monolinguals (Norwegian, German, and Russian). Both studies found that the L3 learners differed significantly from both L2 control groups, suggesting combined CLI from both previously acquired languages. The current study continues this line of research adding a processing dimension to L3 research.

     We use Visual World eye-tracking to investigate processing of English aspectual forms by 2L1 German-Russian children in comparison to age- and proficiency-matched monolinguals. The aspectual systems differ in the three languages involved in the study. In Russian, there is a strict grammatical opposition between perfective and imperfective verbs with speakers uniformly associating perfective forms with completed events, and imperfective forms with ongoing events (see [4]). German lacks verbal aspect and doesn’t grammatically encode the contrast between completed and ongoing events. Finally, English employs a specialized progressive form to refer to ongoing events, while the so-called simple past forms are equally compatible with ongoing and completed event interpretations (see Fig. 1 representing processing data collected with L1 English speakers, from [4]). Ermolina [5] adapted the eye-tracking paradigm developed in [4] to investigate the processing of L2 English Past Progressive (PP) and Simple Past (SP) forms by speakers of L1 Russian and L1 Norwegian (Norwegian is similar to German in that it doesn’t encode grammatical aspect). The results revealed that L1 Russians interpreted English SP forms as denoting completed events, while L1 Norwegians favored ongoing event interpretations (see Fig. 2b from [5]). Both groups showed a similar preference for ongoing events when they heard PP forms (see Fig. 2a). We adapted the paradigm from [4] and tested whether bilingual Russian-German adolescents would be influenced by aspectual systems from both of their languages when processing L3 English. Data collection with 50 bilinguals (8-13 y.o., tested in Berlin) has been completed. The results indicate that when the learners heard PP forms, they preferred looking at pictures of ongoing events – more than when they were listening to SP forms (see Fig. 3). Data collection with L1 German controls is ongoing, and the results will be presented at the conference.

Examples of experimental stimuli:

(a) Grandma was knitting a new jumper.

(b) Grandma  knitted a new jumper.

An experimental trial included an audio preamble which located the narrative in the past (e.g. It was a rainy day), followed by a sentence-picture matching task where the participants were presented with a pair of pictures on a screen: one representing an Ongoing Event (OE), i.e. an action in progress, and one representing a Completed Event (CE), see examples (a-b) from the English task. Each experiment included 24 filers and 24 test trials (12 Past Progressive and 12 Simple Past) involving 48 verbs/event types and visual stimuli. Eye-movements were recorded.

References: 

1. Westergaard et al. (2023). Full Transfer Potential in L3/Ln acquisition: Crosslinguistic influence as a property-by-property process. The Cambridge handbook of L3 acquisition and processing.
2. Jensen et al. (2023). Crosslinguistic influence in L3 acquisition across linguistic modules, IJM.
3. Kolb et al. (2022). Crosslinguistic influence in child L3 English: An empirical study on Russian-German heritage bilinguals. IJB.
4. Minor et al. (2023). Aspect processing across languages: A visual world eye-tracking study. Frontiers in Language Sciences.
5. Ermolina (2022) Acquisition of Verbal Aspect in L2 English by advanced learners with L1 Russian and L1 Norwegian: A web-based eye tracking study. MA thesis, UiT.

Meta-CLI: A web app for a community-augmented meta-analysis of cross-linguistic influence in bilingual children

Elise van Wonderen,¹ Chantal van Dijk^2,3,4, Sharon Unsworth²

¹ University of Amsterdam, ² Radboud University, ³ Universität Stuttgart, ⁴ Technische Universität Braunschweig

e.vanwonderen@uva.nl

Although there is ample evidence for cross-linguistic influence in bilingual children, it is still not well understood when and why cross-linguistic influence occurs. In a recent meta-analysis (van Dijk et al., 2022), we found robust evidence for a small to moderate effect of cross-linguistic influence on the level of morphosyntax in simultaneous and early sequential bilingual children (Hedges’ g = 0.39 [0.21;0.56]). We also found evidence for an effect of language dominance, operationalized as societal language, on the strength of cross-linguistic influence. However, the exact mechanism behind cross-linguistic influence and the role of potential predictors, such as language dominance and language overlap, remains under debate. One of the reasons for this is that most studies conducted to date have been underpowered, which means that many more studies are needed to draw reliable conclusions (i.e., many more than the 26 studies analyzed in van Dijk et al., 2022). Another reason is that studies differ widely in their design, the linguistic property under investigation, and the operationalization of both cross-linguistic influence and its potential predictors, making it difficult to pinpoint the exact circumstances under which cross-linguistic influence occurs.   

In our meta-analysis, we argued that in order to obtain a comprehensive picture of cross-linguistic influence, the research community on child bilingualism should work together and combine results from individual – possibly underpowered – studies. To facilitate this, we propose the use of a community-augmented meta-analysis that will continuously update and extend our meta-analysis on structural cross-linguistic influence in bilingual children. For this purpose, we recently developed a web application called Meta-CLI, accessible via https://cls.ru.nl/meta-CLI/. Meta-CLI has two goals. The first goal is to provide an up-to-date meta-analysis by allowing researchers to add their experimental results to the database from our meta-analysis (van Dijk et al., 2022). The second goal is to allow researchers to conduct and visualize their own meta-analyses on (subsets of) the database, by allowing users to (i) filter the database on 10 different variables, including target language, linguistic property under investigation, and children’s age, and (ii) select one of these 10 variables as a predictor in a meta-regression analysis. Results of these analyses will not only further our understanding of cross-linguistic influence, they can also be used by researchers to perform power analyses when designing future studies. In short, Meta-CLI is a collaborative tool for researchers in the field of child bilingualism where they can share and use experimental data on cross-linguistic influence to overcome the limitations of individual studies. With the current poster we would like to introduce Meta-CLI, demonstrate its functionality and gauge the interest of potential users and contributors.  

Negation processing and working memory in Mandarin-Italian bilingual children

Shenai Hu^a,b, Maria Vender^b, Gaetano Fiorin^c, Denis Delfitto^b

^aXiamen University, ^bUniversity of Verona, ^cUniversity of Trieste

shenai.hu@xmu.edu.cn

maria.vender@univr.it

gaetano.fiorin@units.it

denis.delfitto@univr.it

It is well known that negation processing is cognitively demanding, which is challenging for typically developing children (e.g., Bloom, 1970, 1993; Nordmeyer & Frank, 2014). The findings are less clear with respect to the bilingual children (Hu et al., 2018; Schelletter, 2000). The current study aims to understand the processing of negation in Mandarin-Italian bilingual children, and establish the relationship of children’s working memory capacity and their processing of negation.

            50 Mandarin-Italian bilingual children (MeanAge = 9;4, SD = 1.15), 24 Mandarin controls (MA = 9;8, SD = 1.13) and 16 Italian controls (MA = 9;7, SD = 1.03) completed a truth-value judgment task using E-Prime 2.0. The task was manipulated with Sentence Polarity (affirmative vs. negative) and Truth-value (true vs. false), resulting in four conditions: True affirmative, False affirmative, True negative and False negative. In addition, all the bilingual children completed standardized tests of working memory, i.e., forward digit span and backward digit span tasks (Gong & Cai, 1994; Wechsler, 2005).

            First, negative sentences were harder to process than affirmative ones in all the groups, and True negatives were significantly more difficult than the other conditions in both languages. Second, in Mandarin there was no significant difference between the bilinguals and the Mandarin controls in accuracy (e.g., 74% vs. 72% in True negatives), but the bilinguals were significantly slower than the Mandarin controls (e.g., 3861 ms vs. 3404 ms in True negatives). Third, in Italian the bilinguals were more accurate than the Italian controls, especially in True negatives (70% vs. 33%), but there was no significant difference in response latency between the two groups (e.g., 4336 ms vs. 4128 ms in True negative). Fourth, bilingual children’s working memory capacity was positively correlated with negation processing, regardless of whether the language is Mandarin or Italian.     

            The finding confirms the existence of the asymmetry between affirmative and negative sentences, and the greater difficulties in the process of True negatives, indicating that the difficulties in the processing of negation, in the relevant experimental conditions, have a general validity across the monolingual and bilingual children. This result can be readily interpreted in the light of the “experiential” theories of negation processing, by means of the two-step simulation hypothesis (Kaup et al. 2005, 2006, 2007). Our results confirmed that negation processing is cognitively demanding, requiring a richness of computational resources in terms of memory processing. Finally, it should be emphasized that bilingual children were more accurate in the comprehension of Italian True negatives than Italian monolingual children, a finding that deserves further investigation.

Prediction-based grammatical learning in L2 processing

Duygu F. Şafak & Holger Hopp

TU Braunschweig

d.safak@tu-braunschweig.de

In English, most ditransitive verbs can take both double-object dative (DO) and prepositional-object dative (PO). This optionality is constrained by gradient selectional restrictions on the type of their complements: DO-bias verbs, e.g., pay, tend to occur more frequently with DO, while PO-bias verbs, e.g., send, show a probabilistic tendency to prefer PO [1]. Research on priming in L1 comprehension indicates that the magnitude of structural priming increases when the prime structure goes against its verb’s bias [2-3]. Such surprisal effects of verb bias among child L1 learners and adult L1 speakers are in line with implicit learning models, claiming that prediction error is a key mechanism underlying structural priming [4]. However, it is an open question whether prediction-error-driven implicit learning accounts of priming extend to L2 learners.

            Against this backdrop, we probe whether prediction error constitutes a mechanism of structural priming also in a late-learnt L2. In two experiments combining priming and visual world eye-tracking, we first test whether prediction error, occasioned by verb bias, affects structural priming in L2 comprehension. Second, we examine whether structural priming occurs in the absence of prediction error.

            In Experiment 1, adult L1 German intermediate to advanced L2 learners of English (n = 48) first read aloud written prime sentences crossing Verb Bias (DO-bias vs. PO-bias) and Structure Type (DO vs. PO). Subsequently, they listened to spoken target sentences – with non-biased verbs (e.g., show) – while viewing visual scenes with an agent referent (the tailor), a theme referent (the dress), and a recipient referent (the model); see Figure 1. Cluster-based permutation analyses revealed PO-priming, as evidenced by more looks to the theme than to the recipient during the post-verbal temporal region in target sentences following PO (vs. DO) prime sentences (Figure 2a). Further, PO-priming was modulated by surprisal effects of verb bias, as priming was larger when the prime structure mismatched the bias of prime verb, i.e., after PO prime sentence with DO-bias verbs (Figure 2b). These effects show that L2 learners adapt to the structure of the recently processed prime sentence through learning from their prediction errors.

            In Experiment 2 with a comparable group of L1-German–L2-English learners (n = 48), the materials were the same as in Experiment 1, except that prime sentences were preceded by a context sentence introducing the themes and recipients. Moreover, the full noun-phrase themes and recipients in PO prime sentences were replaced by their pronominal counterparts (Figure 3). DO-bias prime verbs thus elicited no prediction error, given that PO is the only possible structure for any ditransitive verb with two pronominal complements. Unlike in Experiment 1, PO-priming did not reach significance in Experiment 2 (Figure 4). This suggests that, when there is no prediction error to learn from, structural priming may not be strong enough to affect L2 learners’ processing of the target sentences.

            The findings from Experiments 1 and 2 demonstrate that priming in L2 comprehension is driven by prediction error like in L1 comprehension, and further point to an error-based prediction mechanism underpinning L2 structural priming.

References:

1. Şafak, D. F., & Hopp, H. (2023). Cross-linguistic differences in predicting L2 sentence structure: The use of categorical and gradient verb constraints. Studies in Second Language Acquisition, 45, 1234–1260.
2. Fine, A. B., & Jaeger, T. F. (2013). Evidence for implicit learning in syntactic comprehension. Cognitive Science, 37(3), 578–591.
3. Chen, X., Wang, S., & Hartsuiker, R. J. (2022). Error-based structure prediction in language comprehension: Evidence from verb bias effects in a visual-world structural priming paradigm for Mandarin Chinese. Journal of Experimental Psychology: Learning, Memory, and Cognition, 48, 60–71.
4. Dell, G. S., & Chang, F. (2014). The P-chain: Relating sentence production and its disorders to comprehension and acquisition. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1634), 20120394.

Predicting production: Individual differences and possible sources of cross-linguistic influence in a third language

Yevgeniy Melguy^[2]a, Clara Martin^a,b, Arthur Samuel^a,b,c, Efthymia Kapnoula^a,b

^aBasque Center on Cognition, Brain and Language, ^bIkerbasque, ^cStonybrook University

y.melguy@bcbl.eu

Learners of a second language (L2) will often fail to achieve a native-like pronunciation, due to the influence of their L1 phonology. For instance, the Speech Learning Model or SLM (Flege, 1995) claims that if a speaker perceives an L1-L2 sound pair as similar (e.g., /d/ in Spanish dolor “pain” vs. English dollar), learning is unlikely to occur: speakers will produce the L2 sound by substituting the closest L1 equivalent. This study extends this paradigm to L3 acquisition: what is the source of “accent” when multiple phonologies already exist? Studies of L3 pronunciation have yielded mixed findings, with some showing transfer from the L1 (Llama & López-Morelos, 2016), others from the L2 (Llama et al., 2010), and still others from both L1 and L2 (Sypiańska, 2016).

            This study examines the possible sources of cross-linguistic transfer in L3 speakers while also addressing the relationship between perceptual sensitivity and production accuracy. Listeners have been shown to differ in their baseline sensitivity to subphonemic (within-category) differences (Kapnoula et al., 2017, Apfelbaum et al., 2022), with higher sensitivity in the L1 predicting vocabulary size in a non-native language (Kapnoula & Samuel, in press). We ask if  perceptual sensitivity in the L2/L3 predicts L3 production accuracy.To test these questions, we examine sibilant fricatives (see Figure) in L1 Spanish – L2 Basque – L3 English speakers. While English has a two-way sibilant contrast (Collins & Mees, 2003), Basque has a typologically rare 3-way contrast (Hualde et al., 2010). The sound missing from the English inventory (written <s> in Basque) is typically described in the literature as apical (involving the tongue tip) (Hualde et al., 2010), and is the same sound as the <s> in Castilian Spanish (Martínez Celdrán et al., 2003). Crucially, it differs from the English <s>: the latter is laminal (involving the tongue blade), sharing place of articulation with Basque <z>.

            We present preliminary results from 20 early Spanish-Basque bilingual speakers of L3 English, living in Donostia-San Sebastian, Spain. We assessed production by obtaining acoustic measures (center of gravity) of these sounds in each language, and assessed perceptual sensitivity by testing listeners’ perception of phonetic fricative continua in Basque and English (e.g., ship → sip), and then measuring the slope of each listeners’ categorization function to obtain a measure of sensitivity (Kapnoula et al., 2017). Group-level results show distinct patterns in the acoustic realization of L3 English <sh> and <s>:  speakers substitute Basque <x> for English <sh>, but produce English <s> with center of gravity between Basque <x> and <s>. However, individual-level results reveal distinct patterns of transfer across speakers. We discuss these differences in terms of the acoustic similarity of Basque and English sibilants, as well individual speakers’ perceptual sensitivity to within-phoneme differences.

References:

1. Apfelbaum, K. S., Kutlu, E., McMurray, B., & Kapnoula, E. C. (2022). Don’t force it! Gradient speech categorization calls for continuous categorization tasks. The Journal of the Acoustical Society of America, 152(6), 3728–3745. https://doi.org/10.1121/10.0015201
2. Collins, B. D., & Mees, I. (2003). The phonetics of English and Dutch. In The Phonetics of English and Dutch. Brill.
3. Flege, J. E. (1995). Second Language Speech Learning Theory, Findings, and Problems. In W. Strange (Ed.), Speech Perception and Linguistic Experience: Issues in Cross-Language Research. York Press.
4. Hualde, J. I., Lujanbio, O., & Zubiri, J. J. (2010). Goizueta Basque. Journal of the International Phonetic Association, 40(1), 113–127. https://doi.org/10.1017/S0025100309990260
5. Kapnoula, E. C. & Samuel, A.G. (in press). Sensitivity to Subphonemic Differences in First Language Predicts Vocabulary Size in a Foreign Language. Language Learning. DOI: 10.1111/lang.12650
6. Kapnoula, E. C., Winn, M. B., Kong, E. J., Edwards, J., & McMurray, B. (2017). Evaluating the sources and functions of gradiency in phoneme categorization: An individual differences approach. Journal of Experimental Psychology: Human Perception and Performance, 43(9), 1594–1611. https://doi.org/10.1037/xhp0000410
7. Llama, R., Cardoso, W., & Collins, L. (2010). The influence of language distance and language status on the acquisition of L3 phonology. International Journal of Multilingualism, 7(1), 39–57. https://doi.org/10.1080/14790710902972255
8. Llama, R., & López-Morelos, L. P. (2016). VOT production by Spanish heritage speakers in a trilingual context. International Journal of Multilingualism, 13(4), 444–458. https://doi.org/10.1080/14790718.2016.1217602
9. Martínez-Celdrán, E., Fernández-Planas, A. Ma., & Carrera-Sabaté, J. (2003). Castilian Spanish. Journal of the International Phonetic Association, 33(2), 255–259. https://doi.org/10.1017/S0025100303001373
10. Sypiańska, J. (2016). L1 vowels of multilinguals: The applicability of SLM in multilingualism. Research in Language, 14(1), 79–94. https://doi.org/10.1515/rela-2016-0003

Priming ditransitives in L2 Mandarin leads to adjustments in production but not in predictive processing

Yanxin (Alice) Zhu, Theres Grüter

University of Hawai‘i

yanxinz@hawaii.edu

Using a visual world structural priming (VWSP) paradigm, Arai et al. (2007) and Thothathiri & Snedeker (2008a, 2008b) found that reading a double object (DO) or prepositional (PO) dative can lead to subsequent prediction of the same dative construction in real-time comprehension among adult and child native speakers of English. Chen et al. (2022) replicated these findings with native Mandarin speakers; additionally, an inverse frequency effect indicated that priming effects were stronger after DO primes with PO-biased verbs (the unexpected combinations) than with DO biased verbs. Şafak & Hopp (under review) extended the inquiry to L1 German L2 English learners and replicated the construction-level priming effects and inverse frequency effects. These studies support error-driven learning (EDL) accounts for the dative alternation (e.g., Chang et al., 2006; Goldberg, 2019), which suggest that language users predict based on dative verb bias and adjust their linguistic knowledge after encounters of prediction error. None of these VWSP studies, however, included baseline or post-priming phases assessing real-time verb-based prediction in the absence of immediate priming. It thus remains unclear whether these immediate comprehension priming effects translate into longer-term adaptation, as would be expected under EDL accounts. In the present study, we use a pretest-priming-posttest design to address whether structural priming can lead L1 and L2 Mandarin speakers to adapt (1) their predictions in real-time comprehension, and (2) their production of dative constructions.

            Method. L1 (N=65) and intermediate-to-advanced proficiency L2 (N=64, MLEXTALE_CH = 61.3/90, range=39-78, SD=10.1) participants completed a VWSP task where they took turns reading aloud sentences (prime trials) and listening to sentences while looking at visual scenes (target trials) containing three entities (agent, theme, goal, for experimental trials; Fig.1). In the priming phase (but not in baseline and posttest), all primes were DO sentences. The experimental target sentences were DO or PO paired with PO-biased verbs. No lexical boost was utilized. Participants also completed written picture description tasks eliciting ditransitives one day before and after the VWSP task to assess longer-term priming.

            Results. We used LMER models to examine the likelihood of participants looking at the theme vs. goal (log-ratio) during two ambiguous time windows (CR1=verb; CR2=perfective +numeral+ classifier; Fig.1). We interpret changes from CR1 to CR2 as evidence of prediction based on verb information. Neither L1ers nor L2ers showed evidence of prediction based on verb bias at any phase of the task, nor did their looking patterns change from baseline to posttest (Fig.2). However, logistic regression models of the production data indicated both groups produced more DO datives one day after priming (b=1.88, p<.001, Fig.3), and L1ers’ increase (b=2.07, p<.001) was significantly larger than L2ers’ (b=1.97, p=.003).     

            Unlike previous VWSP studies (but see also Chen & Hartsuiker, 2023), we observed no immediate priming effects in real-time comprehension, nor longer-term adaptation of real-time prediction. Nevertheless, the priming treatment led to adaptation in production from pre- to posttest. Notably, the same pattern of change was observed in L1 and L2 speakers. Implications for priming as a learning mechanism in L1 and L2 will be discussed.

Note. Both CRs are fully ambiguous.

Fig.2. Proportion looks for PO-biased verbs under DO priming by condition in VWSP

Fig.3. Productions with PO-biased verbs pre- vs. post-DO-priming

References:

• Arai, van Gompel & Scheepers (2007). Priming ditransitive structures in comprehension. Cognitive Psychology, 54.
• Chang, Dell, & Bock (2006). Becoming syntactic. Psychological Review, 113.
• Chen, Wang & Hartsuiker (2022). Error-based structure prediction in language comprehension: Evidence from verb bias effects in a visual-world structural priming paradigm for Mandarin Chinese. JEP:LMC, 48.
• Chen & Hartsuiker (2023). Structure prediction occurs when it is needed: Evidence from visual-world structural priming in Dutch comprehension. JEP:LMC, 49.
• Goldberg (2019). Explain me this: Creativity, competition, and the partial productivity of constructions. PUP.
• Thothathiri & Snedeker (2008a). Give and take: Syntactic priming during spoken language comprehension. Cognition, 108. 
• Thothathiri & Snedeker (2008b). Syntactic priming during language comprehension in three- and four-year-old children. JML, 58.

Processing Differences in Early and Late Bilinguals as Revealed from Linguistic and Neurolinguistic Data

Prof Hideyuki Taura

Ritsumeikan University, JAPAN

pdf02662@gmail.com

Language proficiency assessment has conventionally targeted language learners’ performance based on such aspects as complexity, accuracy, fluency, and vocabulary. Meanwhile, the last 3 decades have witnessed a rapid advancement in technology, which enables linguists to obtain neuro-linguistic data using fMRI, ERP, and eye-tracking data alongside linguistic data. To date, however, there has been scant research reporting on both aspects of a language learners’ performance, that is, the inclusion of linguistic and neuro-linguistic data. The present study attempts to fill the gap by collecting both linguistic and neuro-linguistic data and examining the correlation between them.

An early Japanese-English bilingual and a highly advanced Japanese learner of English (both females in their thirties) participated in this study. For linguistic analysis, spontaneous narrative and written data were collected, using ‘Frog, where are you?’ (Mayer, 1969) and the Test of Written Languages III (Hammill and Larsen, 1996), respectively. For the neuro-linguistic analysis, fNIRS, ERP, and eye-tracking data were collected while the participants were engaged in four tasks – three production tasks and one comprehension task. They include i) a verbal fluency task, ii) a bilingual Stroop test, and iii) oral and written consecutive interpretation for production along with iv) Garden Path sentence comprehension. The preliminary analysis revealed the possibility of brain activation data being a precursor to linguistic changes. The full results are discussed in terms of language processing similarities and differences between the two types of bilinguals (early vs late) by examining the two types of data (linguistic vs psycho-physiological).

Processing Gender Stereotypes in the Bilingual Brain

Joanna Porkert (University of Groningen), Anna Siyanova-Chanturia (Te Herenga Waka – Victoria University Wellington), Merel Keijzer (University of Groningen)

j.porkert@rug.nl

Research has shown that children being brought up bilingual show less racial bias than their monolingual peers (Singh et al., 2019; 2020). In addition, Singh et al. (2021) found that implicit racial bias was predicted by better cognitive flexibility in bilinguals. It raises the question whether this can also be found in young adults who are either sequential or simultaneous bilinguals, respectively. Moreover, it would be interesting to investigate whether other aspects of social cognition, such as gender stereotyping, can also be predicted by cognitive flexibility.

Psycholinguistic EEG research using the event-related brain potential (ERP) technique has shown that gender stereotypes are a relevant cue during language processing, for example, when establishing coherence in pronoun resolution, such as in “Our aerobics instructor gave himself a break.” (p. 284, Osterhout et al., 1997). Findings have shown that in the case of pronoun resolution, participants tend to perceive gender stereotype violations as grammatical mistake, even though the sentences were grammatical, by showing a P600 effect (typically elicited for morphosyntactic violations). For pronoun resolution in general, individual differences were observed on whether a P600 or Nref (a component typically elicited for ambiguous pronouns) is elicited based on working memory abilities (Nieuwland & van Berkum, 2006). The aim of this project is to investigate gender stereotype violations during pronoun resolution of simultaneous and sequential bilingual young adults with different L1 backgrounds. A secondary aim of this project is to examine individual language processing patterns as past gender stereotype studies found, for example, that participants who score low on sexism tend to show reduced amplitudes in the course of the ERP experiment (Grant et al., 2020). Recent research has also shown that interesting individual ERP patters may become lost when only looking at ERP effects on a group level (Grey, 2023).

We tested 66 university students who were either Dutch or Spanish L1 speakers, using an English reading task. All of our participants were sequential English L2 learners, and around half of our participants were, in addition, Dutch-Frisian (n = 16), or Spanish-Catalan (n = 14) simultaneous bilinguals. The stimuli sentences were sentences with both personal, and reflexive pronouns as critical words (e.g., “The lifeguard threw himself into the water.”). The collected background variables consisted of for instance the language and social background, the LexTale English proficiency score, but also scores in ambivalent sexism (a scale that measures hostile and benevolent sexism), in social dominance (the preference of strong hierarchies in society), or attitudes towards gender equality. Performance in working memory was measured with a Corsi blocks task, and cognitive flexibility with a color-shape switch task.

We hypothesize that abilities in working memory will affect how the pronoun violations will be processed, with higher ERP amplitudes for reflexive in comparison to personal pronouns. We expect that attitudes towards sexism and gender equality will lead to higher and more stable ERP amplitudes. Additionally, we think that cognitive flexibility will correlate with gender stereotyping, and that simultaneous bilinguals will show higher cognitive flexibility.

References:

1. Grant, A., Grey, S., & van Hell, J. G. (2020). Male fashionistas and female football fans: gender stereotypes affect neurophysiological correlates of semantic processing during speech comprehension. Journal of Neurolinguistics, 53. https://doi.org/10.1016/j.jneuroling.2019.100876
2. Grey, S. (2023). Variability in native and nonnative language: an ERP study of semantic and grammar processing. Studies in Second Language Acquisition, 45(1), 137–166. https://doi.org/10.1017/S0272263122000055
3. Nieuwland, M.S., & Van Berkum, J.J.A. (2006). Individual differences and contextual bias in pronoun resolution: evidence from ERPs. Brain Research, 1118(1), 155-167.
4. Singh, L., Moh, Y., Ding, X., Lee, K., & Quinn, P. C. (2021). Cognitive flexibility and parental education differentially predict implicit and explicit racial biases in bilingual children. Journal of Experimental Child Psychology, 204. https://doi.org/10.1016/j.jecp.2020.105059
5. Singh, L., Quinn, P. C., Qian, M., & Lee, K. (2020). Bilingualism is associated with less racial bias in preschool children. Developmental Psychology, 56(5), 888-896. https://doi.org/10.1037/dev0000905
6. Singh, L., Quinn, P. C., Xiao, N. G., & Lee, K. (2019). Monolingual but not bilingual infants demonstrate racial bias in social cue use. Developmental Science, 22(6). https://doi.org/10.1111/desc.12809

Processing of Prosody and Case Marking in Turkish Monolingual and Heritage Language Speakers

Selim Tiryakiol¹, Leyla Zidani-Eroglu², Fatih Bayram³

¹The Arctic University of Norway/Istanbul Medeniyet University

²Central Connecticut State University

³The Arctic University of Norway

selim.tiryakiol@uit.no

This study investigates the sensitivity of Turkish heritage speakers residing in Norway to prosody and overt case marking cues in interpreting verb-medial sentences using a visual world eye-tracking paradigm. We examine how the human parser utilizes various linguistic cues—such as morphological case marking, argument order, and prosody—during predictive language processing. While prior research has primarily focused on the interaction between case marking and word order, our study extends the inquiry to include the role of prosody in argument interpretation, both independently and collectively.

            Participants in this study include school-aged heritage speakers and their input providers, as well as age-matched monolingual peers and adults. We piloted the experiment with 25 adult monolingual Turkish speakers who were presented with images depicting characters (a monster and an alien) performing reversible actions, described in accompanying audio recordings. The experiment featured five critical conditions: Prosodic-SVObiasing, Prosodic-OVSbiasing, Neutral, Neutral-NP1acc marked, and Neutral-NP2acc marked.

            Pilot data reveal that monolingual Turkish adults rely on overt case marking as a cue, directing their gaze to the correct image post-NP1 in the NP1acc marked condition and post-verb in the NP2acc marked condition. Their offline accuracy scores corroborate these observations, consistent with previous studies that underscore the reliability of accusative marking in Turkish. Prosodic cues also emerged as significant, particularly in the SVO biasing condition where a clear effect was noted immediately after NP1 upon hearing rising intonation. This pattern mirrors that seen in the NPacc marked condition, suggesting that prosody may function similarly to overt case marking in real-time sentence interpretation. Conversely, the OVS-biasing prosodic condition showed only a transient effect, with no sustained difference in attention between the target and competitor images.

            As the study progresses, we will refine our experimental design based on these findings and expand our data collection to include both monolingual and heritage speakers in the forthcoming months.

Keywords: Prosody, Overt Case Marking, Predictive Processing, Heritage Language, Eye-Tracking, Turkish

Pronoun resolution in adult L2 learners of German by speakers of null- and overt-subject languages: Cross-linguistic influence or a learning mechanism?

Angelika Golegos¹, Anna Czypionka¹, Theodoros Marinis¹

¹University of Konstanz

angelika.golegos@uni-konstanz.de

Research on pronoun resolution by native speakers of null subject languages (NSLs) who acquire a non-null subject language (NNSL) as a second language (L2) have shown mixed findings: some report similar preferences in L2 learners and L1 speakers (e.g., Cunnings et al., 2017), whereas others reveal differences between the two groups (e.g., Roberts et al., 2008). To date, no studies have investigated if the different outcomes are due to cross-linguistic influence from the L1 to the L2, due to an L2 learning mechanism, or due to a different reason. Our study fills this gap by investigating whether L1 speakers of NSLs (Greek, Italian) and L1 speakers of NNSLs (English, Dutch) transfer resolution preferences to their L2 German (NNSL) or apply a learner mechanism and therefore perform similarly to each other irrespective of their L1s.

In ambiguous contexts, German allows for two different third-person singular masculine pronouns to resolve ambiguity: the personal pronoun er and the demonstrative pronoun (henceforth d-pronoun) der, (1). L1 speakers of German show a strong preference to resolve the d-pronoun towards the object and a moderate preference to resolve the personal pronoun towards the subject (Schumacher et al., 2017). We included an additional condition that has not been investigated before to test the role of prosody, i.e., whether stressing the personal pronoun reverses the resolution preferences from the subject to the object (Féry, 2017).

1. Der Tiger_j will den Igel_k … küssen. Aber er_(j)/der_(k)/ER_(k?) kann nicht stillstehen.

The tiger wants the hedgehog … kiss. But he cannot hold still.

We tested pronoun resolution preferences in 171 L2 learners of German in four groups: 29 L1-Greek, 29 L1-Italian, 29 L1-English, 19 L1-Dutch (data collection ongoing), and 65 L1-German (controls). In an online picture selection task, participants listened to sentences that contained two competing characters, followed by an ambiguous third-person singular masculine pronoun (er, der, stressed-er), see (1). While listening to the sentences, participants saw three images, representing the subject, the object and a distractor. After each sentence, participants answered a comprehension question that forced them to interpret the pronoun towards the subject or object by clicking on the image of the respective character.

            Results showed a significant (p < 0.001) difference between the L2 learners and L1 controls, see Figure 1. Both NSL groups resolved the personal pronoun and the d-pronoun at chance. The two NNSL groups differed significantly from each other (p < 0.001), but showed a tendency in resolving both the personal and d-pronoun towards the subject. Finally, the stressed pronoun was resolved by all learner groups towards the object, unlike the L1 controls who showed no preference. Our study indicates that despite some similarities in resolution preferences based on L1 properties (NSL vs. NNSL) there is no evidence that pronoun resolution is driven only by cross-linguistic influence or by a universal learner strategy across the board. Hence, earlier contradictory findings cannot be reduced to a transfer effect of L1 features. Rather, both seem to play a role, jointly shaping learners’ performance in complex ways.

Figure 1: Subject resolution in L2 learners of German for the personal pronoun er, the d-pronoun der, and the stressed personal pronoun er.

References:

1. Cunnings, I., Fotiadou, G., & Tsimpli, I. (2017). Anaphora     resolution and reanalysis during L2 sentence processing:     Evidence from the Visual World Paradigm. Studies in Second        Language Acquisition, 39(4), 621-652.      doi:10.1017/S0272263116000292
2. Féry, C. 2017. Intonation and Prosodic Structure. Key Topics in      Phonology. Cambridge, New York: Cambridge University Press.
3. Roberts, L., Gullberg, M., & Indefrey, P. (2008). Online pronoun      resolution in L2 discourse: L1 influence and general learner         effects. Studies in Second Language Acquisition, 30(3), 333-357.      doi: 10.1017/S0272263108080480
4. Schumacher, P. B., Roberts, L., & Järvikivi, J. (2017). Agentivity      drives real-time pronoun resolution: Evidence from German er      and der. Lingua, 185, 25-41. doi: 10.1016/j.lingua.2016.07.004

Testing the Switching Experience and Environment Questionnaire (SEEQ): A measure of bilingual switching behaviours

Zlatomira G. Ilchovska^1,2, Ali Mazaheri^2,3, Andrea Krott^2,3

¹School of Psychology, University of Nottingham, UK

²School of Psychology, University of Birmingham, UK

³Centre for Human Brain Health, University of Birmingham, UK

z.g.ilchovska@bham.ac.uk

a.mazaheri@bham.ac.uk

a.krott@bham.ac.uk

Language switching is one of the most important aspects of bilingualism believed to be important for neural and behavioural adaptations (e.g. UBET; DeLuca et al., 2020). It has been measured in various questionnaires. But no questionnaire captures all aspects that have been argued to be of importance. Hartanto & Yang’s (2016; 2020) Code-Switching and Interactional Contexts Questionnaire captures language switching tendencies between-sentences and within-sentences across different daily environments, but does not capture whether these switches were intentional (controlled) or unintentional. Neither does it capture the broader interactional contexts as defined in the Adaptive Control Hypothesis (ACH, Green & Abutalebi, 2013). Rodriguez-Fornells et al.’s (2012) Bilingual Switching Questionnaire captures the reason for language switching, i.e., whether it is contextually relevant (thus, produced in a fully controlled manner) or whether it is produced by mistake, due to lack of control over the other language. Finally, Anderson et al.’s (2017) LSBQ is an excellent instrument for measuring language usage and proficiency, but measures language switching only in three selected contexts (with family, with friends and on social media). It does not distinguish the type, frequency or intentionality of switching.

Consequently, we created a new questionnaire, the Switching Experience and Environment Questionnaire (SEEQ), incorporating what we deemed as the strengths of each of the three questionnaires above, in order to compile an instrument that would better suit bilingualism researchers, particularly those interested in language switching and executive control. Hence, the SEEQ measures type of interactional environment (as defined in the ACH), intentionality (control) and contextuality of the switch, frequency, and type of dense code-switch across four different daily-life contexts. In order to better characterize the bilingual individual, we also included proficiency as well as contextual and life-long usage variables from the LSBQ (Anderson et al., 2017).

We next tested whether the SEEQ captures theoretically important factors of the ACH (Green & Abutalebi, 2013) and the Unified Bilingual Experience Trajectory Model (UBET; DeLuca et al., 2020), namely Dominance/Duration of Bilingualism, Intensity/Diversity of Language Use, Dual-Language Environment and Dense Code-Switching Environment. The Dominance/Duration of Bilingualism factor contained the amount of time a person has been bilingual, as well as their relative proficiency and early-life usage of the two languages. The Intensity/Diversity of Language Use factor contained usage-related variables, such as the usage balance between the two languages. The Dual-Language Environment factor contained a measure of using both languages but in different contexts or with different speakers, while the Dense Code-Switching Environment factor included measures of different types of code-switching, different social contexts of language switching, and unintentional switching. We collected responses from a diverse sample of 231 healthy bilinguals (aged 18-30) and performed a confirmatory factor analysis with our four factors. We found that the model converged well (CFI = 0.870, TLI = 0.853, RMSEA = 0.082. SRMR = 0.125). The SEEQ therefore provides an excellent measure of bilingual language switching, capturing theoretically important aspects of language switching and serving as a tool to test effects of language switching on executive control or brain adaptations.

References:

1. Anderson, J. A. E., Mak, L., Chahi, A. K., & Bialystok, E. (2018). The language and social background questionnaire: Assessing degree of bilingualism in a diverse population. Behavior Research Methods, 50, 250-263. https://doi.org/10.3758/s13428-017-0867-9
2. DeLuca, V., Segaert, K., Mazaheri, A., & Krott, A. (2020). Understanding bilingual brain function and structure changes? U bet! A unified bilingual experience trajectory model. Journal of Neurolinguistics, 56. https://doi.org/10.1016/j.jneuroling.2020.100930 
3. Green, D. W. & Abutalebi, J. (2013). Language control in bilinguals: The adaptive control hypothesis. Journal of Cognitive Psychology, 25, 5, 515-530. https://doi.org/10.1080/20445911.2013.796377
4. Hartanto, A. & Yang, H. (2016). Disparate bilingual experiences modulate task-switching advantages: A diffusion-model analysis of the effects of the interactional context on switch costs. Cognition, 150, 10-19. https://doi.org/10.1016/j.cognition.2016.01.016
5. Hartanto, A. & Yang, H. (2020). The role of bilingual interactional contexts in predicting interindividual variability in executive functions: A latent variable analysis. Journal of Experimental Psychology: General, 149(4), 609-633. https://doi.org/10.1037/xge0000672
6. Rodriguez-Fornells, A., Krämer, U. M., Lorenzo-Seva, U., Festman, J., & Münte, T. F. (2012). Self-assessment of individual differences in language switching. Frontiers in Psychology, 2. https://doi.org/10.3389/fpsyg.2011.00388

The acquisition of gender in adolescent German learners of Spanish: Evidence from production and perception

Clara Terlaak, Sarah Schimke, Johanna Wolf

LMU Munich

C.Terlaak@lmu.de

The acquisition of L2 gender assignment and agreement is a well-known challenge. Even highly proficient L2 learners often struggle with the spontaneous production of gender agreement markings (Dewaele/Véronique 2001; Franceschina 2005; Ayoun 2018). Previous research has identified the transparency and salience of the agreement assignment and marking system in the respective L2 (Arnon/Ramscar 2012) as an important factor of influence. 

            The current study looks at German-speaking instructed adolescent learners of Spanish, and focuses on salience of different agreement targets, as well as on the relation between language perception and production. We ask: 

• whether agreement errors are more likely to disrupt processing when they are marked both on an adjective and an article, using an eyetracking-during-reading paradigm. 
• whether markings that appear not to be salient during reading are more error-prone in spontaneous written production. 
• whether learners who show more sensitivity to agreement markings during reading produce a higher number of correct markings in a gender production task. 

To address these questions, we collected written frog story retellings in a group of adolescent learners of Spanish (n = 51). In a second study targeting the same population, adolescent learners of Spanish (n = 22 so far, data collection is ongoing) read short paragraphs of texts while their eye movements were registered, and completed a written gender assignment and agreement task with lexical items different from the reading task. In the stimuli of the reading task, we manipulated whether critical noun phrases (underlined) contained an agreement error (conditions 2,4) or not (conditions 1,3), and whether agreement was marked on the article only (conditions 1,2) or also on an adjective (conditions 3,4). Note that all items contained nouns with highly regular endings, so that gender assignment should not be an issue. 

1) agreement, adjective invariable 

Allí el escritor triste caminó durante dos horas. 

2) no agreement, adjective invariable 

Allí *la escritor triste caminó durante dos horas. 

3) agreement, adjective variable 

Allí el escritor lindo caminó durante dos horas. 

4) no agreement, adjective variable 

Allí *la escritor *linda caminó durante dos horas. 

“There the_MASC/FEM sad_INVARIABLE /cute_MASC/FEM writer_MASC walked for two hours.” 

Preliminary analyses of data from twenty learners suggest that agreement errors lead to significantly longer first-pass and total reading times on the critical region, and that this effect was not modulated by the presence of agreement on the adjective. Moreover, there was a moderate positive relation between individual learner’s difference score (reading times for ungrammatical minus grammatical items) and performance in the production task. Finally, in line with the eyetracking results, a corpus analysis of the frog story data revealed particular high variability and error rates on adjectives.  

            We conclude that article-noun combinations are salient markers of agreement in L2 Spanish, and that learners who are sensitive to violations during reading also tend to be more successful in producing correct gender markings. We will discuss these results in light of different models of gender acquisition, and will also relate them to results from an ongoing parallel data collection on L2 French. 

Figures

Figure 1: First-pass and total reading times in four experimental conditions. 

Note: Left-side bars in each diagram represent conditions with agreement, right-side bars represent conditions without agreement. Conditions with invariable adjectives are marked in red, and those with variable adjectives in blue.   

Figure 2: Relation between difference score in total reading times and performance in a gender production task.

References:

1. Arnon I. & Ramscar, M. (2012). Granularity and the acquisition of grammatical gender: How order-of-acquisition affects what gets learned. Proceedings of the Annual Meeting of the Cognitive Science Society, 31(31), 2112–2117. 
2. Ayoun, D. (2018). Grammatical gender assignment in French: Dispelling the native speaker myth. Journal of French Language Studies, 28(1), 113-148. 
3. Dewaele, J.-M. & Véronique, D. (2001). Gender assignment and gender agreement in advanced French interlanguage: a cross-sectional study. Bilingualism: Language and Cognition, 4(3), 275-297.  
4. Franceschina, F. (2005). Fossilized Second Language Grammars. The Acquisition of Grammatical Gender. Benjamins [Language Acquisition and Language Disorders, 38]. 

The effect of L1 on the acquisition of definiteness in L2: Evidence from L1-English and L1-Russian bilinguals acquiring L2-Hebrew

Dana Plaut-Forckosh¹, Beradze Marianna¹, Iris Hindi¹, Natalia Meir^1,2

¹The Department of English Literature and Linguistics, Bar-Ilan University, Ramat Gan, Israe.

² The Gonda Multidisciplinary Brain Research Center, Bar-Ilan University, Ramat Gan, Israel

Plautdana@gmail.com

beradzem@gmail.com

irishindi@gmail.com

natalia.meir@gmail.com

Definite use can serve various functions such as anaphoric (maintaining a previously mentioned referent), bridging (assuming part-whole relationships in given information) (Ariel, 1990), or context-dependent (pragmatic enrichment of new information using definite nouns under specific restrictions) (Roberts, 2003). Previous studies show cross-linguistic effects, i.e., effects of L1 properties, for the acquisition of articles in L2-Enlgish (see Ionin et al., 2022; Zdorenko & Paradis, 2008, 2012). The present study investigates the impact of L1 properties on the acquisition of definiteness marking in the second language (L2) by looking into L2-Hebrew of bilingual children with L1-Enlgish and L1-Russian as compared to Hebrew-speaking monolingual child and adult controls.

            The three languages (Hebrew, Russian, and English) vary with respect to definiteness marking. Hebrew marks only definite nouns using the article ha– and the object marker et that precedes only definite objects (Danon, 2001). English has indefinite articles for count nouns a/an, and the definite article the marks count, mass and plural nouns (e.g., Schaeffer & Matthewson, 2005). Russian has no articles in its grammatical system (e.g., Nichols, 1988). Previous studies on definiteness marking in Hebrew show that monolingual children exhibit adultlike production by age 3;6 (Uziel-Karl, 2015; Zur, 1983). English-Hebrew bilingual children were shown to have relatively few article omissions in Hebrew (Altman et al., 2016). For Russian-Hebrew bilinguals the results are mixed; while some studies report omissions of definite articles (e.g., Meir et al., 2017), others report similar proportion of (in)definite use in narratives in typically developing monolinguals and bilinguals (e.g., Fichman et al., 2022).

            Using the “Baby bird” narrative in Hebrew of the Multilingual Assessment Instrument for Narratives (MAIN) (Gagarina et al., 2012), definiteness use was evaluated in four groups: (1) Monolingual Hebrew-speaking adults (N=20), (2) Monolingual Hebrew-speaking children (N=20, ages 5-8), (3) English-Hebrew-speaking bilinguals (N=21, ages 5-10), and (4) Russian-Hebrew-speaking children (N=27, ages 5-8). Narratives were transcribed using the CHILDES project (MacWhinney & Snow, 1990). Coding scheme included Definiteness (+/-), Type of definiteness use (Anaphoric, Bridging, or context), and target use (1/0) (see Table 1).

            Using mixed-effect modelling with Participant as a random effect, the findings show that monolingual children were adult-like (see Fig 1-3), while both bilingual groups were less accurate than adult controls. Furthermore, pairwise comparisons showed that bilinguals with L1-English were on par with monolingual Hebrew-speaking children, while bilinguals with L1-Russian lagged behind. Further error pattern analysis showed that Russian-Hebrew bilinguals were more likely to omit definite article compared to their monolingual Hebrew- and English-Hebrew-speaking peers, specifically in anaphoric environments, corroborating previous findings (e.g., Meir et al., 2017).

            To conclude, the present study corroborates findings of the adultlike production of (in)definiteness system by monolingual Hebrew-speaking children (Uziel-Karl, 2015; Zur, 1983). Furthermore, the study provides evidence for the effects of cross-linguistic influence in L2 article acquisition, showing that L2-Hebrew definiteness is shaped by L1 properties, in line with previous studies for L2-English acquisition (e.g., Ionin et al., 2022; Zdorenko & Paradis, 2008, 2012).

Fig 3. Non-target use per Group.

References:

1. Ariel, M. (1990). Accessing Noun-Phrase Antecedents. London: Routledge.
2. Altman, C., Armon-Lotem, S., Fichman, S., & Walters, J. (2016). Macrostructure, microstructure, and mental state terms in the narratives of English–Hebrew bilingual preschool children with and without specific language impairment. Applied Psycholinguistics, 37(1), 165-193 https://doi.org/10.1017/S0142716415000466
3. Fichman, S., & Altman, C. (2019). Referential cohesion in the narratives of bilingual and monolingual children with typically developing language and with specific language impairment. Journal of Speech, Language, and Hearing Research, 62(1), 123-142.‏ https://doi.org/10.1044/2018_JSLHR-L-18-0054
4. Fichman, S., Walters, J., Melamed, R., & Altman, C. (2022). Reference to characters in narratives of Russian-Hebrew bilingual and Russian and Hebrew monolingual children with Developmental Language Disorder and typical language development. First Language, 42(2), 263-291.‏ https://doi.org/10.1177/0142723720962938
5. Danon, G. (2001). Syntactic definiteness in the grammar of Modern Hebrew. Linguistics, 39(6), 1071-1116.
6. Gagarina, N., Klop, D., Kunnari, S., Tantele, K., Välimaa, T., Balčiūnienė, I., Bohnacker, U., & Walters, J. (2012). MAIN: Multilingual Assessment Instrument for Narratives. ZAS Papers in Linguistics, 56, 1–140. https://doi.org/10.21248/zaspil.56.2019.414
7. Ionin, T., Choi, S. H., & Liu, Q. (2022). What transfers (or doesn’t) in the second language acquisition of English articles by learners from article-less native languages?. Linguistic Approaches to Bilingualism, 12(2), 133-162.‏
8. Meir, N., Walters, J., & Armon-Lotem, S. (2017). Bi-directional cross-linguistic influence in bilingual Russian-Hebrew children. Linguistic Approaches to Bilingualism, 7(5), 514-553.‏ https://doi.org/10.1075/lab.15007.mei
9. Nichols, J. (1988). Russian prose. Clause combining in grammar and discourse, 18, 399.‏
10. Roberts, C. (2003). Uniqueness in definite noun phrases. Linguistics and philosophy, 26, 287-350.‏
11. Schaeffer, J., & Matthewson, L. (2005). Grammar and pragmatics in the acquisition of article systems. Natural Language & Linguistic Theory, 23(1), 53-101.‏
12. Uziel-Karl, S. (2015). The development of differential object marking in child Hebrew. Revue Roumaine Linguistique, 4, 339-350.‏‏ http://dspace.bcu-iasi.ro/handle/123456789/15866
13. Zur, B. (1983). On the acquisition of definiteness by Hebrew speaking children [Master’s thesis]. Tel Aviv University. (Hebrew).
14. Zdorenko, T., & Paradis, J. (2008). The acquisition of articles in child second language English: fluctuation, transfer or both?. Second language research, 24(2), 227-250. ‏ https://doi.org/10.1177/0267658307086302
15. Zdorenko, T., & Paradis, J. (2012). Articles in child L2 English:When L1 and L2 acquisition meet at the interface. First language, 32(1-2), 38-62.‏ https://doi.org/10.1177/0142723710396797  

The Effect of Metalinguistic Awareness on Theory of Mind – preliminary findings

Shijun Yu (Iris), Dr. Andrea Krott, Dr. Ian Apperly

University of Birmingham, UK

sxy186@student.bham.ac.uk

Objective. Bilingual children have precocious Theory of Mind (ToM) skills, that is skills to distinguish between one’s own and other’s knowledge, desires and intentions (Schroeder, 2018; Diaz & Farrar, 2018). It has been argued that this is due to enhanced metalinguistic awareness, that is bilinguals’ experience to link two representations for the same referent (two names for the same object). This study examined the effect of metalinguistic awareness training on ToM.

            Methods. We tested 32 participants (target: 34) three- to four-year-old monolingual children in the United Kingdom. They took part in a pre-test, a training period of eight sessions (twice per week for 4 weeks) and a post-test. For the pre- and post-test, children were tested on three ToM tasks (Director task, Sally Anne, Unexpected content), an executive function (EF) task (Blue dog/red dog), vocabulary (British Picture Vocabulary Scale), and number knowledge (Numeral identification). For the training sessions, children were randomly assigned to either metalinguistic training (experimental group) or numerical training (active control group). For the metalinguistic training, we used the Synonym Judgement task, which had previously been found to be related to enhanced ToM skills (Diaz & Farrar, 2018; Fan et al., 2015; Hsin & Snow, 2017). For the numerical training, we used “the Great Race” task (Ramani & Siegler, 2011). We expected metalinguistic awareness training to improve ToM, with EF and vocabulary knowledge as potential moderators.

            Results. Tables 1 show preliminary results. A two-way ANOVA was carried out on the Director Task, the Sally Anne, and the Unexpected Content task with group as between-subject factor and timepoint as within-subject factor. For the Sally Anne task we found a significant effect of group (F (1) = 9.54 p = .003), with children in the metalinguistic training group scoring higher at both time points. For the Director task, we found a significant effect of time (F (1) = 9.99, p = .003), with lower scores post training compared to pre training. For the Unexpected content task, we found neither an effect of group (F (1) = 2.66, p = .11) nor time (F (1) = 0.001, p = .97). In addition, there was no significant interaction between groups and time on any of the tasks.

            Conclusion. These preliminary results suggest that metalinguistic awareness training does not lead to increased ToM skills. But our training might not have been extensive enough to lead to changes, especially when compared to the extensive experience of a bilingual child.

            Further analyses will investigate whether children improve on the training task across sessions and whether improvement on the training task is related to changes between pre- and post-test. We will also test whether children who followed the numerical trained exhibit improvements in numeral knowledge. Furthermore, we will investigate whether inhibitory or vocabulary skills pre-intervention might interact with the effect of the intervention, and whether improvements in ToM (or numeral knowledge) might be related to improvements in inhibitory or vocabulary skills.

Keywords: ToM, Metalinguistic Awareness, Executive Function, Preschool children, Monolinguals

Table 1. Descriptive Statistics for Metalinguistic and Numerical Training Group at Pre- and Post-test.

References:

1. Diaz, V., & Farrar, M. J. (2018). The missing explanation of the false-belief advantage in bilingual children: a longitudinal study. Dev Sci, 21(4), e12594. https://doi.org/10.1111/desc.12594
2. Fan, S. P., Liberman, Z., Keysar, B., & Kinzler, K. D. (2015). The Exposure Advantage: Early Exposure to a Multilingual Environment Promotes Effective Communication. Psychol Sci, 26(7), 1090-1097. https://doi.org/10.1177/0956797615574699
3. Hsin, L., & Snow, C. (2017). Social perspective taking: a benefit of bilingualism in academic writing. Reading and Writing, 30(6), 1193-1214. https://doi.org/10.1007/s11145-016-9718-9
4. Ramani, G. B., & Siegler, R. S. (2011). Reducing the gap in numerical knowledge between low- and middle-income preschoolers. Journal of Applied Developmental Psychology, 32(3), 146-159. https://doi.org/10.1016/j.appdev.2011.02.005
5. Schroeder, S. R. (2018). Do Bilinguals Have an Advantage in Theory of Mind? A Meta-Analysis. Frontiers in Communication, 3. https://doi.org/10.3389/fcomm.2018.00036

The Multilingual Picture Database

Dr Christos Pliatsikas*

School of Psychology and Clinical Language Sciences

University of Reading

c.pliatsikas@reading.ac.uk

* Co-authors:

Jon Andoni Duñabeitia^1,2, Ana Baciero³, Kyriakos Antoniou^{4, 5}, Mark Antoniou⁶, Esra Ataman⁷, Cristina Baus⁸, Michal Ben-Shachar⁹, Ozan Can Çağlar¹⁰, Jan Chromý¹¹, Montserrat Comesaña¹², Maroš Filip¹³, Dušica Filipović Đurđević¹⁴, Margaret Gillon Dowens¹⁵, Anna Hatzidaki¹⁶, Jiří Januška¹⁷, Zuraini Jusoh¹⁸, Rama Kanj¹⁹, Say Young Kim²⁰, Bilal Kırkıcı¹⁰, Alina Leminen^21,22, Terje Lohndal^23,24, Ngee Thai Yap²⁵, Hanna Renvall^26,27, Jason Rothman^1,24, Phaedra Royle²⁸, Mikel Santesteban²⁹, Yamila Sevilla³⁰, Natalia Slioussar^31,32, Awel Vaughan-Evans³³, Zofia Wodniecka³⁴, Stefanie Wulff^35,24

1. Centro de Investigación Nebrija en Cognición (CINC), Universidad Nebrija, Madrid, Spain.

2. UiT The Arctic University of Norway; Tromsø, Norway.

3. Bournemouth University; Poole, United Kingdom.

4. Department of Rehabilitation Sciences, Cyprus University of Technology, Limassol, Cyprus

5. Hellenic Open University, Greece

6. The MARCS Institute for Brain, Behaviour and Development, Western Sydney University, Penrith, NSW, Australia

7. School of Psychological Sciences and Centre for Reading, Macquarie University, Sydney, Australia

8. Department of Cognition, Development and Educational Psychology, Universitat de Barcelona, Barcelona, Spain

9. Department of English Literature and Linguistics and The Gonda Multidisciplinary Brain Research Center, Bar-Ilan University, Ramat-Gan, Israel

10. Department of Foreign Language Education, Middle East Technical University, Ankara, Turkey

11. Institute of Czech Language and Theory of Communication, Faculty of Arts, Charles University, Praha, Czech Republic

12. Research Unit in Human Cognition, CIPsi, School of Psychology, University of Minho, Braga

13. Department of Linguistics, Faculty of Arts, Charles University, Praha, Czech Republic

14. Laboratory for Experimental Psychology, and Department of Psychology, Faculty of Philosophy, University of Belgrade, Beograd, Serbia

15. School of Education and English, University of Nottingham Ningbo China, Ningbo, China

16. School of Philosophy, Department of English Language and Literature, National and Kapodistrian University of Athens, Athens, Greece

17. Department of Central European Studies, Faculty of Arts, Charles University, Praha, Czech Republic

18. Malay Language Department, Faculty of Modern Languages and Communication, Universiti Putra Malaysia, Serdang, Malaysia

20. Department of English Language and Literature, and Hanyang Institute for Phonetics and Cognitive Sciences of Language, Hanyang University, Seoul, Republic of Korea

21. C-unit, Laurea University of Applied Sciences, Vantaa, Finland

22. Cognitive Brain Research Unit, Department of Psychology and Logopedics, University of Helsinki, Helsinki, Finland

23. Department of Language and Literature, NTNU Norwegian University of Science and Technology, Trondheim, Norway

24. AcqVA Aurora Center, Institute of Language and Culture, UiT the Arctic University of Norway

25. Department of English, Faculty of Modern Languages and Communication, Universiti Putra Malaysia, Serdang, Selangor, Malaysia

26. Department of Neuroscience and Biomedical Engineering, Aalto University, Espoo, Finland

27. BioMag Laboratory, HUS Diagnostic Center, Helsinki University Hospital, University of Helsinki and Aalto University, Helsinki, Finland

28. School of Speech-language Pathology and Audiology, University of Montreal, Centre for Research on Brain, Language and Music (CRBLM), and Interdisciplinary Centre for Brain and Learning Research (​​CIRCA), Montréal (Québec), Canada

29. The Bilingual Mind Research Group, Department of Linguistics and Basque Studies, University of the Basque Country UPV/EHU, Vitoria-Gasteiz, Spain

30. Instituto de Lingüística, Facultad de Filosofía y Letras, Universidad de Buenos Aires, Consejo Nacional de Investigaciones Científicas y Técnicas (Conicet), Buenos Aires, Argentina

31. School of Linguistics, Higher School of Economics, Moscow, Russia

32. Saint Petersburg State University, Russia

33. School of Human and Behavioural Sciences, Prifysgol Bangor, Bangor, Wales

34. Institute of Psychology, Jagiellonian University, Krakow, Poland

35. University of Florida, United States

The growing interdisciplinary research field of psycholinguistics is in constant need of new and up-to-date tools which will allow researchers to answer complex questions, but also expand on languages other than English, which dominates the field. One type of such tools are picture datasets which provide naming norms for everyday objects. However, existing databases tend to be small in terms of the number of items they include, and have also been normed in a limited number of languages, despite the recent boom in multilingualism research. In this paper we present the Multilingual Picture (Multipic) database, containing naming norms and familiarity scores for 500 coloured pictures, in thirty-three languages or language varieties from around the world (Table 1). The data was validated with standard methods that have been used for existing picture datasets. This is the first dataset to provide naming norms, and translation equivalents, for such a variety of languages; as such, it will be of particular value to psycholinguists and other interested researchers. The dataset has been made freely available. We welcome expressions of interest to norm Multipic in more languages and language varieties.

The presence of orthography enhances regressive crosslinguistic influence in Spanish-Basque-English trilinguals

Antje Stoehr¹, Christoforos Souganidis², Trisha Thomas^1,3, Jessi Jacobsen¹, Clara D. Martin^1,4

¹Basque Center on Cognition, Brain and Language, ²HiTZ Basque Center for Language Technology – Aholab Signal Processing Laboratory, University of the Basque Country UPV/EHU, ³University of Michigan, ⁴Ikerbasque Basque Foundation for Science

a.stoehr@bcbl.eu

Orthography is omnipresent in instructed language learning, and its effects on L2 speech production receive growing attention. Orthography facilitates word learning and phonological accuracy but hinders phonetic accuracy when L1 and L2 sound-letter mappings mismatch (for a recent review: Hayes-Harb & Barrios, 2021). No research has yet investigated orthographic influence in trilingual speech production, which is an important topic considering that virtually all adult participants in research studies are literate and that the majority of high school students in the European Union speaks at least two non-native languages (Eurostat, 2021). However, it is established that trilinguals’ languages interact progressively (L1/L2→L3) or regressively (L3→L1/L2). In regressive crosslinguistic influence (rCLI), L3→L2 influence is stronger than L3→L1 influence (Cabrelli Amaro, 2017a, 2017b). In the present study, we combined the research fields on orthographic effects on speech production and rCLI by testing whether the speech elicitation method affected rCLI in 55 Spanish-Basque-English trilinguals. Participants were young adults (M_age=25.15 years) who acquired Spanish from birth, Basque in early childhood (M_AoABasque=2.8 years) and English as a third language slightly later (M_AoAEnglish=5.8 years). We measured voice onset time (VOT) production elicited in two tasks: picture naming and reading aloud. Spanish and Basque have similar VOT (prevoiced voiced plosives; short lag voiceless plosives), while English has longer VOT (short lag voiced plosives; aspirated voiceless plosives). Importantly, Spanish, Basque, and English plosives map onto the same letters. If the presence of orthography modulates rCLI from English, it would result in longer VOT in L2-Basque in the reading aloud task than in the picture naming task. We analyzed all data in mixed-effects models in RStudio (RStudio Team, 2022). As expected, in the picture naming task, participants produced prevoiced voiced plosives in Spanish (94% prevoiced) and Basque (94% prevoiced) and fewer prevoiced voiced plosives in English (73% prevoiced); and voiceless plosives with short positive VOT in Spanish (29 ms) and Basque (30 ms) and significantly longer VOT in English (42 ms). In the reading aloud task, VOT of voiceless plosives increased in all languages compared to the picture naming task (Spanish=30 ms; Basque=31 ms; English=45 ms), showing a language-independent task effect. Furthermore, in the reading aloud task, participants produced fewer L2-Basque voiced plosives with prevoicing (91%) than in the picture naming task, while no between-task differences were observed for L1-Spanish or L3-English. This means that L2-Basque voiced plosives were produced less native-like in the reading aloud task than in the picture naming task. These results suggest that the online influence of orthography increases rCLI from L3-English to L2-Basque, but that rCLI appears limited to less stable phonetic categories, as no rCLI was observed in voiceless plosives. We argue that voiced plosives are relatively less stable than voiceless plosives in Spanish and Basque due to their allophonic variation and later acquisition age (Hualde, 1991; Macken & Barton, 1980). We discuss the methodological implications of these findings and how they inform two research fields: 1) orthographic effects in L2/L3 production and 2) multilingual crosslinguistic interactions.

References:

1. Cabrelli Amaro, J. (2017a). L3 morphosyntactic effects on L1 vs. L2 systems: The differential stability hypothesis. In T. Angelovska & A. Hahn (Eds.), L3 syntactic transfer (pp. 173–194). John Benjamins.
2. Cabrelli Amaro, J. (2017b). Testing the Phonological Permeability Hypothesis: L3 phonological effects on L1 versus L2 systems. International Journal of Bilingualism, 21,698−717.
3. Eurostat. (2021). Foreign language learning statistics [Data file]. Retrieved from https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Foreign_language_learning_statistics&oldid=542152
4. Hayes-Harb, R., & Barrios, S. (2021). The influence of orthography in second language phonological acquisition. Language Teaching, 54, 297−326.
5. Hualde, J. I. (1991). Basque phonology. Routledge.
6. Macken, M. A., & Barton, D. (1980). The acquisition of the voicing contrast in Spanish: a phonetic and phonological study of word-initial stop consonants. Journal of Child Language, 7, 433−458.
7. RStudio Team. (2022). RStudio: Integrated development environment for R. RStudio (Version IDE 2022.02.2+485) [Computer software]. PBC. http://rstudio.com

The Processing of Subject Pronouns in L2 English: An Online Visual World Eye-tracking Study

Linghui Diao (University of York), Leah Roberts (University of York)

linghui.diao@york.ac.uk

Pronoun comprehension can be challenging for second language (L2) speakers, especially in contexts where the referent for the pronoun is ambiguous, for instance, where a pronoun can grammatically refer to either of two potential antecedents in the preceding discourse (Roberts, Gullberg & Indefrey, 2008). Most L2 acquisition studies have focused on speakers of a non-pro-drop language, such as English, with obligatory overt pronouns, who are learning a pro-drop system, with null pronouns (e.g., Chinese). In languages that allow pronoun omission, explicit pronouns often signal a shift in discussion topic (Shibata & Yashima, 2014), a feature not typical in non-pro-drop languages. Yet, studies investigating pronoun resolution with L2 speakers from a pro-drop first language (L1) background acquiring the assumed less complex non-pro-drop system are limited and yield mixed results. Some studies find that L2 learners employ native-like strategies effectively in straightforward contexts (e.g., “John met Paul while he was in high school”), but struggle in more complex scenarios where equally prominent antecedents are presented through conjunctions (e.g., “Carlos and Martin are at the office. While Carlos is working, he is eating lunch”), leading to variability in pronoun resolution (Contemori, Asiri & Perea Irigoyen, 2019; Contemori & Dussias, 2020; Cunnings, Fotiadou & Tsimpli, 2017; Roberts et al., 2008). This study focuses on 50 native Mandarin Chinese speakers, who are L1-pro-drop users, and compares them with 50 native English speakers using distinct methods: the eye-tracking during a visual world paradigm tapping into real-time processing and implicit knowledge, and a picture selection task to assess non-real-time comprehension and the explicit knowledge of pronouns. Participants first listen to 24 sets of critical sentences while viewing corresponding displays where two possible antecedents of pronouns and a distracted item/launch pad are introduced, and following each item, they are prompted to answer follow-up questions by selecting pictures of referents on the screen. 72 filler sentences are included, with half of these followed by yes/no questions. The research also explores to what extent the order of mention and the dynamics of topic continuity or topic shift impacts pronoun resolution (sample materials provided below).

Stimuli Samples

(1) Mother Sentence: The doctor (female) and the president (female) are in the office.

a. First Mention: While the doctor is working on the report, she is listening to a song.

b. Second Mention: While the president is working on the report, she is listening to a song.

(2) Mother Sentence: The doctor (male) and the president (female) are in the office.

a. Topic continuation: While the doctor is working on the report, he is listening to a song.

b. Topic shift: While the doctor is working on the report, she is listening to a song.

The role of individual differences in the online processing of L2 English articles by L1 Spanish learners and L1 English controls

Jelena O’Reilly (University of York), Verónica García-Castro (University of Costa Rica)

jelena.oreilly@york.ac.uk

Although English articles (a[an]/the) are arguably one of the hardest features to acquire for second language (L2) learners, research suggests that learners from first language (L1) backgrounds that also grammaticalise articles, can be highly accurate at article suppliance even at lower proficiencies (Ionin et al., 2008). However, apart from the L1, not much is known about which other individual characteristics facilitate L2 English article acquisition. Studies in other areas of SLA suggest that gains in the declarative memory subserve semantic learning (e.g., the mental lexicon) while gains in the procedural memory aid syntactic learning (Morgan-Short, 2014), but the role of the declarative/procedural memory has never been studied in relation to L2 English article acquisition.

Here we present preliminary findings of a larger study (pre-registration) with 70 L1 Spanish L2 English learners (whose L1 has articles) and 40 L1 English controls to contribute to our understanding of (1) whether L1 Spanish L2 English learners benefit form L1 transfer and are sensitive to online article violations already at intermediate proficiency, and (2) to what extent the declarative/procedural memory systems facilitate this online processing of L2 English articles. The explicit and implicit knowledge of English articles (both omission: I saw cat., and substitution: I saw a/the cat.) was assessed using acceptability judgement (AJ) and self-paced reading (SPR) tasks. For the SPR and AJ tasks, we developed 48 scenarios (24 indefinite and 24 definite) with the critical item consisting of singular noun in anaphoric direct object position. The individual differences in declarative and procedural memory were assessed via yes/no vocabulary size test (Meara & Miralpeix, 2017), and Tower of Hanoi respectively.

The preliminary data for 28 L1 Spanish L2 English participants have been analysed using mixed-effects regression with by-participant and -item random slopes for within- measures and intercepts for between- measures (Barr, 2013).

References:

1. Barr, D. J. (2013). Random effects structure for testing interactions in linear mixed-effects models. Frontiers in Psychology, 4(1). https://doi.org/10.3389/FPSYG.2013.00328
2. Ionin, T., Zubizarreta, M. L., & Maldonado, S. B. (2008). Sources of linguistic knowledge in the second language acquisition of English articles. Lingua, 118(4), 554–576. https://doi.org/10.1016/j.lingua.2006.11.012
3. Meara, P. M., & Miralpeix, I. (2017). Tools for researching vocabulary. Multilingual Matters.
4. Morgan-Short, K. (2014). Electrophysiological approaches to understanding second language acquisition: A field reaching its potential. Annual Review of Applied Linguistics, 34, 15–36. https://doi.org/10.1017/S026719051400004X

The role of lexical overlap and cognate facilitation in initial child foreign language speech segmentation

Katie Von Holzen, Marie Schnieders, Sophia Wulfert, Holger Hopp

English Linguistics, Technical University of Braunschweig

k.von-holzen@tu-braunschweig.de

Adults initially struggle to segment word forms from continuous speech in a foreign language (FL) before they have systematic exposure to it, but do better when words partially overlap in form and meaning with their L1 equivalents (i.e. cognates: English: /bʌtər/; German: /bʊtɐ/; noncognates: English: /mɪrər/; German: /ʃpi:gl/; Shoemaker & Rast, 2013). The listener already has a representation for the sequence of sounds from their L1, which may facilitate segmentation and recognition of the form-similar FL word. Lexical access improves with age (e.g. Kim & Na, 2008), making it more likely that adults can take advantage of their existing representations to segment foreign speech. To assess the extent to which developing L1 knowledge affects segmentation, we examine whether L1 lexical overlap facilitates the segmentation of English speech among German 6- to 9-year-old primary-school students before they receive instruction in English. 

            In a word recognition study, English-German cognate and noncognate word pairs (n = 160) were embedded into an otherwise identical English sentence frame (You pushed some butter/mirror mustom to the exit). 113 German 6- to 9-year-olds listened to the sentence followed by an isolated probe-word that either did (target) or did not (lure) appear in the sentence. Participants indicated via button press whether they had heard the probe word in the sentence (word acceptance). 

            A general-linear mixed model revealed a main effect of word status (target vs. lure; β = 1.18, SE = 0.04, p < .001) but no main effect of or interaction with cognate status (p’s > .05). Word acceptance was higher for target compared with lure probe words (Figure 1) but this pattern did not differ between cognate and noncognate words. Word acceptance for both targets (EMM = -0.27, SE = 0.12, p = .02) and lures was significantly below chance (EMM = -1.46, SE = 0.12, p < .0001). 

            Although word acceptance was higher for targets compared with lures, this difference was driven by high performance in rejecting the lures as words, suggesting that memory traces of the strings are sufficient to recognize the mismatch with lure words, but not to recognize the match with the target words. This interpretation is supported by the lack of facilitated recognition for cognate words, suggesting that listeners did not activate L1 lexical representations purely on the basis of phonological overlap.  

            In a follow-up experiment we will examine whether first activating the L1 German lexical entry will strengthen the representation for segmented cognate words and facilitate recognition. Participants will first preview a pair of pictures before completing the segmentation task. The pictures will depict the cognate and noncognate targets in the paired English sentence frames. We expect that participants will implicitly activate the label for these objects in German (Meyer et al., 2007). If we find improved recognition of target cognate words, relative to chance and non-cognate targets, it may indicate that early FL learners rely on L1 lexical entries in initial FL segmentation, but only when these entries have been preactivated, thus narrowing the candidate set of L1 words.

Keywords: foreign language learning, speech segmentation, lexical overlap 

Figure 1. Word acceptance (%) for cognate and non-cognate target and lure words.

References:

1. Kim, H., & Na, D. L. (2008). A normative study of the Boston naming test in 3- to 14-year-old Korean children. Clinical Neuropsychologist, 22(1), 84–97. https://doi.org/10.1080/13854040601064526 
2. Meyer, A. S., Belke, E., Telling, A. L., & Humphreys, G. W. (2007). Early activation of object names in visual search. Psychonomic Bulletin & Review, 14(4), 710–716. https://doi.org/10.3758/BF03196826 
3. Shoemaker, E., & Rast, R. (2013). Extracting words from the speech stream at first exposure. Second Language Research, 29(2), 165–183. https://doi.org/10.1177/0267658313479360 

Verb aspect processing in monolingual and bilingual heritage speakers of Turkish

Onur Özsoy, Nisa Büyükyıldırım, Özce Özceçelik

Leibniz-Center General Linguistics (ZAS)

oezsoy@leibniz-zas.de

Introduction. We explore the incremental processing of grammatical aspect in Turkish among monolinguals and bilingual heritage speakers, conceptually replicating recent Visual World Paradigm (VWP) eye-tracking studies that demonstrated incremental sub-word level processing in monolingual English, Russian and Spanish (Minor et al., 2022, 2023). In addition, we ask whether aspectual distinctions will be recognized by heritage speakers of Turkish who have shown reduced sensitivity to TAM-morphology in previous studies (Arslan et al., 2017; Coşkun Kunduz, 2018). We also included language proficiency (C-tests and self-rating assessment) and processing speed (WAIS-IV subtest) to account for the expected individual variation.

Hypotheses. We expect a substantial difference in how aspect is processed in Turkish. Participants will favor perfective aspect for completed events whereas imperfective aspect will be preferred for ongoing events. Additionally, we anticipate that Turkish heritage speakers will show a reduced effect in comparison to monolinguals. We also predict that processing speed and proficiency levels will be able to predict how participants process and comprehend aspectual information.

Methods. We conducted a picture selection task and VWP eye-tracking using the same design as Minor et al. (2022, 2023). Twenty-five bilingual heritage speakers and 28 monolingual speakers saw images of completed and ongoing events side by side. There were 24 critical stimuli sentences in two conditions (Perfective and Imperfective), and 20 unrelated fillers. After each item, participants were asked to select the image that corresponded to the sentence.

Results. Monolinguals achieved 95.5% mean accuracy (SD=1.8%) and heritage speakers 89.9% (SD=3.2%). Analysis using binomial Generalized Linear Mixed-Effects models (bGLMM) to predict accuracy found significant effects of Aspect β=-2.1 (SE=0.7, p<0.01) and Proficiency β=0.1 (SE=0.05, p<0.05). This indicates that participants’ accuracy drops in the perfective condition compared to the imperfective condition. Cluster-based permutation analysis confirmed incremental processing of aspectual information (Figure 1), but it also points to a reduced effect compared to monolinguals. We ran an additional bGLMM which shows that participants with higher processing speed and proficiency levels were more sensitive to aspectual differences.

Discussion. Monolingual and heritage speakers process and comprehend overt marking in the imperfective better than the default unmarked perfective. This study broadens our understanding of incremental processing in heritage speakers who show more variance than monolinguals which can be explained by individual proficiency levels and processing speed.

Figure 1. The subfigures show the eye-tracking results for the heritage speaker group as looks to the Target and Competitor images under the Perfective and Imperfective conditions. The grey shading highlights significant clusters of time bins. The red line represents looks to the Target, and the blue line represents looks to the Competitor. The dashed vertical line indicates the mean verb offsets.

References:

1. Arslan, S., De Kok, D., & Bastiaanse, R. (2017). Processing grammatical evidentiality and    time reference in Turkish heritage and monolingual speakers. Bilingualism: Language and Cognition, 20(3), 457–472.
2. Coşkun Kunduz, A. (2018). The second language acquisition of inflectional morphology in Turkish. MA thesis, Boğaziçi University.
3. Minor, S., Mitrofanova, N., Guajardo, G., Vos, M., & Ramchand, G. (2023). Aspect processing across languages: A visual world eye-tracking study. Frontiers in Language Sciences, 1, 1052205.
4. Minor, S., Mitrofanova, N., & Ramchand, G. (2022). Fine-grained time course of verb aspect processing. Plos one, 17(2), e0264132.

Verb-bias effects in L1-to-L2 and L2-to-L1 cross-linguistic structural priming

Chantal van Dijk (Technische Universität Braunschweig & University of Stuttgart), Holger Hopp (Technische Universität Braunschweig)

c.vandijk@tu-braunschweig.de

 h.hopp@tu-braunschweig.de

Introduction. Structural priming, the tendency of speakers to repeat a structure they just heard (e.g., Bock, 1986) has been found to be long-lasting, indicative of learning (e.g., Jaeger & Snider, 2013). A key piece of evidence of learning is that less preferred structures prime more strongly and therefore yield more learning (i.e. inverse-preference effects; Ferreira & Bock, 2006). Whilst there is some evidence that structural priming between languages can be longer-lasting as well (Ahn & Ferreira, 2023; Hwang & Shin, 2019) it is unclear to what extent structural preferences affect cross-linguistic priming and whether their effects are different in L1-to-L2 priming compared to L2-to-L1 priming. In this study, we investigated and compared L1-to-L2 and L2-to-L1 priming of ditransitive structures in German and in English native speakers in a German-to-English cross-linguistic priming task. Both languages allow for prepositional object (PO; the clown mails a postcard to the pirate) and double object structures (DO; the clown mails the pirate a postcard). To test for preference effects on priming, we manipulated the structural biases of the German prime verbs. To test for learning we investigated cumulative priming effects over the course of the experiment and changes in participants’ production of ditransitive structures in a post-test compared to a baseline task. We asked the following research questions: 

RQ1) Do prime verb biases modulate immediate L1-to-L2 and L2-to-L1 structural priming? 

RQ2) Does L1-to-L2 and L2-to-L1 priming of ditransitive structures result in learning? 

            Method. Ninety-two L1-German-L2-English adults and 89 L1-English-L2-German adults participated. In a German-to-English priming task, a spoken DO or PO prime preceded each picture description. There were 4 English target verbs, 4 strongly-DO-biased and 4 weakly-DO-biased German prime verbs which were PO-biased in their English translation (Şafak & Hopp, 2023). Baseline tasks measured PO/DO production preferences for English and German ditransitive prime and target verbs through picture descriptions. A post-test measured English DO/PO production afterwards. We predicted stronger immediate PO priming the more DO-biased German prime verbs were. Furthermore, we expected adaptation towards the less preferred PO structure over the course of the experiment, and we expected this adaptation to persist in the post-test. 

            Results. Generalized linear mixed effects models showed verb biases to modulate immediate priming effects in both the L1 German and the L1 English group. However, for the L1 German group, prime-verb biases in German predicted short-term priming, while, for the L1 English group, the prime-verbs’ translation-equivalents’ biases in English predicted short-term priming. In addition, cumulative priming held for PO structures in the L1 German group and for DO structures in the L1 English group. In sync with cumulative priming, longer-term learning from baseline to post-test was observed in terms of increased production of POs in the L1 German group and increased production of DOs in the L1 English group. (see Figure 1). 

            Discussion. Our results show that verb biases modulated immediate L1-to-L2 and L2-to-L1 priming. For both groups, what mattered were participants’ verb biases in their dominant L1, even when they were primed in their L2. Participants were found to adapt towards the structure they initially preferred less as per their L1 verb biases. Such cumulative priming carried over to an unprimed post-test, indicative of learning. Our results are largely in line with within-language priming studies (e.g., Bernolet & Hartsuiker, 2010; Kaan & Chun, 2018) and suggest that (longer-term) cross-linguistic structural priming both from L1-to-L2 and L2-to-L1 reflects an implicit error-based learning mechanism (e.g., Chang et al., 2006). Critically, even though the same priming mechanisms underlie learning and the two groups received the same input in the priming phase, the two groups adapted to opposite structures as a consequence of their different L1 biases: more POs for the L1 German group, and more DOs for the L1 English group. Hence, this study shows that different preferences push cross-linguistic priming with the same primes in opposite directions, leading to substantially different learning outcomes. 

Figure 1. Mean proportions (in numbers on the bars) and standard errors of L1-English and L1-German participants’ PO productions in the baseline (pre-test), priming phase (after PO versus DO primes) and post-test. 

References:

1. Ahn, D., & Ferreira, V. S. (2024). Shared vs separate structural representations: Evidence from cumulative cross-language structural priming. Quarterly Journal of Experimental Psychology, 77(1), 174-190. https://doi.org/10.1177/17470218231160942 
2. Bernolet, S., & Hartsuiker, R. (2010). Does verb bias modulate syntactic priming? Cognition, 114(3), 455–461. https://doi.org/ 10.1016/j.cognition.2009.11.005 
3. Bock, J. K. (1986). Syntactic persistence in language production. Cognitive Psychology, 18, 355–387. https://doi.org/10.1016/0010-0285(86)90004-6 
4. Chang, F., Dell, G. S., & Bock, K. (2006). Becoming syntactic. Psychological Review, 113(2), 234-272. https://doi.org/10.1037/0033 
5. Hwang, H. & Shin, J.-A.(2019). Cumulative effects of syntactic experience in a between- and a within-language context: Evidence for implicit learning. Journal of Memory and Language, 109, 104054, https://doi.org/10.1016/j.jml.2019.104054 
6. Jaeger, T. F. & Snider, N. E. (2013). Alignment as a consequence of expectation adaptation: Syntactic priming is affected by the prime’s prediction error given both prior and recent experience. Cognition, 127(1), 57-83. https://doi.org/10.1016/j.cognition.2012.10.013. 
7. Kaan, E., Chun, E. (2018). Priming and adaptation in native speakers and second-language learners. Bilingualism: Language and Cognition, 21(2), 228-242. https://doi.org/10.1017/S1366728916001231 
8. Şafak, D. F., & Hopp, H. (2023). Cross-linguistic differences in predicting L2 sentence structure: The use of categorical and gradient verb constraints. Studies in Second Language Acquisition, 45(5), 1234–1260. https://doi.org/10.1017/S0272263123000256 

Visual Event Representation Facilitates the Processing of Grammatical Case by Russian-German Bilingual Children

Serge Minor¹, Natalia Mitrofanova¹ & Marit Westergaard^1,2 

¹UiT The Arctic University of Norway, ²NTNU 

sergey.minor@uit.no

The development of psycholinguistic methods over the last decades has allowed for a deeper and qualitatively new understanding of language processing as it unfolds in real time. Converging evidence suggests that language processing is highly incremental and involves interconnected sub-systems or ‘modules’ (phonological, syntactic, lexical, etc.) which are activated in parallel to decode and anticipate the incoming linguistic signal (Brodbeck et al. 2022; Hagoort 2008; Huettig 2015). However, less is known about the parallel processing of linguistic and non-linguistic information, especially the interaction between non-linguistic (e.g., visual) and grammatical cues.

            In this study, we investigate how subtle changes in visual representations can affect the processing of grammatical case cues in heritage Russian by German-Russian bilingual children (N=50, 8-13 y.o.). The linguistic manipulation followed previous designs (Kamide et al. 2003) contrasting SVO sentences – with NP1 marked with nominative case and NP2 marked with accusative, and OVS sentences – where the case marking was reversed. For example, 1) ‘The hare (acc) will now find the fox (nom)’; 2) ‘The hare (nom) will now find the cabbage (acc)’.

            Unlike previous studies, we contrasted three types of visual displays: representing individual referents (potential agents/themes of the action; e.g., ‘the fox’ vs ‘the cabbage’), pairs of referents (the referent of NP1 + potential agent/theme; e.g., ‘the hare + the fox’ vs ‘the hare + the cabbage’), and events (representing interactions between the referent of NP1 and the potential agent/theme; e.g., ‘the fox finding the hare’ vs ‘the hare finding the cabbage’). We found that the participants were significantly more sensitive to the case manipulation when they viewed depictions of events compared to the other types of visual display. No significant differences in effect size were observed between the displays representing individual referents vs pairs of referents. This suggests that the participants were able to quickly integrate the thematic role information signalled by grammatical case with the visual cues in the event representations, facilitating the identification of the target picture. On the other hand, just co-presenting the pairs of referents together doesn’t lead to facilitation in grammatical case processing (as opposed to presenting them in an event representation with clear visual cues to thematic roles).

Selected references:

1. Kamide et al. (2003). Integration of Syntactic and Semantic Information in Predictive Processing: Cross-Linguistic Evidence from German and English. J.of Psycholing Res.
2. Brodbeck et al. (2022). Parallel processing in speech perception with local and global representations of linguistic context. Elife. Hagoort (2008). The fractionation of spoken language understanding by measuring electrical and magnetic brain signals. Philos Trans R Soc Lond B Biol Sci.
3. Huettig (2015). Four central questions about prediction in language processing. Brain Res. 

Well-being and motivation in later life intervention studies – a pilot study comparing language learning and a combined physical-cognitive course

Louisa Richter (University of Bamberg, University of Groningen)

Jascha Rüsseler (University of Bamberg)

Greg Poarch (University of Groningen)

Merel Keijzer (University of Groningen)

louisa.richter@uni-bamberg.de

Investigating ways of promoting healthy aging has never been more pressing than in our rapidly aging society. An emergent field is now dedicated to investigating the impact of later life language learning, due to the influence that life-long bilingualism has been shown to have on cognitive reserve (for a review see Berkes, & Bialystok, 2022). In the absence of lifelong bilingualism, novel language learning has been proposed as a particularly effective later life training to enhance cognition, as the activated brain regions overlap with areas often affected by age-related cognitive decline or indeed brain apathy (Antoniou et al. 2013; Antoniou, & Wright, 2017). To date, investigations into later life language learning and the cognitive effects to ensue from it have produced mixed results, pointing to a need for replication to increase comparability across studies (for a review see Pot et al., 2019; van der Ploeg et al., 2020).  

Later life language learning also appears to be a promising method for increasing overall well-being in elderly due to being a socially and cognitively stimulating experience. However, little is known about the modulating factors as well as the link between well-being and motivation in (language) interventions for later life learners. Alongside high self-reported well-being, the motivation of the participants of the 4-week intensive English course for beginners in the study of Pfenninger, & Polz (2018) increased continuously over the timespan of the intervention. In the study of van der Ploeg et al. (2023), motivation did not significantly change over the course of a 3-month online English language intervention for later life learners, but an increase in well-being was registered in the daily diary data. 

Though some studies have started comparing later life language learning to other interventions to enhance cognition, it remains unclear which type of intervention is most beneficial and feasible for elderly people as well as most effective in increasing well-being.  

In our pilot study, we investigated the effect of a 4-week language intervention (n=3; learning English in a non-anglophone environment) compared to a combined physical and cognitive training (n=4; playing cognitively demanding board games and movement) on cognition and well-being in elderly with varying degrees of cognitive decline. We report findings from the WHO quality of life questionnaire as well as data from a daily survey (tapping into well-being, motivation, and perceived positive impact of the course), where we find a connection between fluctuations in well-being and motivation.  

References: 

1. Antoniou, M., & Wright, S. (2017). Uncovering the mechanisms responsible for why language learning may promote healthy cognitive aging. Frontiers in Psychology, 8, 2217-2217. doi:10.3389/fpsyg.2017.02217 
2. Antoniou, M., Wong, P., & Gunasekera, G. (2013). Foreign language training as cognitive therapy for age-related cognitive decline: A hypothesis for future research. Neuroscience and Biobehavioral Reviews, 37(10), 2689-2698. doi:10.1016/j.neubiorev.2013.09.004 
3. Berkes, M., & Bialystok, E. (2022). Bilingualism as a contributor to cognitive reserve: what it can do and what it cannot do. American Journal of Alzheimer’s Disease & Other Dementias®, 37, 15333175221091417. 
4. Pfenninger, S., & Polz, S. (2018). Foreign language learning in the third age: A pilot feasibility study on cognitive, socio-affective and linguistic drivers and benefits in relation to previous bilingualism of the learner. Journal of the European Second Language Association, 2(1), 1-1. doi:10.22599/jesla.36 
5. Pot, A., Porkert, J., & Keijzer, M. (2019). The bidirectional in bilingual: Cognitive, social and linguistic effects of and on third-age language learning. Behavioral Sciences (basel, Switzerland), 9(9). doi:10.3390/bs9090098 
6. van der Ploeg, M., Keijzer, M., & Lowie, W. (2020). Methodological concerns and their solutions in third-age language learning studies. Dutch Journal of Applied Linguistics, 9(1-2), 97-108. 
7. van der Ploeg, M., Keijzer, M., & Lowie, W. (2023). Language learning, motivation, and well-being in later life. Social Sciences & Humanities Open, 8(1), 100749. 

What you hear is not what you see: irony comprehension in monolingual and multilingual children  

Kimberley Mulder, Elise van Wonderen, Britt Daize & Josje Verhagen

Amsterdam Center of Language and Communication, University of Amsterdam, Amsterdam, The Netherlands

k.mulder2@uva.nl

Research suggests that multilingual children are more sensitive than their monolingual peers to non-verbal information, such as eye gaze and tone of voice, at least in referential communication tasks (e.g., Verhagen, Grassman & Kuntay, 2017; Yow & Markman, 2011a, 2011b). Irony provides an interesting case to examine whether and to what degree comprehension of another speaker is facilitated by heightened sensitivity to non-verbal cues. In an ironic sentence, the verbal message is in contradiction with the non-verbal information (i.e., when you say “What a delicious ice cream” while looking angry or disappointed). To understand an ironic sentence, a listener must thus interpret and integrate both verbal and non-verbal information.  

            In the current study, we investigate whether multilingual children comprehend irony more readily than monolingual peers due to a potentially higher sensitivity to non-verbal information in multilingual children. We also investigate whether comprehension is modulated by familiarity with speaker behavior. 60 Monolingual Dutch and 60 multilingual (Dutch + other language(s)) children between the ages of 5 and 7 completed an irony judgement task. In this task, the children watched videos of 4 speakers producing utterances of the type ‘What a + adjective + noun’ (e.g., ‘What a beautiful car!’) in either an ironic or non-ironic manner and indicated for each video whether the speaker was being ironic or not. The experiment contained 2 main blocks with 2 speakers each, and an intervening speaker familiarization block that contained only ironic (or only non-ironic) utterances by one of the speakers that would appear shortly after in the second main block. We measured accuracy and response latencies. In addition, all children performed a set of tasks measuring receptive vocabulary, basic emotion recognition, detection of real and apparent emotions, response inhibition and working memory. A caregiver questionnaire was administered to the children’s parents to measure language use and exposure. 

            In contrast to our hypothesis, preliminary results (based on 100 children) reveal that monolingual children were faster and more accurate than multilingual children in processing ironic but not non-ironic sentences. Second, previous exposure to a speaker improved the ability to read their intention, but only when earlier encounters were purely ironic. This effect did not differ between monolinguals and multilinguals. In conclusion, multilingual children’s enhanced sensitivity to non-verbal cues in referential communication tasks does not seem to extend to their understanding of ironic sentences. However, we have not yet analyzed performance on the additional tasks in relation to irony comprehension in monolinguals and multilinguals. The results from this study contribute to a better understanding of the different factors underlying irony comprehension, and whether or not exposure to multiple languages plays a role in understanding non-literal language. 

References:  

1. Verhagen, J., Grassmann, S., & Küntay, A. C. (2017). Monolingual and bilingual children’s resolution of referential conflicts: Effects of bilingualism and relative language proficiency. Cognitive Development, 41, 10-18. 
2. Yow, W. Q., & Markman, E. M. (2011a). Young bilingual children’s heightened sensitivity to referential cues. Journal of cognition and development, 12(1), 12-31. 
3. Yow, W. Q., & Markman, E. M. (2011b). Bilingualism and children’s use of paralinguistic cues to interpret emotion in speech. Bilingualism: Language and Cognition, 14(4), 562-569. 

When the dative becomes less reliable – L1 attrition in a multilingual context

Judith Schlenter¹, Marit Westergaard^1,2

¹UiT The Arctic University of Norway, ²NTNU Norwegian University of Science and Technology

judith.schlenter@gmail.com

In two experiments, we investigated the use of morphological case for argument interpretation in a group of German L1 speakers (n=34) in Norway. All expats were regularly exposed to their later-learned languages Norwegian (~43%) and English (~30%). Unlike German, both Norwegian and English lack case marking on full NPs, and word order is a reliable cue to argument interpretation. Like in German, [+definite] is realized on pre-nominal articles in English, while in Norwegian it is realized post-nominally as a suffix; see overview below.

Following models such as the Competition Model [1] or the Attrition via Acquisition model [2], cross-linguistic influence from additional languages on the L1 should occur especially in contact situations in which languages display similar, but also slightly different, properties.

Participants were presented with two event scenes, one showing an event as described by a transitive or ditransitive sentence, see Table 1, and another showing the same event, but the roles of agent and patient (transitive events) or recipient and theme (ditransitive events) were reversed. The task was to select the target scene. Picture selection was combined with the recording of eye movements (Experiment 1) or mouse movements (Experiment 2). In both experiments, we found no strong evidence of an influence from Norwegian and/or English, that is, no overreliance on word order. When presented with an OVS sentence, participants used the case cue to revise any initial expectation towards an NP1-as-agent interpretation. Experiment 1 showed that the expats used accusative case in transitive and ditransitive sentences to predictively fixate on the target scene. However, unlike a previously tested group of L1 speakers in Germany [3], they did not or to a limited extent use dative case as a predictive cue, indicating effects of attrition specific to dative morphology. For an illustration, see Figure 1.

Our finding that dative morphology is prone to L1 attrition is in line with previous research showing a specific vulnerability of the dative in (child) L2 attrition [4] as well as heritage German [e.g., 5]. Importantly, unlike in previous research, in our study this cannot be explained in terms of incomplete or differential acquisition, as our participants were adults when they moved away from a German-speaking environment.

Table 1. Example items for the transitive and ditransitive structure set: The critical time window for a prediction effect in the eye-tracking experiment is indicated in grey.

Figure 1. Fixations to the target (solid, red) and competitor scenes(dot dashed, blue) from article of object1 offset (i.e., der) for the previously tested L1 group in Germany (top) and the L1 group in Norway (bottom). Dashed vertical lines indicate the average onset of object2.

References:

1. MacWhinney, B. (2019). Language attrition and the Competition Model. In M. S. Schmid & Köpke (Eds.), The Oxford Handbook of Language Attrition (pp. 6–17). Oxford University Press.
2. Hicks, G., & Domínguez, L. (2020). A model for L1 grammatical attrition. Second Language Research, 36(2), 143–165.
3. Schlenter, J., & Westergaard, M. (to appear). What eye and mouse movements tell us about expectations towards argument order: An eye- and mouse-tracking study in German. Acta Psychologica.
4. Flores, C. (2015). Losing a language in childhood: A longitudinal case study on language attrition. Journal of Child Language, 42(3), 562–590.
5. Yager, L., Hellmold, N., Joo, H.‑A., Putnam, M. T., Rossi, E., Stafford, C., & Salmons, J. (2015). New structural patterns in moribund grammar: Case marking in heritage German. Frontiers in Psychology, 6, 1716.

Within and Across-Language Priming and the Lexical Boost Across Development in a Structurally Biased Language

Alina Kholodova, Fenia Karkaletsou, Shanley Allen

RPTU Kaiserslautern-Landau

kholo@rptu.de

Abstract priming is the tendency to reproduce previously heard structures (Bock, 1986) and is enhanced when the verb is repeated across prime and target (i.e., lexical boost; Pickering & Branigan, 1998). Further, priming can lead to structural adaptation across time (i.e., cumulative priming; Jaeger & Snider).  However, there is extremely little research on these core priming effects across development in bilingual children both within and across languages as well as conflicting research in monolingual children. Besides, we lack research on these effects in languages where the two structural alternatives behave differently in bias strength leading to prediction error for infrequent structures (Chang et al., 2000; 2006).

In the present study, we studied core priming effects across growing age (3-4, 5-6, 7-8 and adults) in German speakers (N=193), in early bilingual German L2 speakers with a different L1 (N=164) within German and, extended this study to prime English-German speaking bilinguals across the two languages. We have gathered data from English-German speaking adults (N=30) by now and are currently recruiting English-German speaking children in the same age ranges. The participants described video clips in turns with the experimenter with double object datives (Dora sent Boots the rabbit – DO) or prepositional object datives (Dora sent the rabbit to Boots – PO) either in German or from English to German. We manipulated verb condition by presenting primes and targets either in the same (SV) or different verb (DV) condition to test for the lexical boost effect (within language)and the translation equivalent boost (across-languages). Crucially, in contrast to English, German is a DO biased language in which children hardly ever produce POs. We tested this by implementing a baseline session prior to priming.

Within German, we found a strong DO bias in monolingual participants (~80%) and a somewhat weaker DO bias in bilingual children (~70%). In line with Chang et al. (2000; 2006), both groups showed abstract priming and cumulative priming effects (which decreased with growing age) across all ages for the PO but not for the DO structure. Interestingly, we found a significantly higher priming effect in the youngest monolingual children compared to the older groups due to the initially strongest DO bias (90%) which must have caused stronger prediction error. The lexical boost effect also emerged across development in both groups but developed somewhat earlier in the bilingual group (see Figure 1). 

Preliminary results for priming from English to German in adults show the opposite pattern: more adaptation for the DO structure immediately and across time which is the somewhat less preferred option in English. We also find a translation equivalent boost effect (see Figure 2). By the time of the conference, we intend to acquire bilingual child data and answer the following questions: will English-German speaking children show 1) similar priming trajectories as monolingual and bilingual children do within German? 2) a translation equivalent boost effect which will also grow across age? We intend to discuss our results within the framework of current priming accounts in children and adults.

References:

1. Bock, J. K. (1986). Syntactic persistence in language production. Cognitive Psychology, 18(3), 355–387. https://doi.org/10.1016/0010-0285(86)90004-6
2. Bock, K., & Griffin, Z. M. (2000). The persistence of structual priming: Transient activation or implicit learning? Journal of Experimental Psychology: General, 129(2), 177–192. https://doi.org/10.10371/D096-3445.129.2.1T/The
3. Chang, F., Dell, G. S., Bock, J. K., & Griffin, Z. M. (2000). Structural priming as implicit learning: A comparison of models of sentence production. Journal of Psycholinguistic Research, 29(2), 217–230.
4. Chang, F., Dell, G. S., & Bock, K. (2006). Becoming syntactic. Psychological Review, 113(2), 234–272. https://doi.org/10.1037/0033-295X.113.2.234
5. Jaeger, T. F., & Snider, N. (2008). Implicit learning and syntactic persistence: Surprisal and Cumulativity. Proceedings of the 30th Annual Meeting of the Cognitive Science Society (CogSci08), 1061–1066.
6. Pickering, M. J., & Branigan, H. P. (1998). The representation of verbs: Evidence from syntactic priming in language production. Journal of Memory and Language, 39(4), 633–651. https://doi.org/10.1006/jmla.1998.2592

---

[1] For more details and for references, see Schlechtweg et al. (2023).