Abstract Bilingualism research sits at the intersection of linguistics, cognitive neuroscience, and psychology, yet these disciplines often operate in relative isolation, limiting explanatory power and theoretical progress. This paper argues that deeper, more purposeful cross-fertilization among these fields is both possible and necessary to address some of the most pressing questions in the study of bilingualism. After establishing a shared foundation through a concise history of mind–brain–language inquiry, we outline the componential architecture of language — spanning phonology, syntax, morphology, semantics, and pragmatics — to underscore the linguistic sophistication required for meaningful neuroscientific investigation. We then review the neuroanatomical substrates of language representation, processing, and cognitive control, emphasizing distributed network models over classical localizationist accounts. The contribution of electroencephalography (EEG) is examined, including event-related potentials, oscillatory dynamics, and resting-state measures, highlighting their capacity to capture how bilingual experience reshapes linguistic processing and neural function over time and at the individual level. Two exemplars operationalize our convergence argument. Exemplar 1 demonstrates how early brain states measured via EEG can predict individual differences in bilingual linguistic developmental trajectories, offering prospective, mechanistic insights that retrospective adult studies cannot. Exemplar 2 illustrates how linguistically principled formalization of relative linguistic typological distance (RLTD) is indispensable for studies seeking to link cross-language structural differences to bilingualism-induced neurocognitive adaptations. Together, these cases show that linguistics needs neuroscience to ground its constructs in neural mechanisms, while neuroscience needs linguistics to properly characterize what is being processed. Genuine interdisciplinary integration, not parallel inquiry, is the path forward for bilingualism research.
This mixed-methods study investigates how bilingual exposure and Theory of Mind (ToM) in autistic and neurotypical children relate to two dimensions of narrative coherence (i.e., referential and causal coherence). Study 1 analysed oral narratives from 34 children (7 autistic, 27 neurotypical) using regression modelling. Results indicated that diagnostic group predicted referential coherence, with autistic children scoring lower after controlling for age and vocabulary. Both first- and second-order ToM were associated with referential coherence in autistic children but not neurotypical peers. No group difference emerged for causal coherence, which was predicted by age, vocabulary, and heritage language exposure. An exploratory follow-up of eight children approximately one year later found that all four children who improved in causal coherence were bilingual, a pattern descriptively consistent with the results of Study 1. These findings highlight the value of examining referential and causal coherence as distinct dimensions with different predictor profiles, and of modelling bilingual exposure as a continuous construct in research on children's narrative development.
Abstract This study investigated the role of typological similarity and structural overlap in the acquisition of morphosyntax–discourse interface phenomena in L2 Korean. Two experiments examined plural marking agreement and information structure among Korean native speakers, L1-Chinese learners, and L1-Japanese learners. Experiment 1 showed that only native speakers, and not the L2 learners, showed robust sensitivity to plural marking violations, highlighting the persistent difficulty associated with integrating morphology and discourse. Importantly, the non-convergence of L1-Japanese learners, despite the typological proximity between Japanese and Korean, raised the question of whether structural overlap, rather than broad typological similarity, might be the decisive factor in interface acquisition. To address this issue, Experiment 2 investigated information structure, a domain where Japanese and Korean share fine-grained correspondences. In this experiment, L1-Japanese learners successfully aligned with the given-before-new ordering principle, whereas L1-Chinese learners did not. Taken together, the findings suggest that fine-grained structural overlap may provide a relatively facilitative advantage in L2 interface acquisition. These results point to the need to view crosslinguistic influence at external interfaces as a dynamic interaction of structural and typological factors.
Theoretical accounts of second language (L2) reading differ in whether sentence context compensates for inefficiencies in L2 lexical processing (Compensatory Processing Hypothesis) or is constrained by them (Lexical Bottleneck Hypothesis). This eye-tracking study tested these competing predictions in 80 Chinese learners of English by examining how contextual predictability interacts with word frequency and orthographic neighbourhood size, and whether these effects are modulated by L2 proficiency. Results showed additive effects of frequency and predictability in early processing measures (first fixation duration and gaze duration). In contrast, total fixation duration and skipping rates revealed significant interactions such that contextual facilitation was strongest for high-frequency words. This pattern challenges the compensatory view and instead suggests that efficient lexical retrieval is a prerequisite for effective context integration. Proficiency also modulated the effect of neighbourhood size in first fixation duration and the interaction between neighbourhood size and predictability in total fixation duration, indicating that the use of lexical and contextual information differs across proficiency levels. Overall, the findings support a lexical bottleneck account of L2 reading, whereby lexical quality shapes both the time course and the effectiveness of context use.
This study investigated the role of typological similarity and structural overlap in the acquisition of morphosyntax-discourse interface phenomena in L2 Korean. Two experiments examined plural marking agreement and information structure among Korean native speakers, L1-Chinese learners, and L1-Japanese learners. Experiment 1 showed that only native speakers, and not the L2 learners, showed robust sensitivity to plural marking violations, highlighting the persistent difficulty associated with integrating morphology and discourse. Importantly, the non-convergence of L1-Japanese learners, despite the typological proximity between Japanese and Korean, raised the question of whether structural overlap, rather than broad typological similarity, might be the decisive factor in interface acquisition. To address this issue, Experiment 2 investigated information structure, a domain where Japanese and Korean share fine-grained correspondences. In this experiment, L1-Japanese learners successfully aligned with the given-before-new ordering principle, whereas L1-Chinese learners did not. Taken together, the findings suggest that fine-grained structural overlap may provide a relatively facilitative advantage in L2 interface acquisition. These results point to the need to view crosslinguistic influence at external interfaces as a dynamic interaction of structural and typological factors.
Research on grammatical gender in bilinguals has mostly examined L1 effects on L2, with less focus on reverse influence or individual differences. Using visual world eye-tracking, we tested bidirectional interactions between Polish (L1) and German (L2) gender in 57 late unbalanced Polish-German bilinguals. Participants performed spoken word recognition in both languages with picture pairs whose nouns shared gender in the target language, with targets either congruent or incongruent with the other language's gender. We also measured gender-representation stability, operationalised as mean accuracy and certainty in gender assignment. Mixed-effects analyses revealed significant gender congruency effects in both languages, which were stronger among participants with lower mean assignment accuracy. In L2 German, effects were additionally modulated by certainty. Gender congruency effects in L1 were overall weaker than in L2. These findings suggest integrated cross-language gender representations, modulated by representational stability, and are discussed within parasitic accounts of the bilingual lexicon.
Heritage language research has predominantly examined immigrant communities where cross border relocation reduces input and typically renders the heritage language 'weaker'. This study instead examines heritage Slovene in a non immigrant, border shift context, where speakers remain in place but are incorporated into a new state creating an 'in situ' heritage setting with different input dynamics. We compare two such Italian communities, Gorizia-Trieste and Natisone, which differ in the degree of schooling and broader institutional support for Slovene. Using a production task targeting Genitive of Negation and pronominal clitic placement, we compare heritage and monolingual speakers. The Gorizia-Trieste group benefiting from sustained institutional backing patterns closely with monolinguals, whereas the Natisone group shows restructuring in Genitive of Negation and elevated clitic omission, consistent with reduced input and increased processing demands. These results indicate that in border shift settings, strong institutional support can mitigate heritage divergence and challenge simple 'strong vs. weak language' characterizations, with some communities approaching balanced bilingual profiles.
Previous research has shown that past-tense inflection is particularly difficult for child learners of English whose L1 is Chinese, with persistent omission often attributed to cross-linguistic influence (CLI). Building on this literature, this study investigates how Chinese-English bilingual children in the United States (US, n = 4) realize English past tense in naturalistic production between 2;8 and 6;11, focusing on CLI from their heritage language. We examined 11,682 utterances and identified 1,052 lexical, copular, and auxiliary verb tokens in obligatory contexts for past-tense marking. All four children displayed optional past-tense inflection in some or all observational periods, along with overregularization and a previously unreported pattern of overmarking. The production patterns suggest that the persistent difficulty of the Chinese-L1 learners in an English-dominant context is a result of the lack of tense marking in Chinese and the overall isolating typology of the language, rather than transfer from specific phonological, morphosyntactic, or lexical elements. Our findings show that heritage bilingual acquisition is a manifestation of language contact at the individual level, and shed new light on how language contact may drive language change at the communal level.
This paper reports on a study investigating variation in negative existential constructions in varieties of Russian spoken by bilinguals from different regions of Russia. Non-standard marking of the negated subject in such constructions is reported in the literature as a common feature of varieties of bilingual Russian and is usually explained as the result of pattern borrowing due to contact with languages that, unlike Russian, do not have any special marking of the subject in such constructions. In this study, we analyze data from seven spoken corpora of bilingual Russian and compare them to data from a spoken corpus of monolingual Russian. The results of the analysis show that pattern borrowing can account for part of the variation patterns attested in our data, but not for all of them. By applying a quantitative variationist method to data analysis, we show that variation is motivated by an interaction of multiple linguistic and extralinguistic factors, not all of which are related to language contact.
This study examined the interplay between cross-linguistic influence, structural priming, and language variation. Specifically, it investigated whether Spanish–Majorcan Catalan bilinguals process expressed and unexpressed object constructions similarly, and whether they show evidence of cross-linguistic influence and synchronic change arising from sustained contact between the two languages. While previous research has examined null-object representations in monolinguals, much less is known about how bilinguals of languages that (dis)allow object omission, process such constructions. To address this gap, forty-three Spanish–Majorcan Catalan bilinguals with varying dominance profiles completed a comprehension-based structural priming task in Spanish. The task included both grammatical and ungrammatical Spanish sentences, the latter incorporating constructions modeled on the prescriptive Catalan partitive clitic en to introduce Catalan-influenced syntax into Spanish. Results revealed differential priming effects for expressed vs. unexpressed objects. Priming effects with ungrammatical constructions further supported an implicit learning account of structural priming. Finally, although not mediated by priming, cross-linguistic influence from Catalan to Spanish indicated ongoing synchronic change in the bilingual grammar of this population. This study underscores the role of individual linguistic experience in shaping bilingual syntactic processing and highlights the need for more experience-sensitive approaches to understanding how contact-induced grammatical convergence develops in bilingual speakers.
Previous research on bilingual speech production has mainly examined individuals whose first languages (L1s) differ in vowel inventories, often focusing on individual vowels and revealing distinct phonetic realizations for cross-language similar sounds. However, little is known about whether bilinguals with identical vowel inventories phonetically distinguish between their two L1s. This study examined the acoustic characteristics of 38 Basque-Spanish early bilinguals' production of the vowels /a, e, i, o, u/ and tested whether code-switching frequency and amount of language use influence individual vowels and vowel space in each language. The results showed that /i/ was lower and /i, o, u/ were more fronted in Basque than in Spanish. In addition, bilinguals exhibited a larger vowel space in Spanish compared to Basque, with frequent code-switching diminishing this difference regardless of language use. These findings provide updated acoustic norms for Gipuzkoan Basque and Northern Castilian Spanish vowels and demonstrate that code-switching experience shapes phonetic realizations in early bilinguals whose L1s share identical vowel inventories.
This article reports on an empirical study investigating the second language (L2) acquisition of the Chinese perfective markers le and guo and their aspectual/temporal features by English-speaking learners, under the framework of the Feature Reassembly Hypothesis ( Lardiere, 2008 , 2009 ). English lacks dedicated perfective markers, whereas Chinese is rich in aspect marking but has no morphologically realised tense. This study provides a feature-based account of the Chinese perfective markers le and guo and their English corresponding forms (the simple past tense -ed and the present perfect). An acceptability judgment task, a multiple-choice task, and an online sentence completeness judgment task were conducted on 65 L2 learners of three proficiency levels and 25 Chinese natives. Our findings are largely compatible with the predictions of the FRH. It is found that a complex mapping between the L1 tense and L2 aspect markers can lead to persistent learning challenges. Le is more difficult to acquire than guo as the acquisition of le involves a more intricate feature reassembly process. L2 learners must disentangle and reconfigure aspectual and temporal features in the L2 feature sets, while also distinguishing between different aspect markers in L2 Chinese.
Cho (2022) investigated the computation of conversational implicatures (CIs) in two English determiners, the and that. Although both can be used to indicate uniqueness, only that carries an implication of contrast, which distinguishes between two nouns from the same category. It was found that L1-Korean L2-English speakers did not compute this implication due to functional overlap between the English determiners and the Korean demonstrative ku 'that.' Replicating Cho (2022), the present study examined the role of learning contexts in the computation of the implication of contrast by L1-Thai EFL (English as a Foreign Language) and ESL (English as a Second Language) speakers. Since CIs lie at the semantics-pragmatics interface and positive effects of naturalistic exposure in pragmatic development have been reported, L1-Thai ESL learners were predicted to benefit from naturalistic exposure L2 English, overriding L1 transfer, and to show sensitivity to the implication. EFL learners, by contrast, would show non-targetlike performance due to L1 transfer. Results from an acceptability judgment task revealed that, unlike L1-English speakers, both L1-Thai EFL and ESL learners did not compute the implication of contrast, corroborating the original study and revealing no significant effect of naturalistic exposure on the computation of CIs.
