ABSTRACT. Three main accounts have been proposed to explain heritage speaker acquisition: incomplete acquisition (Montrul, 2008; Polinsky, 2006; Silva-Corvalan, 2003), attrition (Polinsky, 2011) and parental input effects (Pascual y Cabo, 2018; Montrul & Sanchez-Walker, 2013; Pires & Rothman, 2009). Few studies have focused on children heritage speakers (Laskowski, 2009: Montrul 2016) although this group of speakers experience a major shift in their linguistic input and social context around the age of 4: they start school in the majority language. In this study we focus on the acquisition of Case by a group of 30 Polish heritage speaker children (HSC) in England in order to try to tease apart these three possible accounts.

Starting school is a crucial event in the life of HSC: it signals the exposure to the majority language and a significant change in their linguistic input; it also signals a change in their social environment. Hence, we investigate whether starting school in the majority language is a key event with significant consequences for heritage language acquisition, i.e. whether the heritage children’s grammatical system is complete or not before they enter mainstream education in the majority language, and/or whether any changes can be observed after one year of schooling. We focus on a group of 30 young Polish heritage children in the UK and investigate their development of Case marking during their first year in primary school. Polish marks Case on nouns with specific morphological forms in a consistent and uniform fashion which is not found in English. Previous studies have shown that adult Polish heritage speakers have problems with particular Case markings (e.g. locative) (Koźmińska, 2015; Wolski-Moskoff, 2019), but not others (i.e. nominative). We tested the productive command of these forms using a narrative retelling task and appropriate grammatical knowledge using an acceptability judgement task. Sociograms were used to investigate the heritage children’s social networks. The children were tested twice, at the start and at the end of their first year in primary school. The Bilingual Language Experience Calculator (Unsworth, 2013) was used to calculate how much input (quantity and quality) they receive.

Out of the three tested cases, (nominative, genitive and locative), nominative remains ‘intact’ throughout the first year. There is evidence of attrition for genitive, as 20% of the children show lower accuracy with this case that was 100% accurate at the start. Locative is the most challenging case as the mean accuracy rate is already lower at the beginning of the school year than for the other two cases (92.23%), but it decreases further as the year progresses to 79.89% which points to both incomplete acquisition and attrition for this case. These results reveal that attrition of certain morphosyntactic forms does occur after only 9 months in school but that this change, which is only attested in some cases and for some children, is modulated by both linguistic and social factors.

ABSTRACT. Aspectual morphology is an area of vulnerability in adult heritage speakers (Montrul, 2016; Polinsky & Scontras, 2019). This study traces such vulnerability to childhood by examining Turkish evidentiality system in child Turkish heritage speakers in the U.S. and the parental generation, who are the main input providers. If child heritage speakers show significant variability in their knowledge of evidentiality as compared to their parents, we hypothesize that the main cause of variability in heritage speakers’ ultimate attainment is insufficient input. However, if child heritage speakers are monolingual-like, then variability at a later age could be due to potential changes in the knowledge of evidentiality in later years. Finally, if first-generation immigrants show different performance from adults in the homeland, then parental input quality can be assumed to contribute significantly to morphological variability in heritage speakers (Coşkun Kunduz & Montrul, 2022).

In Turkish, there are two past tense morphemes to encode evidentiality, namely the direct evidential – DI and the indirect evidential -mIş. As shown in (1), -DI indicates that the speaker has witnessed the event, whereas in (2), -mIş expresses that the speaker has indirect information (hearsay or inference) (Aksu-Koç, 1988). –mIş in Turkish is also used in traditional children’s narratives and the expression of admiration. Research has shown that the acquisition of evidentiality is a late achievement and that Turkish-speaking children do not fully acquire Turkish evidential paradigm until after the age of 7 (Özturk & Papafragou, 2016). A small number of studies also revealed that Turkish adult heritage speakers show variability in their processing and production of evidential markers in Turkish (Arslan et al., 2017).

(1) Ali ara-dı. Ali call-D.PAST-3SG ‘Ali called.’ (2) Ali ara-mış. Ali call-M.PAST-3SG ‘Ali called.’

If longitudinal studies are not possible, one way to understand the root of this variability in young adult heritage speakers is to examine children. Our study contributes to this goal by investigating the acquisition of evidentiality in 20 first-generation immigrants, 20 adult and 20 child (aged 7–14) Turkish heritage speakers, and the monolingual group including 20 Turkish-speaking adults, 20 7–14-year-old and 20 3–6-year-old Turkish-speaking children in Turkey who completed a story retelling task (SRT) and a picture selection task (PST). Results showed that the parental generation patterned with the older children and the adult Turkish speakers in the homeland in the PST (Figure 2), but not in the SRT (Figure 1). However, the child and adult heritage speakers showed variability in both tasks, patterning with the 3-6 year-old Turkish children. Moreover, the performance of each child-parent pair did not necessarily follow the same pattern (Figure 3). These findings suggest that the variability in heritage evidentiality is more likely due to insufficient input in the early years of heritage language development than to changes in parental input. The innovations of this study lie in the comparison of child heritage speakers of Turkish to their input providers (their parents) who are first-generation immigrants and the use of comprehension and production measures.

ABSTRACT. This study applies a mixed methods approach to heritage language socialization. It examines the understudied role of overheard speech in the maintenance of third-generation heritage Spanish in a Mexican family in L.A. Quantitative measures reveal how much the family uses Spanish longitudinally. Qualitative analyses describe the use of overheard Spanish through the prism of multigenerational, and multiparty interactions. Based on longitudinal, and naturalistic video-recorded data collected through ethnography, quantitative and qualitative analyses were carried out to obtain a snapshot of the quantity and quality of heritage Spanish (over)heard input in a third-generation bilingual language learning environment. 24 hours of spontaneous interactions spread across three sampling periods were transcribed on CLAN to create a plurilingual corpus. They were further coded on Excel according to additional sociolinguistic categories. The data allowed the analyses of nearly 30,000 heritage Spanish, English, and language mixed utterances. The quantitative data shows dynamic linguistic practices. Spanish was used most in the first sampling period, before its use significantly dropped in the last two periods both at the family, and the individual level. The qualitative data provides evidence that the third-generation target-child is finely attuned to overheard Spanish, and motivated to use it especially when it is addressed to her younger brother. Overheard input not only helps her use heritage Spanish, but it also allows her to temporarily enter, and sometimes influence the direction of the multigenerational, and multiparty social encounters. This paper delves into the understudied role of overheard input in studies of heritage language socialization. Through the creation of a plurilingual corpus it explores the language practices of one multigenerational Mexican family, who belong to a well-established population in multicultural California. Furthermore, it provides an account of how the dynamic discursive practices in multiparty interaction support the meaning making process, and how it helps a third-generation child use heritage Spanish. The findings contribute to the ongoing discussion of what it means to be bilingual in general, and definitions of heritage bilingualism in particular. The results underscore the role of overhead speech in heritage language socialization. They also suggest that heritage bilingualism is not just about abstract grammatical rules, but rather about inheriting bilingual and bicultural practices constructed by, with, and for her community of practice.

ABSTRACT. The migration to the Baltic states was always both popular (in the search for a better quality of life in the Soviet “West”) and encouraged by Soviet authorities. The share of the Russian-speaking population is notable in Estonia: 29%. Estonian is a “medium-sized” national language that shares key domains with English as a global language. Family is a site where languages are managed: language practices constructed by family members are further negotiated with the larger society (Fishman, 1966; King et al., 2008; Schwartz, 2010; Spolsky, 2012). In my paper, I approach the topic at the micro level. The families were chosen to reflect the sociolinguistic diversity of Estonian regions: the bilingual capital Tallinn, the dominantly Russian-speaking Ida-Viru County (in north-eastern Estonia) and the dominantly Estonian-speaking southern Estonia. This study is based on an in-depth analysis of a variety of sources, including qualitative sociological materials (semi-structured in-depth interviews with 30 couples from Russian-Estonian mixed families and participant observations) and quantitative statistical and demographic data on self-reported language behaviour and language ideologies, revealing the “context” of community types. The couples all represent the traditional family model. There were also three different age groups (25–39, 40–59 and over 60 years old). The interview consisted of three parts: (1) the socialization of each individual, (2) the socialization of the couple, and (3) the socialization of the children. This paper primarily focuses on the language policies that bilingual Russian-Estonian families follow in relation to the maintenance of Russian as a heritage language, in order to identify social variables which either favour or hinder this process. A more concrete task is to search for commonalities and specifics of each family type within broad categories of the mainstream attitudes towards Russian as a heritage language. The aim of the paper is to analyse the sociolinguistic situation of the Russian language in Estonia and to examine the factors which have defined the maintenance of Russian as a heritage language. The paper also investigates translanguaging practices of Russian-Estonian families, including their perceptions of and attitudes towards their multilingual communication. The languages are used creatively to convey or negotiate meaning and identity. This paper presents results from ethnographic fieldwork studies conducted in different regions of Estonia, and thus offers important conclusions about sociolinguistic variation in heritage language maintenance and loss. It provides evidence of how social milieu and different sociolinguistic backgrounds may affect all processes related to heritage language transmission: management, maintenance, use and proficiency. References Fishman, Joshua A. (1966) Language Loyalty in the United States: The Maintenance and Perpetuation of Non-English Mother Tongues by American Ethnic and Religious Groups. Hague, The Netherlands: Mouton & Co. King, Kendall A., Lyn Fogle and Aubrey Logan-Terry (2008) Family Language Policy. Language and Linguistics Compass 2, 907–922. Schwartz, Mila (2010) Family language policy: Core issues of an emerging field. Applied Linguistics Review 1, 171–192. Spolsky, Bernard (2012) Family language policy: The critical domain. Journal of Multilingual and Multicultural Development 33, 3–11.

ABSTRACT. In this ongoing study, we examine heritage language (HL) and child second language (cL2) development in the European refugee context. When examined as adults, HL bilinguals tend to show significant differences in their HL performance and competence from typical monolingual peers, and from one another (e.g., Rothman 2009, Rothman & Treffers-Daller 2014). Given the lack of studies, not enough is known about HL development in childhood and adolescence (e.g., Meisel 2011, Montrul 2012, Kupisch and Rothman 2016), and even less is known about HL continued development in refugee contexts.

This empirical study investigates linguistic and extra-linguistic variables affecting the development of HL Syrian Arabic and cL2 German in Germany.

The goal is to understand the extent to which HL exposure leads to HL maintenance and its impact on the path of cL2 development. We investigate the following research questions: RQ1. To what extent do experiential variables and HL support affect HL trajectories and outcomes? RQ2. To what extent does HL support affect cL2 development?

We test the following hypotheses: (i) there is an inverse relationship between cL2 growth and continued growth in the HL (ii) the general trend in (i) can be attenuated by increased exposure, use and support for the HL, and (iii) heritage speakers who receive formal training in their HL will have a sharper growth trajectory.

Data collection is still ongoing. Two groups of 6-to 12-year-old children (n=60) are included in this study. The two groups are Syrian Arabic heritage speakers with and without HL literacy. The division between children who do and do not receive education in the HL allows us to understand if support for the HL outside the home has, among other variables, an impact on the developmental trajectories of both HL and L2.

We investigate grammatical gender as it has been shown to be a vulnerable domain for heritage speakers, i.e., a decrease of target-likeness in the HL and/or protracting delays or lingering difficulties with gender marking in the L2. The following three experiments are used: Peabody Picture Vocabulary Test for density of vocabulary, an eye-tracking experiment on grammatical gender to test comprehension and subsequent use of gender for predictive processing, and the Quantifying Bilingual Experience Questionnaire (Q-BEx) (DeCat et al. 2021) to collect information on experiential variables pertaining to Syrian Arabic and German exposure/use which are used as regressors in our modelling of the results.

Our preliminary findings suggest that on the group level the heritage speaker children do not use gender cues predictively in Syrian Arabic (see figure 1) and German (see figure 2). However, on the individual-level we find a variation showing that individual children use gender cues predictively in Syrian Arabic, German, none or both of the languages. In order to answer our research questions, we are currently collecting more data, as we do not have enough statistical power so far to find HL support or other experiential variables to predict trajectories.

(see attached file for references and figures)

ABSTRACT. Differential acquisition of Aspect in Spanish by Spanish-English Heritage bilinguals is a key argument in support of incomplete acquisition in heritage bilinguals (Montrul 2002, 2006, Montrul & Perpiñan 2011). The evidence has come primarily from studies of the production and interpretation of preterite/imperfective morphology (see Montrul 2022 for a recent overview). Less studied has been the telicity marker se. We explore the effects of age of acquisition (AoA) of English, level of lexical access in Spanish, and English dominance on the telic interpretations of se among heritage speakers. Following Beavers (2011:352), we define a predicate as telic if and only if for any event e described by it, the predicate does not describe any subevent of e. This requires that the theme have a quantity and the predicate’s scale be bounded (the scalar maximum is reached) (Martínez Vera 2022, Beavers 2011). In English, telicity arises from the combination of the verb scalar meaning and the theme’s quantity, e.g., in eat the apple, eat has an extent scale, and the apple indicates fixed quantity; context plays a role in determining (a)telicity. In Spanish, while this strategy is present, there also are lexical means affecting telicity. Aspectual se imposes a requirement that the verb’s scale be bounded, and the theme’s quantity be fixed. Thus, comerse la manzana ‘eat up the apple’ is telic, since se indicates that the event ends up when the apple is consumed. Comer la manzana ‘eat the apple’, in contrast, is not subject to this requirement. Twenty-six Spanish-English heritage bilinguals, 36 monolingually raised English speakers, and 25 monolingually-raised Spanish speakers completed a forced-choice picture selection task to determine whether they interpreted Spanish sentences with and without se as telic or atelic. They also completed an English version with sentences like The girl ate the apple. Lexical access was measured using the MiNT (Gollan et al. 2012), and dominance was measured using the Bilingual Language Profile Questionnaire (Gertken et al. 2014) Results show that se generates telic interpretations for the Heritage and monolingually-raised groups (β=1.22084, SE=0.45889, 2.660, 0.0078). There were no differences between the two groups regarding the effect of se (β=0.39341, SE=-0.27904, z=-1.410, p=0.1586), and among the Heritage bilinguals, se had a more substantial effect on telic interpretations (β=1.2675, SE=0.4269, z=2.969, p=0.00299). Among the heritage bilinguals, there were no effects of English AoA. Results were close to showing an effect of Spanish MiNT high scores on lower proportions of telic interpretations of items without se (β=0.3713, SE=0.1953, z=1.902, p=0.05724). There were no dominance effects. Self-reported oral comprehension in Spanish and the desire to be perceived as native speakers also had an effect on telic interpretations of se. These results suggest that the heritage group did not differ significantly from their monolingually-raised counterparts and showed no evidence of incomplete acquisition.

References Beavers, J. On affectedness. 2011. Natural Language and Linguistic Theory 29(2): 335-370. Gertken LM, Amengual M, Birdsong D. 2014. Assessing language dominance with the bilingual language profile. In Measuring L2 Proficiency: Perspectives from SLA, ed. P Leclercq, A Edmonds, H Hilton, pp. 208–25. Bristol, UK: Multilingual Matters. Gollan, T. H., Weissberger, G. H., Runnqvist, E., Montoya, R. I., & Cera, C. M. (2012). Self-ratings of spoken language dominance: A Multilingual Naming Test (MINT) and preliminary norms for young and aging Spanish–English bilinguals. Bilingualism: language and cognition, 15(3), 594-615. Martínez Vera, G. 2022. Revisiting aspectual se in Spanish: telicity, statives and maximization. The Linguistic Review 39(1): 159-202. Toth, P., & Guijarro-Fuentes, P. (2013). The impact of instruction on second-language implicit knowledge: Evidence against encapsulation. Applied Psycholinguistics, 34(6), 1163-1193. doi:10.1017/S0142716412000197 Montrul, S. (2002). Incomplete acquisition and attrition of Spanish tense/aspect distinctions in adult bilinguals. Bilingualism: Language and Cognition, 5(1), 39-68. Montrul, Silvina. 2006. Incomplete acquisition as a feature of L2 and bilingual grammars. Inquiries in linguistic development. In honor of Lydia White,ed. by Slabakova Roumyana, Montrul Silvina, and Prévost P., 335–360. Amsterdam, The Netherlands: John Benjamins. Montrul, S., & Perpiñán, S. (2011). Assessing differences and similarities between instructed heritage language learners and L2 learners in their knowledge of Spanish tense-aspect and mood (TAM) morphology. Heritage Language Journal, 8(1), 90-133.

ABSTRACT. Recent studies claim that heritage speakers (HSs) are more likely to differ from monolinguals in formal situations than in informal ones since HSs are typically exposed to colloquial variants of their heritage language (HL) (Flores & Rinke, 2020; Lohndal et al., 2019). The current study addresses an effect of different formalities and modes on the word order choice by HSs of Russian in the U.S and monolingual speakers of Russian. Standard Russian is an SVO language that allows great word order variation which is discourse driven. SVO and OVS orders are typically more frequent in the formal written and spoken situations while SOV order is usually more frequent in the informal spoken situations (Sirotinina, 2003; Slioussar, 2007). Word order in HLs was reported to be prone to change and transfer from the majority languages (MLs) (Polinsky, 2018). Studies on word order in heritage Russian in the US generally report on the increase of the SVO order use and the reduction of word order flexibility in productions of HSs (Polinsky, 2006; Laleko & Dubinina, 2018). Although word order in heritage Russian received a great deal of attention, the role of different formalities and modes remains unclear. I will present semi-spontaneous data of HSs and monolinguals elicited according to the “Language Situations” method (Wiese, 2020). Participants were asked to describe a fictional car accident in two formalities (formal and informal) and two modes (spoken and written). The productions of 40 HSs and 40 monolinguals (320 productions overall) were analysed. The data were manually annotated for word order. Three word orders (SVO, OVS, SOV) were analyzed using the binomial generalized linear mixed-effects models. Each word order was modelled as a function of Speaker group (HSs/monolinguals), formality (formal/informal) and mode (spoken/written). First, the results of the study revealed that some word orders were influenced by formality and mode in productions of HSs and monolinguals in a similar manner. For instance, SOV order was more frequent in the informal and spoken situations compared to the formal and written ones (HSs formal vs informal: z=-3.58, p < .01; monolinguals formal vs informal: z=-7.11, p< .0001; spoken vs written: z=3.01, p< .01). Second, the two groups were found to differ from each other regarding their choice of some word orders. HSs produced more SVO utterances in the written mode than monolinguals (z=-2.88, p<.05), but the two groups did not differ from each in the spoken mode (z=-1.03, n.s). Besides, HSs produced less OVS orders in the formal situations than monolinguals (z=4.49, p<.0001), but there was no difference between the groups in the informal situations (z=0.95, n.s). The abovedescribed results showed that HSs did not differ from monolinguals across the board, but only in the formal and written situations. Thus, the differences between the two groups may be attributed to the differences in the language situations rather than the effect of language contact (confirming previous studies, e.g., Rothman 2007, Kupisch & Rothman, 2018).

ABSTRACT. In Turkish, if a noun phrase (NP) in the direct object position is marked with the accusative case marker –(y)I, then the NP is definite (Kitab-ı aldım ‘I took the book’). However, if the NP in the direct object position is proceeded by the word bir ‘one/a(n)’, then the NP is indefinite (Bir kitap aldım ‘I took a/one book’). Indefinite NPs with the accusative case marker –(y)I are specific and refer to entities that were mentioned in the preceding discourse (Ben bir kitab-ı aldım ‘I took one of the books) (Göksel & Kerslake, 2005). Previous studies with Turkish heritage speakers (HS) have shown that they use the indefinite marker bir more frequently than the monolingually-raised Turkish native speakers (MS) (Backus et al., 2011; Felser & Arslan, 2019). These results indicate that definiteness is a vulnerable phenomenon for Turkish HS as they have to integrate different language modules (i.e. morphosyntax and pragmatics). The current study investigates if the correct use of definiteness is difficult for HS (n=59) during real time sentence processing by exploring their reading times (RTs) and end of sentence responses in comparison to Turkish MS (n=39). All experimental sentences had a context sentence with either a plural object (e.g. Masanın üzerinde kalın kitaplar vardı. “There were thick books on the table.”) or a singular object. (e.g. Masanın üzerinde kalın bir kitap vardı. “There was a thick book on the table.”). The context sentence is followed by the experimental sentences with a definite marker (DS) or an indefinite marker (IS) (see Table 1). The experiment was designed on the web-based platform Ibex Farm. To counter the problems of word length and individual differences in reading, residual RTs (RRTs) were analysed for the critical region “CR” and the “Spillover” region (see Table 1). The RRTs data in the “CR” reveals a significant definiteness effect (slower RRTs for definite sentences) and a marginally significant definiteness*plurality*group interaction (p=.056). The marginal interaction indicates a significant definiteness and plurality interaction for the MS but not for the HS group. In the “Spillover” region, there was no significant group difference or any significant interactions with the factor “group”. The end of sentence data reveals a significant group effect (the HS with more accepting responses), a significant definiteness*group and a significant definiteness*plurality*group interaction. The definiteness*group interaction reveals that the HS group gives significantly more accepting responses both in definite and indefinite sentences. The three-way interaction shows a significant definiteness*plurality interaction for the MS but not for the HS group. These results suggest that HS experience problems when they have to integrate morphosyntactic and pragmatic information as they behave differently from the MS group. The HS group’s RRTs and end of sentence responses are not affected by the manipulation of definiteness and plurality in the same way as the MS group. HS face difficulties with definiteness because integrating morphosyntactic and pragmatic information is cognitively demanding (Rothman & Slabakova, 2011) and their processing resources are limited (Polinsky & Scontras, 2020) when compared to the MS group.

ABSTRACT. Effect of transfer can be unidirectional, either from L1 to L2, or from L2 to L1, and bidirectional, when both languages of a bilingual are affected (Pavlenko & Jarvis, 2002). The current study aimed to evaluated the directionality of transfer in the domain of pragmatics in heritage language (HL) speakers. Despite the rapid development of HL research, the sphere of HL speech act pragmatics is still in its infancy and covers mainly Spanish and Russian as HLs (see Pinto & Raschio, 2007; Dubinina, 2011; Dubinina & Malamud, 2017; Xiao-Desai, 2019). Moreover, research comparing both languages of HL speakers is rather limited (Scontras et al., 2017; Kupisch et al., 2014; Stangen et al., 2015).

The current study sought to investigate the directionality of transfer in three groups of bilinguals: (a) HL adult speakers (HL-speakers) who have acquired English as their HL and Hebrew as their dominant societal language (SL) (n=20); (b) Hebrew-dominant speakers (HEB-DOM) who were born to Hebrew-speaking families and raised in Israel, and thus English is the language of their schooling (n=20), and (c) English-dominant speakers (ENG-DOM) who were born to English-speaking families and immigrated to Israel from an English-speaking country after the age of eighteen, and thus Hebrew is their L2 (n=20). For the purposes of the study, pragmatic competence in both languages of bilinguals was evaluated. The parallel tasks in English and in Hebrew presented 18 scenarios eliciting requests and 18 scenarios eliciting apologies in each language.

Previous studies report cross-linguistic and cross-cultural differences in request and apology making strategies between English and Hebrew. In order to capture the differences between the two languages, the participants' requests were analyzed for the choice of the syntactic structure of the head act/s (interrogative / imperative / declarative / mixed), the use of modals, and the use of 'please'. Their apologies were analyzed for expressions of apology (IFID's), number of propositions added (i.e., offering explanation, taking responsibility, offering repair or compensation, and promising forbearance), and the use of adverbial intensifiers.

The results indicated that the dominant groups (END-DOM and HEB-DOM) systematically transferred request and apology making strategies from their dominant language into their weaker one, thus pointing at the effects of unidirectional pragmatic transfer. However, the picture was more complex for the HL-speakers; in some cases, their strategies paired up with a dominant group, while in other cases they developed a unique and hybrid linguistic style reflecting both their HL and their SL in order to accommodate to their ever-changing communication needs, thus showing effects of bi-directional transfer.

From a theoretical perspective, the study sheds light on the pragmatic competence of HL-speakers and dominant bilinguals in order to provide a greater understanding of the directionality of transfer. From a pedagogical perspective, the study contributes to the field of teaching pragmatic skills to HL speakers and L2 learners and helps educators develop research-supported curricula that facilitate appropriate politeness strategies.

ABSTRACT. One of the salient characteristics of Heritage Language (HL) speakers (people whose “home” language as children was different from the surrounding societal language (SL)) is that while their HL eventually deviates somewhat from the home baseline in the morphosyntactic and lexical domains (due to a combination of diminished input and cross-linguistic influence), the phonological domain seems to be relatively unaffected (Chang, 2021); Polinsky, 2018)). The HL speaker’s (HS) accent in their HL is often very native-like, and is almost undetectable to the casual listener, to the extent that they are often surprised by the appearance of non-native-like errors in grammar and vocabulary, which don’t seem to fit the accentless production. This tendency is usually attributed to the fact that phonology is acquired very early on in the child’s linguistic development. But there is also evidence of less easily discernible deviations in phonological production of HSs, both in the HL and in the SL (Chang, 2021; Polinsky, 2018; Kupisch et al., 2014; Kupisch et al., 2021).

The current study (in progress) examines accentedness in the unique dyad of adult HL-English speakers in Israel (SL Hebrew), by recruiting raters of different linguistic backgrounds to evaluate their naturalistic speech. The trajectory of HL English often differs from other HLs due to English’s prestigious status, its ubiquity in the international media, and the fact that it is often (e.g., in Israel) part of the child’s formal education.

This research attempts to determine what sound patterns (if any) typically mark the speaker as an HS (rather than a monolingual native English speaker). For example, it has been suggested (Polinsky, 2018) that subtle accentedness is more likely to appear in sounds for which the differences between the two languages are less salient, where the speaker is more likely to employ “leveling” rather than small differentiations. We also examine to what extent the linguistic background of the raters affects detectability – we compare the judgments of monolingual (English) raters with those of Hebrew-dominant native Hebrew speakers in Israel, and with those of English-dominant bilinguals in Israel. Finally, we examine what individual characteristics of the speaker (e.g., childhood input patterns, age of onset of SL, number of English-speaking parents) may predict the level of accentedness.

Naturalistic speech segments in English (descriptions of an incident in a parking lot portrayed in a video provided by the RUEG Corpus (Wiese et al., 2019)) were recorded from 80 speakers divided into four groups: balanced bilinguals with Hebrew preference born in Israel with one native-English-speaking parent; with two English-speaking parents; serial bilinguals who immigrated to Israel as adults; and monolingual English-speakers (control). Raters are asked to rate the level of accentedness (none, light, heavy), their level of certainty, and the aspects of the segment they identified as accented.

We expect a statistical analysis of these responses will shed light on the factors which affect the trajectory of the HL-English speaker in the phonetic/phonological domains, and to contribute to our understanding of accentedness in HL speakers in general.

ABSTRACT. Despite the rapid development of heritage language (hereafter HL) research, the domain of pragmatics remains somewhat neglected (Polinsky, 2018; Dubinina, 2021). Across different language domains, HL speakers are reported to show divergences from baseline speakers, i.e., speakers using this language in the country of origin and/or bilingual speakers who are dominant in this language (Montrul 2016; Polinsky 2018). Previous research on the pragmatic skills of HL speakers has mainly concentrated on speech acts of making requests, in HLs in contact with English as a dominant Societal Language (hereafter SL) (Dubinina & Malamud, 2017, Pinto & Raschio, 2007). The current study investigates cross-linguistic and cross-cultural influence from the SL on production of requests among HL-Russian speakers dominant in SL-Hebrew residing in Israel compared to HL-Russian speakers dominant in SL-German in Germany. The properties of requests, such as the form of address, honorifics, and the use of morphosyntactic particles, contrast in the two societal languages, and therefore they are good candidates for studying cross-linguistic influence under the pressure of the dominant languages. For example, Russian differentiates between two forms of address: formal situations require a polite V-form (e.g., vy ‘you.PL’), while the T-form (e.g., ty ‘you.SG’) is used when the interlocutor is known to the speaker, in informal situations with friends or family, to someone younger, or to someone of a lower social status. German, like Russian, differentiates between formal and informal forms of address, whereas Hebrew does not. We hypothesize that if the comparison of HL-Russian in contact with two different SLs (German and Hebrew) reveals differences between the groups, it will strengthen the evidence for cross-linguistic and cross-cultural influence. Conversely, if no difference between the two groups is found on the investigated parameters, the results will testify to the effect of diminished input. For the purposes of this study, a total of 75 participants equally split into three groups were recruited: adult Russian-Hebrew bilinguals (SL-HE), adult Russian-German bilinguals (SL-DE) and a control group of monolingual Russian-speaking participants (MONO). The participants’ requests are analyzed for the presence of alerters, head acts, and supporting moves. Next, requests are coded for the choice of (a) address forms (T vs. V); (b) morphosyntactic realization: clause type (Declarative, Interrogative, and Imperative) of head acts, the presence of negation particle ne, subjunctive particle by and the interrogative particle li, and (c) lexical strategies (the use of požalujsta ‘please’ and the use of možno-požalujsta ‘possible please’) This is an ongoing study, the data for all three groups have been collected and are currently being analyzed. Our study will be the first to investigate pragmatic phenomena in the HL in contact with two typologically different languages. This will shed light on the pragmatic development and maintenance of HL speakers and provide a better understanding of the mechanisms responsible for shaping requests realizations. Words (464)

References: Dubinina, I. Y., & Malamud, S. A. (2017). Emergent communicative norms in a contact language: Indirect requests in heritage Russian. Linguistics, 55(1), 67-116. Pinto, D., & R. Raschio. (2007). A comparative study of requests in heritage speaker Spanish, L1 Spanish, and L1 English. International Journal of Bilingualism 11, 135– 55. Polinsky, M. (2018). Heritage languages and their speakers (Vol. 159). Cambridge University Press. Blum-Kulka, S., J. House, and G. Kasper (eds.). (1989). Cross-Cultural Pragmatics: Requests and Apologies. New York: Ablex Pub

ABSTRACT. Documentation on the proficiency of heritage speakers (HS) using the ACTFL (American Council for the Teaching of Foreign Languages) proficiency Guidelines (2012) dates back to a 2014 study by Swender et al., followed by a description of writing proficiency for HSs of Chinese, Korean and Spanish (Author 1 and XX, 2017 and 2018). A more recent study by Author 1 and Author 2 (2020) investigated the relationship between oral and writing proficiency of Spanish heritage language learners (HLL), who are HSs enrolled in a heritage language (HL) course. Results of all three studies show support for some commonly-held assumptions (e.g., that earlier age of arrival correlates with higher proficiency in both speaking and writing), but also contributed some less intuitive findings. For example, the 2020 study reports that while Spanish HLLs do tend to be better speakers than writers, the distance between proficiency in the two modalities is, in most cases, no more than one sublevel. More recent research has found a correlation between proficiency ratings and complexity measures (e.g., Kisselev et al., 2021; Kuiken & Vedder, 2019; Park, 2017).

These lines of research are fundamental and informative, and they also create the need for a new line of investigation: taking into account the proficiency of the dominant language. It is well known that, in the U.S., learners enter post-secondary education with varying degrees of writing proficiency in the majority language (English). Knowing this, how can we be certain that what prevents writers from being rated at higher levels of proficiency in their HL is indeed the heritage status of their language, and not instead their overall writing capabilities?

This study attempts to fill this gap in the literature by investigating both languages of university-level bilingual HLLs. We purport that understanding writing proficiency in the two languages helps shed light on the possibility of knowledge transfer across languages. We endeavor to answer the following two questions: (1) Is it accurate to assume that HLLs who received most of their education in the U.S. write better in English (their majority language) than in Spanish (their HL)? (2) If HLLs are in fact better writers in English, how large is the average gap between writing abilities in the two languages?

We collected both English and Spanish writing samples using an instrument modeled after the ACTFL WPT (Writing Proficiency Test) from 103 Spanish HLLs enrolled in university-level language courses designated for HLLs at four- and two-year campuses in the U.S. We also collected biographical data and self-assessments of writing proficiency. The writing samples were double-rated by certified ACTFL raters, and the final rating values were analyzed for statistical significance via Chi square and t-tests. Results show an unexpectedly small gap in proficiency differences across languages, as well as surprising heterogeneity in the majority language. Additional results suggest a relationship between heritage proficiency and student profile type (four- or two-year campus), and the observation that heritage learner self-assessments may be influenced by language attitudes. This presentation will also discuss practical implications of these findings.

References Author 1, & XX. (2017). Removed for anonymization. Author 1, & XX. (2018). Removed for anonymization. Author 1, & Author 2. (2020). Removed for anonymization. ACTFL. (2012). ACTFL proficiency guidelines [Electronic version]. https://www.actfl.org/resources/actfl-proficiency-guidelines-2012. Kisselev, O., Klimov, A., & Kopotev, M. (2021). Syntactic Complexity Measures as Indices of Language Proficiency in Writing: Focus on Heritage Learners of Russian, Heritage Language Journal, 18(3), 1-30. Kuiken, & Vedder, I. (2019). Syntactic complexity across proficiency and languages: L2 and L1 writing in Dutch, Italian and Spanish. International Journal of Applied Linguistics, 29(2), 192–210. Laleko, O., & Kisselev, O. (2021). Heritage language complexity matters: The editors’ introduction to the special issue. Heritage Language Journal, 18(2), 1–8. Park, J. H. (2017). Syntactic complexity as a predictor of second language writing proficiency and writing quality. Michigan State University: Second Language Studies.

Swender, E., Martin, C.L., Rivera‐Martinez, M., & Kagan, O.E. (2014). Exploring oral proficiency profiles of heritage speakers of Russian and Spanish. Foreign Language Annals, 47, 423–446. doi:10.1111/flan.12098

ABSTRACT. Heritage speakers often exhibit structural differences from baseline grammars across linguistic domains (Montrul, 2010). The present study examines Spanish heritage speakers’ (SHS) knowledge of argument structure; specifically, SHSs’ knowledge of resultatives in English and Spanish, in comparison with monolingually raised native speakers of Spanish (MRS) and English (MRE). Crucially, resultatives occur in English but not in Spanish (Pylkkänen, 2008). Resultatives predicate a state of an NP which is a result of the action described by the verb (Hovav & Levin, 2001). Examples (1) and (2) show resultatives. (1) John hammered the metal flat. (2) The lake froze solid. Resultatives are subject to a Direct Object Restriction (Levin & Hovav, 1995), meaning they can only be predicated of objects. As such, they are grammatical with transitive verbs, as in (1) and unaccusative verbs, as in (2). They are ungrammatical with unergatives, as in (3), unless a fake reflexive object is added as in (4). (3) *She sang hoarse. (4) She sang herself hoarse. Given SHSs’ dominance shift to English, they are predicted to have nativelike knowledge of English resultatives. This prediction is supported by previous research which found that SHSs have nativelike knowledge of the dominant language (Montrul, 2006; Montrul & Ionin, 2012). Assuming dominant language influence, SHSs are predicted to incorrectly accept resultatives in Spanish. With respect to dominance, greater English dominance should result in greater acceptance of resultatives in Spanish, while greater Spanish dominance should result in greater rejection of resultatives in Spanish. 19 Spanish heritage speakers, 25 monolingually raised Spanish native speakers and 29 monolingually raised English native speakers completed the study (Table 1). Spanish speaking groups completed the Bilingual Language Profile (Birdsong et al., 2012), Spanish and English cloze tests, and 2, 56-item acceptability judgment tasks, one each in English and Spanish. For the Spanish task, conditions were *Resultative and Resultative which contained items like (1) and their grammatical Spanish counterparts. For the English task, conditions were Transitive Resultative, items like (1), and *Unergative Resultative, items like (3). The MRE group completed only English tasks. According to the statistical results, overall predictions were partially supported. For the Spanish task (Figure 2), there were significant differences between the SHS and the MRS groups. The SHS group accepted significantly more ungrammatical resultatives in Spanish, indicating dominant language influence. For the English task (Figure 1), results diverged from predictions. There were no significant differences found between the SHS and MRS groups. However, there were significant differences found between the SHS group and the MRE group, in the transitive resultative condition. Interestingly, this difference was due to the SHS group demonstrating more categorical judgments in English. This finding confirms the tendency attested in heritage phonetics/phonology in which HSs, “tend to amplify properties that separate their two languages.” (Polinsky, 2018, p. 162) Finally, across both tasks dominance and proficiency were not significant predictors, which may be attributable to the nature of our sample: the high level of English proficiency of the MRS group and the relatively balanced bilingualism of the SHS group.

ABSTRACT. This study investigates the use of filler particles by heritage speakers of Russian in two language contact setting and relates findings to the possible transfer of filler particles as words. Filler particles (FPs) such as uh and uhm in English are not simply symptoms of speech planning but signal discourse boundaries as well as delays in speech. They have therefore been claimed to have word status. Perceptually FPs are not very salient due to a reduced phonetic form and their frequency is not reliably estimated in perception. Their acquisition in a second language may therefore depend on the amount of input. Heritage speakers (HS), bilingual speakers of a heritage language (HL) and the majority language (ML) are exposed to different amounts and registers of input in their two languages. As a result, HS have shown to differ from monolinguals in different areas of their grammar. This study investigates the frequency and segmental form of FPs produced by HS of Russian living in the US and Germany in comparison to monolingual speakers in all three languages. This is done based on data from the RUEG corpus [7] which contains semi-spontaneous narrations in formal and informal register of ML and HL speakers with different language background. Language specific FPs were extracted from the corpus: 3 640 FPs produced in 363 Russian narrations, 2212 FPs in 295 English narrations and 2294 FPs in German 304 narrations. The FP frequency was normalized by 100 words within each narration. The language specific FP forms were categorized as either vocalic (V), nasal (N) or vocalic-nasal (VN; vowel quality was not taken into consideration here). The mean frequency of FPs is comparable for ML speakers with or without HL background while the mean normalized number of FPs in Russian is higher for HS in Germany and the US. Regarding the segmental form language specific preferences emerged as is illustrated in figure 1. In ML English VN forms are preferred over V forms. A similar preference can be seen in the ML German data. The Russian data, however, reveals a different FP use by different speaker groups. Monolingual ML Russian speakers prefer the V form over N and VN forms. HL Russian speakers also predominantly produce V forms but also produce higher number of VN forms. The proportion of VN forms in Russian is highest for HS Russian speakers living in Germany. The analysis of FP use in HL revealed differences between HS and monolinguals. The higher number of FPs in HL Russian could be related to higher cognitive load monitoring two languages at the same time. The different FP forms in HL Russian are evidence for a possible transfer of nasalized FP variants from the respective ML to the HL Russian. The analysis of FP use in ML language, however, shows little differences in the frequency and segmental form. This supports the status of ML speakers with a HL as native speakers of this language.

ABSTRACT. This study investigates the use of overt (OS) and null pronominal subjects (NS) in the oral narratives of ten children speaking Bulgarian as their heritage and dominant language in Germany. Bulgarian is a consistent null-subject language, in which the alternation of OS and NS is dependent on grammatical and discourse conditions (Genevska-Hanke 2019). NS are typically used for topic-continuity (TC), while OS are associated with topic-shift or focus (Sorace 2005). These conditions show cross-linguistic microvariation and are acquired late in L1-acquisition (Tsimpli, 2014, Shin&Cairns, 2009), whereas syntactic aspects of NS come early, Belletti&Guasti (2015). They are well-studied in different populations and found to be vulnerable in bilinguals, even in the absence of cross-linguistic difference (Di Domenico&Baroncini 2018). For bilingual and monolingual children, Kupersmitt&Berman (2001) and Paradis&Navarro (2003) found similar subject use in elicited narratives, but see Serratrice et al. (2004), Sorace et. al. (2009) for reporting differences. Concentrating on Spanish heritage children, Montrul (2018) found overproduction of OS as well as use of ambiguous OS and NS. Focusing only on Bulgarian-dominant children we hope to clarify whether dominance can account for similarities between monolingual and heritage children and explain conflicting results of previous research. We investigate Bulgarian as the heritage language of 4-8-year-old simultaneous bilinguals (N=10) with German as majority language. All are Bulgarian-dominant, where dominance is calculated as an experiential index from questionnaire data (Tuller 2015). Their oral narratives elicited with the MAIN protocols for telling and retelling (Gagarina et al. 2012) were compared to those of Bulgarian monolingual children (N=10) and adults (N=10). 50 narratives were analyzed (adults only told the stories) for non-/ambiguous OS/NS in TC- and shift contexts. Results are shown in figure 1. First, all speakers used OS and NS to a comparable extent. Second, the analysis of TC and -shift contexts showed that both groups of children slightly overproduced OS in TC and also used more ambiguous OS than adults (example 1). These patterns, however, failed to reach statistical significance, see Kupersmitt&Berman 2001/Serratrice 2007. Third, the heritage children used significantly more ambiguous NS (example 2) than the monolinguals, p < 0.001, in-line with Montrul (2018). Fourth, adults also produced some OS in TC, and marginally ambiguous OS and NS. The first result supports early acquisition of NS syntax. The second and third indicate that relevant discourse constraints are not yet in place, as Shin&Cairns (2009) showed for Spanish monolingual acquisition. The fourth result, the overproduction of OS in TC and the use of ambiguous OS by adults, speaks to the gradient nature of the phenomenon. Bilingualism, development and cross-linguistic influence (the fact that German is a semi-null- subject and topic-drop language, see Hamann 1996, Roberts&Holmberg 2010) seem to amplify its effects. The second result, the overall similar performance of bilingual and monolingual children, may be attributed to the fact that these heritage children were L1-dominant unlike those in Montrul (2018). The fact that both, ours and Montrul’s heritage children, overused ambiguous NS, seems to indicate that this is an effect of bilingualism and possibly L2-influence.

ABSTRACT. The current study investigates the morphosyntactic processing of adjectival concord (in case and number) and subject-verb agreement (in person) by mono- and bilingual speakers of Russian. The focus of the study is on the potential factors that may trigger divergence in Heritage Language (HL) speakers from the baseline. HL speakers’ grammars undergo internal restructuring (e.g. Rothman, 2009, Benmamoun, Montrul, & Polinsky, 2013) under (a) transfer from the dominant societal language, SL (also known as cross-linguistic influence, CLI below) and/or (b) diminished input. In the study we considered effects of cross-linguistic influence, i.e., the influence of the dominant Societal Majority Language on HL, limited input (as indexed by Age of onset of bilingualism, AoB), and processing limitations due to working memory constraints in the morphosyntax processing. An auditory acceptability judgment task was performed by a total of 119 adult participants split into four groups: (1) Monolingual Russian-speaking controls residing in the former USSR (MonoControl), (2) Immigrant Controls, i.e. Russian-Hebrew bilinguals with AoB after the age of 13 (IMMControl); (3) bilinguals with AoB between 5-13 (BI-Late); and (4) bilinguals with AoB before the age of 5 (BI-Early). The latter group represents HL speakers, while BI-Late can be identified as 1.5 generation speakers. The AJT assessed the sensitivity to morphosyntactic violations (adjectival concord in case and number, subject-verb agreement in person). Russian and Hebrew show similarities and differences in the mapping and bundling of morphosyntactic features. In Russian, features of number, gender, and case of the head-noun are copied onto the modifying adjective, while in Hebrew the features of gender, number, and definiteness agree. As for the subject-verb agreement, in both languages, number, gender, and person features are copied from the subject to the agreeing verb. The three investigated phenomena are considered early-acquired phenomena in Russian-speaking monolinguals (Gvozdev, 1961). In order to manipulate the working memory load, all the agreement violations were tested in split and non-split constructions (see Stimuli examples in Appendix, Table 1). We did not find effects of cross-linguistic influence, as no difference was found for number concord as compared to case concord and person agreement on the verb, or extra memory load; at the same time, the effects of AoB were robust (see Accuracy scores in Appendix, Figure 1). Additionally, HL speakers (BI-Early group) differed from the other groups in poor performance on adjectival concord, but patterned with the others in person agreement. This observation confirms that this feature [person] is more robust than other agreement features in HL grammars. This difference can be attributed either to the different strengths of features that have to be matched (the person feature, being inherently indexical, is better maintained in bilingual grammars, see Benmamoun et al., 2013; Polinsky, 2018), or to structural differences between the clausal and nominal domain. Future work, where the same features are manipulated in both domains, is needed to tease these explanations apart; since only gender and number are matched in both domains, such work should concentrate on those features.

ABSTRACT. A wide range of tools have been used to assess the language proficiency of bilingual speakers, including heritage language speakers. There are subjective forms of assessment, such as self-reports or self-assessments, and objective forms of assessment, such as analysis of written production elicitations, samples of spontaneous spoken conversations, oral narratives, and standardized tests. Two types of tools were not that often combined. Although self-assessment ratings across single or multiple language domains are a widely used method, their accuracy, reliability, and usefulness have been debated (e.g., Brantmeier, 2006; Treffers-Daller, 2015). Thus, there is a need for indices that combine subjective and objective measures and allow for a more in-depth examination of the relationship between these measures (Treffers-Daller, & Korybski, 2015).

This study aims to contribute to this line of research by investigating the relationship between self-assessment data and language sample analysis. We focus on heritage language speakers of Italian that live in Istria County, a statutory bilingual county in Croatia. Italophone heritage language varieties in this area (the most prominent being Istrovenetian) date back several centuries and are still used in everyday spoken communication, while standard Italian is mainly used in formal contexts and education (Blagoni et al., 2016). The use of self-assessment and language sample analysis will allow us to obtain information about Istrovenetian, since other objective measures (e.g., standardized tests) are rarely used to assess spoken language varieties.

Measures of lexical diversity and syntactic complexity were used to examine whether they reflect the self-assessment of language proficiency of 30 functional Istrovenetian heritage speakers in four domains (reading, writing, speaking, and understanding). Participants’ data were extracted from the Istrian spoken multilingual corpus (Hržica et al., in preparation), which was created using the conversational sampling method. Measures of lexical diversity (Moving-average type-token ratio – MATTR; Covington, & McFall, 2010) and syntactic complexity (Mean length of communication unit – MLCU; Heilmann, 2010) were calculated using Computerized Language Analysis (CLAN – MacWhinney, 2000).

A Pearson correlation was calculated to test the relationship between self-assessment and language measures. The average self-assessment score correlated positively and moderately with MLCU (r(30) = .513, p < .005). MATTR did not correlate with the average self-assessment score, but the results indicate a trend suggesting this relationship (r(30) = .351, p = 0.57). A Spearman correlation was calculated for each domain of self-assessment. MATTR showed weak positive correlations with self-assessment of understanding (rs(30) = .364, p < .05) and speaking (rs(30) = .388, p < .05), but not with self-assessment of reading and writing. MLCU showed moderate positive correlations with self-assessment of reading (rs(30) =.484, p < .05) and writing (rs(30) = .563, p < .005), and weak positive correlations with speaking (rs(30) = .373, p < .05) and understanding (rs(30) = .364, p < .05). Descriptive statistics are presented in Table 1.

The correlations found between measures of lexical diversity and syntactic complexity and self-assessment scores indicate that these measures are related, suggesting that heritage language speakers may be able to objectively assess their language proficiency in spoken language variety.

ABSTRACT. Purpose: Statistics Canada (2017) showed that 1 in 5 Canadians are immigrants. Although Canada celebrates individuals’ diverse cultures and ethnicities, people who immigrate to Canada may still experience acculturation. Acculturation is the gradual process of adapting to a new cultural environment upon relocating from one’s heritage culture to a new and prevailing culture of the society in which one resides (Ward & Berry, 2001). Enculturation is when people maintain their heritage culture’s values and traditions. Immigrants usually integrate themselves into a new culture by adopting the host culture’s traditions and practices while still maintaining some of the values of their own heritage culture (Berry, 1994). Thus, many immigrants who come to Canada will speak more than two languages, their heritage language (L1) and English. Language dominance is a crucial factor to examined in terms of individuals’ proficiencies in the two languages as well. Looking at language dominance is essential as it discloses the bilingual learners’ proficiency in each language and ability in each language beyond subjective evaluations (Treffers-Daller, 2019). The present study aims to look at individuals’ heritage language proficiency in relation to their acculturation levels. The study also examines language dominance in youth and adolescents with immigrant backgrounds and their culture affiliation towards their heritage culture and the mainstream culture in which they reside, specifically Canada.

Methods: Undergraduate students who speak a language other than English as their first language (N=177, Females=124, Mage=19.73, SD=3.15) participated in the present study. The Family Language questionnaire, Bilingual Dominance Scale, Acculturation Rating Scale, Vocabulary Size Test, and the Verbal Fluency test were utilized in the present study. These measures assessed participants’ language dominance (L1 dominant, L2 (English) dominant, or balanced bilingual), self-reported L1 and English proficiencies, as well as cultural affiliation and identification.

Results: Preliminary analyses show that acculturation was related to lower self-reported L1 speaking r(176)=-.194, p=.010 and L1 writing skills, r(176)=-.212, p=.005. Language dominance was significantly different based on participants’ enculturation, F (2, 145)=14.12, p<.001, partial η2= .16, and participants’ acculturation, F (2, 146)=17.83, p<.001, partial η2=. 20. Enculturation was significantly higher between L1 dominant (M=55.31, SD=8.94) and balanced bilinguals (M= 49.48, SD=8.07) and between L1 dominant and L2 (English) dominant participants (M=46.34, SD=9.12). Furthermore, acculturation was lower for L1 dominant (M= 39.71, SD=7.36) as compared to L2 dominant speakers (M=46.31, SD=9.12) as well as between L2 dominant speakers and balanced bilinguals (M= 42.00, SD=6.03).

Conclusions: The findings suggest that acculturation is negatively associated with individuals’ L1 speaking and writing skills, implying that higher acculturation is related to worse self-reported L1 skills. In addition, those who are L1 dominant had higher enculturation as compared to balanced bilinguals and L2 dominant speakers. On the other hand, L2 dominant speakers had higher acculturation as compared to the other speakers. The findings of the present study highlight the link between enculturation and heritage language maintenance.

ABSTRACT. The Turkish accusative morpheme marks specificity and definiteness on direct objects. Nonspecific direct objects are marked with a null-form. Specific direct objects are marked with the suffix -(y)I (Türk & Caha, 2022) (1). Most scholars agree that Turkish applies Differential Object Marking (DOM) (von Heusinger & Kornfilt, 2005; von Heusinger et al., 2019).

Compared to the accusative, the Turkish dative has received comparably less attention in linguistic research. On the one hand, it is an adverbial case marker, marking the goal of an action. At the same time, it is the “oblique object” (Göksel & Kerslake 2005: 177) of verbs of emotion, and third, it is “the unmarked case for the arguments of transitive verbs that have non-macroroles” (Haig 1996: 46), for example when it marks the causee in causative constructions (see also Tonyalı 2015).

L1 acquisition studies of Turkish show that while the canonical use of the accusative is already established at 18 months in infants, the non-canonical of the dative continues in early childhood (Ketrez & Aksu-Koç, 2009). The few studies of case marking in Turkish in contact situations note conspicuous alternations between dative and accusative in children acquiring Turkish as a heritage language. They attribute this to transfer effects, language exposure and/or “incomplete acquisition” (Coşkun Kunduz & Montrul, 2022, Gökgöz et al., 2020). A closer look at the nature of the alternations is lacking, however.

In our talk, we take a broader view on the marking of accusative and dative in different varieties of Turkish. We investigate Turkish in Turkey spoken by monolinguals and by Kurmanji-Turkish bilinguals, as well as Turkish spoken in Germany and the U.S., including bilinguals as well as Turkish-Kurmanji-German trilingual speakers. Across all participants (n=226), data were elicited in two different modes (oral and written) and two settings (informal and formal) (https://hu.berlin/LangSit). Additionally, each participant was asked to fill in an extensive questionnaire that contained questions about participants’ educational background, the use of languages, their family background, etc.

A preliminary investigation (data from 188 participants) shows that in most non-canonical uses, speakers indeed alternate between dative and accusative. In particular, the accusative-marking DOM strategy is overgeneralized. We will investigate all non-canonical data sets in detail, inquire into their systematicity and ask which of the case functions appear to undergo change in what kind of grammatical or lexical contexts. Analyzing the data from different groups gives us an opportunity to explore the role of the contact language (German, English, Kurmanji) as well as the societal status of Turkish (heritage vs. majority) and register differences on the use of accusatives and datives. Finally, we correlate our findings with the metadata variables from the questionnaire, such as index of language use and exposure, and self-rated proficiency.

Our study sheds light both on the case grammar of Turkish as well as on dynamics of change of this system in contact situations. This will broaden our understanding of the role of extra-linguistic variables in heritage language maintenance as well as the development of Turkish in relation to its societal status and in different language contact scenarios.

ABSTRACT. This presentation considers the role of extra-linguistic variables in the loss of heritage Doukhobor Russian among Canadian Doukhobors (Spirit Wrestlers), a religious and ethnic minority group of mixed ethnic origins who immigrated to Canada in 1899 from the Russian Empire (Tarasoff, 1984). The presentation provides a brief outline of the unique structural features of the Doukhbor Russian (DR) variety and describes a research study which is aimed at analyzing the reasons for the variety discontinuation. The research question of the study is: What are the reasons for the decline of the variety, as they are seen by the community members? The materials come from interviews with 40 speakers of Doukhobor Russian (DR) who are bilingual in English and Doukhobor Russian. The interviews were recorded between 2012 and 2018 in British Columbia and Saskatchewan. The theoretical approaches as well as methods are rooted in sociolinguistic and ethnolinguistic studies (e.g., Brown & Bousquette, 2018) and involve manual thematic coding of the interview transcripts. The results demonstrate that, first, for most participants, there is a strong emotional connection with DR, which is still important for their childhood and early adulthood memories, molen’e (prayer services) and for understanding Doukhobor beliefs expressed in the psalms. On the other hand, about 1/3 of the participants see no merit in maintaining the language which is rapidly losing grounds, and they recommend switching to translations of the psalms and hymns into English, as the ideas are more important than the language. Second, the reasons for language discontinuation are described by the participants as being based on demographic weakening of the group, mixed marriages, lack of jobs in the region and young people leaving rural communities. The participants also mention severe discrimination by the colonial Canadian government in its strive to assimilate the non-conforming minority. Hundreds of Doukhobors (whose major belief is pacifism) were imprisoned for refusals to serve in the army during WWI and WWII. Other examples of discrimination include cancellations of Doukhobor land registries, attempts to confiscate their privately owned lands, and throwing them into prisons for protests. The most heinous action by the government against some of the Doukhobors in the 20th century was taking children away from the parents and placing them into juvenile detention centers until the age of 15 or into foster care. Discriminatory attitudes against the Doukhobors continue among general public to this day. The current political situation further contributes to a general negative perception of the Russian language. One more reason for the discontinuation of the heritage language is its perception as being inferior to Standard Russian. Our results agree with the previous evidence of discrimination against this ethnic, religious and linguistic minority group in Canada (e.g., Maloff, 1948; Tarasoff, 1984; Verigin, 2019). However, our study is the first one which connects the generational language shift with discrimination, colonialism, as well as with other demographic and socio-economic factors. In conclusion, the article outlines potential implications for language revitalization.

Selected References Brown, J. R., & Bousquette, J. (2018). Heritage Languages in North America: Sociolinguistic Approaches, Journal of Language Contact, 11(2), 201-207. doi: https://doi.org/10.1163/19552629-01102002

Maloff, P. (1948). Doukhobortsy, ikh istoriia, zhizn’ i bor’ba [Doukhobors, their history, life and struggles]. North Kildonan: J. Regehr.

Tarasoff, K. (1984). Plakun-Trava. Grand Forks: Mir.

Verigin, G. V. (2019). The Chronicles of Spirit Wrestlers’ Immigration to Canada: God is not in might, but in truth. Ed: Makarova, V. and L. Ewashen. Cham: Palgrave/Macmillan.

ABSTRACT. Adyghe, the Circassian language spoken in Northern Israel, is a Heritage language that is maintained by a unique community of immigrants from the Caucasus. These immigrants settled in two villages in the Galilee during the period of the Ottoman Empire, and for over five generations children born to this small community of speakers acquire Adyghe as a first language. At the same time, the majority of the community are proficient speaker-writers of Hebrew and Arabic, the two dominant Semitic languages of the area. As in other Heritage language communities (cf. Montrul, 2016), adult members of the Adyghe community are continuously raising voices of concern regarding their children’s proficiency in the language. In light of this concern, the current study set out to examine children’s command of morpho-syntactic features of Adyghe as an Ergative language, focusing particularly on agreement marking in the clausal domain. Specifically, we explored the production of number agreement markers. In Adyghe, number agreement on the verb is controlled by the subject as well as the object nominals, and interacts with the absolutive and ergative case markings (Lander & Testelets, 2017). For the purpose of the study we designed an experiment to elicit simple clauses, based on 30 short novel videos that depict different types of events, with varying numbers of participants either in the subject or the object position. Four children at the age of 4-6 years and their eight parents were asked to describe what is occurring in each video. The resulting corpus consisted of 720 clauses which were coded for the occurrence of number agreement. A total of 301 clauses were found to involve plural marking on the verb, controlled either by the subject or by the object nominals (see Appendix). The children’s data were compared to the adult productions using a Jaccard similarity coefficient test. Our results show that Adyghe-speaking children who are assumed to have already acquired the ergative structure (Stoll & Bavin, 2013) show command of the morpho-syntactic number agreement features of their language similarly to adults. Both children and adults produced plural markings on the verb controlled by subject as well as object nominals. When the plural marking was not produced, this was related to the occurrence of the absolutive case. This pattern of usage was more frequent among the children compared to the adults. Nonetheless, no significant differences were found between the children and their parents in this agreement domain. At the same time, some qualitative differences between these productions were found, that are compatible with the acquisition process of an Ergative language (Allen & Schröder, 2003; Austin, 2013). This is the first study to explore the acquisition of Adyghe. Our findings indicate that in contrast to the concerns expressed by the adult community, Adyghe-speaking children show typical patterns of usage of morpho-syntax in the clausal domain. We conclude by considering the implication of these results for the study of Heritage languages in general, and for Ergative languages in particular.

ABSTRACT. Background: Classifiers are morphological units that mark the class of nouns. In Mandarin Chinese, classifiers are obligatory when a counting of the noun is present (example 1). Importantly, Mandarin classifiers encode both lexical-semantic and lexicalized grammatical information, allowing us to examine the relative weighting of different linguistic informational cues in sentence processing. The sortal classifier, tiao, generally associates with nouns that conform to the description “slender, long-shape, often flexible” -- semantic association. For example, shengzi “rope” takes the classifer tiao. However, not all things that are “slender, long and flexible” take the same classifier because for any given “semantics” there can be a number of classifiers, which are idiosyncratically (lexically) associated with distinct nouns despite them sharing overlapping semantic features, for example, shoubiao “wristwatch” takes the classifer zhi. Thus, there is no one-to-one mapping between specific classifiers and potential candidates given a semantic class. Moreover, there are simply expectations. That is cases where the lexically associated classifier does not respect the general semantics with which it is associated. For example, gou “dog” co-occurs with the tiao classifier but is clearly not a long, slender, flexible object. Recent literature [1] using eye-tracking visual world paradigm showed that L1 Mandarin speakers rely more on the lexicalization of classifiers, while L2 speakers rely more on semantic information. In other words, L2 learners are more likely to accept/be distracted by tiao for both rope (correctly) and wristwatch (wrongly) and not for dog (wrongly) because they focus on the semantic association of the classifier to the object. The authors attributed the L1/L2 differences to L2 speakers’ needs in promoting processing efficiency such that each relies more on the cue which is a more reliable source of knowledge to them (the lexical entry for natives, the semantic tendencies for L2ers given weaker lexical connections). This current study examines how heritage speaker Mandarin bilinguals (HSs)—native speakers of Mandarin who are, however, dominant in a different (their societal) language—process Mandarin classifiers. Methods: We adopt the visual word eye-tracking task used in [1], targeting three Mandarin sortal classifiers, i.e., tiao, “∼long, flexible”; zhi, “∼stick-like, long”; and zhang, “∼flat, spread open”. Participants are shown visual stimuli with three objects (Figure 1): the target, the competitor, and a distractor, while listening to sentences (audio stimuli, example 2). For the target, we selected the noun, e.g., gou “dog”, that can co-occur with the classifier, e.g., tiao, (lexicalized grammatical match; G+), but does not belong to the prototypical semantic class the classifier is associated with, (semantic mismatch; S-). For the competitors, we manipulated the grammatical and semantic (mis)matching between the nouns and the classifiers, i.e., G-S- (pingguo “apple”), G-S+ (shoubiao “wristwatch”), and G+S+ (shengzi “rope). A total of 12 experimental trials, with 4 items in each of the 3 conditions were made and rotated across three lists. Audio stimuli were constructed by concatenating extracted tokens of na yi “which one”, the classifier, shi “is”, and the target noun. Preliminary results, predictions and discussions: 13 Mandarin-English HSs have participated in the study to date (50+ are reasonably expected by the time of presentation). So far, when the competitor(s) conform to the grammatical (G+) and/or the semantic (S+) constraints of the classifier(s), HSs direct their eye gaze less to the target (figure 2). This suggests as a whole group, HSs use both grammatical and semantic cues in processing classifiers. However, we expect individual differences, i.e., certain HSs are more like native homeland Chinese speakers while others diverge towards what is described by L2 learners, as a function of their bilingual experience. For example, HSs with more input/use in Mandarin will rely more on the grammatical cue, otherwise, on the semantic cue. With a significant sample size, we will regress composite scores for exposure and usage to see what, if any, exponents of bilingual experience explain individual variation in HSs.

ABSTRACT. This ongoing study investigates the acquisition and use of English as a foreign language in a primary school setting in Norway (grade 5, age 10-11). L2 learners of English (L1 Norwegian) are compared to L3 learners (heterogeneous group of heritage bilinguals in Norway) to examine whether there are differences between these two groups based on their English production. This research aims to answer which of the previously acquired languages is the source for cross-linguistic influence (CLI) by comparing L3 to L2 learners, partially in line with the subtractive language group design model (Westergaard et al. 2023). We add to the current debate of whether CLI occurs based on typological primacy (e.g., Rothman 2015), linguistic proximity (e.g., Westergaard et al. 2017), or further factors.

The grammatical properties tested pertain to adverb placement and V2 in (1) non-subject initial declaratives (e.g., Now the boy plays the guitar), and (2) subject-initial declaratives (e.g., The girl often builds the castle). English and Norwegian exhibit different word order patterns, which may potentially cause non-facilitation from Norwegian in the learners’ English production. The L3 learners’ heritage languages (HLs) vary and may either pattern with English ((1) ASV; (2) SAV) or Norwegian ((1) AVS; (2) SVA) or demonstrate a mixed pattern. For example, German patterns with Norwegian, and Russian with English. French and Finnish share one property with English (1) and the other with Norwegian (2).

There are three possible outcomes: (i) If CLI occurs based on typological primacy and if the majority language Norwegian is typologically closer to L3 English than the HL, we predict CLI from Norwegian only and no significant differences between L2 and L3 learners. (ii) If CLI occurs based on typological primacy and the HL is typologically closer, we predict CLI from the HL only. If the HL and Norwegian differ for the particular property, we expect significant differences between L2 and L3 learners for this condition, e.g., for both experimental conditions in HL Russian. (iii) If CLI occurs based on linguistic proximity, we expect CLI from both the HL and Norwegian depending on structural similarity.

Data collection is still ongoing. We collect two types of production data (sentence repetition, sentence ordering). Production data is currently underrepresented in L3 acquisition studies. The English sentence repetition task is administered in a one-on-one format and includes 48 short sentences, 6 grammatical and 6 ungrammatical per condition (1) and (2) (experimental conditions) and for wh-questions and determiner omission as filler conditions. The English sentence ordering task is a paper-pencil-task and includes 20 sentences overall, 10 per experimental condition with 5 grammatical and 5 ungrammatical sentences. In addition, we conduct a mini sentence ordering task in Norwegian and an oral version in the most frequent HLs. Furthermore, the study controls for a range of other confounds, such as proficiency in English (adapted version of the British Picture Vocabulary Scale (BPVS); Dunn et al. 2009) and background variables assessed via the Q-Bex questionnaire (De Cat et al. 2022).

  References

De Cat, C., Kašćelan, D., Prévost, P., Serratrice, L., Tuller, L., & Unsworth, S. (2022). Quantifying Bilingual EXperience (Q-BEx): questionnaire manual and documentation. DOI 10.17605/OSF.IO/V7EC8

Dunn, L. M., Dunn, D. M., Styles, B., & Sewell, J. (2009). Dunn, Lloyd M., Dunn, Douglas M. and Styles, Ben (2009). The British Picture Vocabulary Scale - Third Edition. London: GL Assessment.

Rothman, J. (2015). Linguistic and cognitive motivations for the Typological Primacy Model (TPM) of third language (L3) transfer: Timing of acquisition and proficiency consider aed. Bilingualism: Language and Cognition, 18, 179-190.

Westergaard, M., Mitrofanova, N., Mykhaylyk, R., & Rodina, Y. (2017). Crosslinguistic influence in the acquisition of a third language: The Linguistic Proximity Model. International Journal of Bilingualism, 21(6), 1-17.

Westergaard, M., Mitrofanova, N., Rodina, Y., & Slabakova, R. (2023). Full Transfer Potential in L3/Ln acquisition: Crosslinguistic influence as a property-by-property process. In Cabrelli, J., Chaouch-Orozco, A., González Alonso, J., Pereira Soares, S. M., Puig-Mayenco, E., & Rothman, J. (Eds.) The Cambridge Handbook of Third Language Acquisition and Processing. Cambridge University Press.

ABSTRACT. In this study, we present data from middle school students in Norway who are Norwegian-English bilinguals and who are beginner learners of French, Spanish or German as foreign languages. We will investigate whether English can exert either beneficial or detrimental influence on these three target languages. Listhaug, Busterud & Dahl (2021) have tested the influence of L2 English on word order phenomena in L3 French and found non-facilitative cross-linguistic influence (CLI) from L2 English when it comes to sentence adverbials. They also observed more CLI at lower levels of English proficiency. Stadt et al. (2018, 2020) on the other hand, who tested L1 Dutch - L2 English - L3 German/French, found more CLI from English to French at higher levels of L2 proficiency. They also observed more CLI from English to French than to German.

Our study adds to the existing body of research by comparing CLI from English in three different target languages: French, German and Spanish. Similar to the studies mentioned above, our target phenomenon is adverb placement, but we will focus on four different conditions, depending on the type of clause (main versus subordinate) and the type of verb (copula versus lexical verb), given that the five languages involved in our study behave differently from one another in all these conditions. Spanish always allows both word orders and therefore overlaps with all languages in all conditions, but French and German overlap with English and Norwegian for only some of the conditions (see table 1).

In addition to the question whether we observe influence from English in these conditions, we are also interested in examining which, if any, extra-linguistic variables such as L2 proficiency/dominance, exposure and use, best predict the extent of the effect. We are particularly interested in the role of age of onset of English, which is why we compare children who acquired English from birth as a heritage language, to children who became exposed to English later in childhood as an early L2.

Data collection is still ongoing. We test the above mentioned structures using a grammaticality judgment task and a self paced reading task, which is administered online during school hours. The tasks contain grammatical and ungrammatical sentences as in example (1) (from condition 3). In addition, we measure the participants’ proficiency level in Norwegian, English and the target language using the PPVT (Dunn & Dunn, 2007), as well as their knowledge of the target structure in Norwegian and English.

References: Dunn, L. M., & Dunn, D. M. (2007). PPVT-4: Peabody picture vocabulary test. Pearson Assessments. Listhaug, K. F., Busterud, G., & Dahl, A. (2021). French as a third language in Norway: The influence of the prior languages in the acquisition of word order. Oslo Studies in Language, 12(1), 125-144. Stadt, R., Hulk, A., & Sleeman, P. (2018). The influence of L1 Dutch and L2 English on L3 French: A longitudinal study. Journal of the European Second Language Association, 2(1). Stadt, R., Hulk, A., & Sleeman, P. (2020). L2 influence in L3 acquisition: The role of the L3. In Formal Linguistics and Language Education (pp. 203-221). Springer, Cham.

Example (1): French: Ma sœur Julia, qui est toujours heureuse, rit tout le temps au travail. German: Meine Schwester Julia, die immer glücklich ist, lacht die ganze Zeit bei der Arbeit. Spanish: Mi hermana Julia, que (siempre) está (siempre) feliz, se ríe todo el tiempo en el trabajo. Norwegian: Søstera mi Julia, som alltid er glad, ler hele tiden på jobb. English: My sister Julia, who is always happy, laughs all the time at work.

ABSTRACT. Heritage speaker (HS) bilinguals are native speakers of their heritage language (cf. Rothman & Treffers-Daller, 2014). Like homeland speakers of the same language, HSs acquire their heritage language (HL) early and naturalistically. Differently, however, they often do so in a context of significantly reduced input and/or opportunities (over the lifespan) to use the language in a comparatively similar way. It is no surprise then that a substantial amount of research over the past decades has documented significant differences between HS and homeland native speakers (Montrul, 2016, Polinsky, 2018; Polinsky & Scontras, 2020) across a wide array of grammatical domains. Whether or not this is equally true in linguistic processing is not as clear, given the disproportionate number of studies utilizing offline behavioral methods. For example, behavioral HS studies would conclude that grammatical gender is vulnerable in Spanish as an HL in the context of the United States (US) (e.g., Montrul, Foote & Perpiñan, 2008). However, will neural signatures during grammatical processing confirm or problematize such a conclusion (see Bayram et al. 2022)? Herein, we seek to fill the gap in neurolinguistic methods applied to HL processing in an otherwise well-studied domain of grammar. To do this, in this study, we focused on grammatical gender processing in Spanish, which has been well researched at both the behavioral and neural (EEG/ERP) levels across homeland native speakers and L2 learners, but only behaviorally in HSs (no neuroimaging studies exist). Given the significant number of ERP studies on grammatical gender with Spanish homeland natives, which converge on showing at least a robust P600, the brain response typically associated with morphosyntactic processing (see Morgan-Short, 2014), for gender agreement anomalies, we have a solid basis of comparison for the current study. A total of 44 US Spanish-English HSs completed a behavioral gender assessment task and a grammaticality judgment task (GJT) while their brain activity was recorded using EEG. The behavioral gender assignment task was aimed to examine each participant’s assigned gender value for the complete set of nouns used in the EEG/ERP experiment. The EEG/ERP GJT task comprised grammatical and ungrammatical sentences that included gender violations, where morphological transparency was manipulated. Our results indicated that, at the neural level, gender violations elicited a significant P600 effect (p=< .001**) across all relevant conditions and, crucially, that transparency matters during online gender processing, with transparent nouns eliciting a more positive response than opaque ones (see Figure 1). Notably, the results from our study provide converging behavioral and brain evidence revealing that our HSs’ grammatical representations, as well as processing strategies for gender, are qualitatively similar to that of other types of Spanish native speakers contrary to what has been often found in the behavioral literature for HSs’ grammatical gender processing thus far. Crucially, our results highlight the importance of and need to incorporate online processing methods, such as EEG, to examine both the universality and diversity behind HL processing in a multidimensional, fine-grained and ecologically valid manner not only for gender but across all domains of HL grammar(s).

ABSTRACT. The study is a part of a large project on the investigation of the language interaction in Southern Siberia. The project has collected a corpus of texts from native speakers of Turkic languages (Tatar, Shor, Khakass). The respondents’ native languages can be defined as heritage languages [1], [2], which is confirmed by the data of the questionnaire on the linguistic and social experience. Data from two questionnaires were used: the sociolinguistic questionnaire [3: 221–231] and the bilingual language experience questionnaire [4: 940–967]. This type of linguistic existence of modern Shor, Siberian Tatar and Khakas bilinguals is characterized by the displacement of the native language by the Russian language, which dominates in all spheres of communication. Such a functional distribution of languages determines a relatively small number of cases of interference. [5], [6]. Material and Method The corpus includes morphologically annotated transcripts of more than 400 hours of speech in Russian by Mystem (Yandex programme). Also, 40 hours of speech are annotated manually by deviations from the language standard (DLS). Respondents differed in their native language, the age group (from 23 to 81) and the level of education. Further, discursive deviations (Disc)– hesitations, formal-semantic and functional incompleteness of statements; syntactic (Synt) – different types of violations of syntactic coherence and constructive correctness of statements and deviations in the field of morphology (Morph) and vocabulary (Lex) in Russian speech were contrasted. The data was analyzed using quantitative and qualificative methods in R-Studio. Results We examined the influence of the heritage bilingualism on Russian language usage and whether age and education influence on the deviations. The analysis showed, that the deviations in the field of vocabulary and morphology are much less frequent: Median – Disc 0.050; Synt 0.020 VS Lex 0.004; Morph 0.001. Correlation analysis revealed a positive correlation between morphological and syntactic (0.534), morphological and lexical (0.602) DLS. The Kruskal–Wallis test showed statistically significant differences only in the syntax DLS between the age groups (picture 1, p = 0. 04) and the education level (picture 2, p = 0. 02) of the respondents. Qualificative analysis showed: 1) errors in the use of the grammatical gender prevail in all the syntactic positions, which is due to its absence in the native Turkic languages of the bilinguals; 2) differences in the structure of the grammatical number of the languages resulted in the prevalence of the deviations from the language standard using the number forms of the material and collective nouns; 3) among the deviations in the field of case, substitutions quantitatively dominate: nominative case → prepositional case, nominative → accusative; dative → accusative. These differences are predetermined by differences in the formal-semantic structure of the category, in the system of oppositions of particular grammatical meanings, in the ways of their formal marking, and in the interactions with other grammatical and lexico-grammatical categories. This study was supported by the Tomsk State University Development Programme (Priority-2030)

ABSTRACT. Due to increasing global mobility, research into how immigrants acquire and use multiple languages has become a major topic in the past few decades. However, comparatively little is known about how the developmental trajectories of heritage languages (HLs) are shaped alongside the societal languages in their children who are heritage speakers (HSs). CLICK is the first study to investigate how different societal languages leave their traces on the developmental trajectories of the same HL. This was done by studying HSs of Persian/Farsi in contact with different societally dominant languages.

CLICK explores whether or not the syntactic status of pronominal subjects (null and overt subjects in Persian and Spanish versus overt subjects in French and English) affects the development and maintenance of related properties in Persian as a HL. In doing so, it focusses on the anaphoric interpretation of null and overt pronouns , which requires the processing of grammatical and discourse information to be coordinated. The motivation is rooted in the fact that interfaces, especially external ones such as syntax-discourse, have been found to be vulnerable domains in language development and loss. 1,2 Sixty stimuli were developed in Persian with different type of anaphora (forward, backward and mid) with overt and null pronouns and while the participants were listening to the stimuli, their eye movements across two pictures were monitored.

1) The mother kisses her daughter, while she/pro is singing a song. (forward) 2) While she/pro is singing a song, the mother kisses her daughter. (backward) 3) The mother, while she/pro is singing a song, kisses her daughter. (mid)

For instance, in above stimuli, they saw two pictures in both the mother kisses the daughter, while in one the mother is singing, and in the other the daughter is singing. In addition, since this study strives to combine online and offline measures in the same task to better capture underlying grammars of HSs in different language contact situations, their offline choices of pictures after listening to the stimuli were reordered. Twenty-two HSs of Persian with different societal languages (including French, Spanish, English, German and Norwegian) as well as their own mothers participated in the study. These HSs were all bilinguals who were born or immigrated to their host countries at a young age, while their mothers were all grew up in Iran and emigrated during adulthood.

The preliminary analyses show that Position of Antecedent Strategy (PAS) is confirmed by the HSs and their mothers, according to which the null pronoun prefers an antecedent in the canonical subject position, while the overt pronoun prefers an antecedent which is not in canonical subject position.1 In addition, the HSs and mothers do not seem to differ in their eye movements and offline results in different anaphora apart from mid anaphora. This may have been because this type of anaphora is common in Persian for mothers while the HSs may have had more difficulty processing this grammatical structure.

ABSTRACT. It has been argued that certain words can “trigger” intrasentential code-switching (de Bot et al., 2009). While some researchers suggest that cognates establish triggering at the lexical level (Broersma, 2009), others have argued that words that lack direct translations are “more natural to switch” (González-Vilbazo et al., 2013, p. 8). Yet to be tested experimentally is to what extent lexical items influence the acceptability of mixed utterances. The current study investigates this methodological consideration for code-switching research by having early US Spanish-English bilinguals complete an acceptability judgment task with a 7-point Likert scale, directly comparing cognates (e.g., sopa ‘soup’) and culturally specific items (e.g., pozole ‘traditional Mexican soup’) in otherwise identical grammatical switched sentences (N=24). Data collection is still ongoing, but preliminary results (after standardizing the ratings into z-scores) suggest that culturally specific Spanish words receive the higher (more acceptable) ratings (M=0.87) than Spanish cognates (M=0.75), whereas the ratings in English are comparable for both the former (M=0.62) and the later (M=0.60). These findings show that judgment tasks can be affected by lexical item choice, but the effect can be asymmetrical, at least in the case of Spanish-English bilinguals in the US. The reason, we suspect, is that culturally specific terms are held on to strongly for questions of identity (Callesano, 2020).

ABSTRACT. Since code-switching (CS) is rule-governed, it can uncover patterns present in bilingual grammars that remain hidden in unilingual speech. Despite extensive research, few uncontested rules have been identified to explain the regularities underlying mixed speech. To a great extent, the lack of consensus is due to existing data originating from non-comparable methodologies and populations. In this respect, community norms are an understudied factor influencing CS patterns. The fact that exposure to and use of a bilingual structure can influence speakers in production and intuition underscores the importance of considering context-specific CS norms, in line with theories supporting the emergent nature of CS grammars (see Backus 2015). The main research question of this study therefore revolves around community norms: can we identify pluralization patterns of code-switched nouns that are unique to one bilingual community (e.g. family, friend group, school) when compared to other bilinguals of the same language pair? If so, the findings may support a usage-based approach to CS. Predictions of restrictive or universalist approaches, such as the Matrix Language Frame (Myers-Scotton 1993), Phonetic Form Interface Condition (MacSwan & Colina 2014), Distributed Morphology (cf. Grimstad et al. 2014), i.a. are also tested. The languages of the bilinguals involved (Hungarian-Dutch, Hungarian-German, Hungarian-English) use different pluralization systems: English chooses a suffix (/-z/, /-s/, /-ɪz/) based on phonological considerations (Bauer et al. 2015), Dutch plural suffixes (-s, -en) depend mostly on prosody (Booij 2002), German uses suffixes (-er, -(e)n, -e, -s, -Ø), sometimes combined with Umlaut (Hentschel & Weydt 1994), and Hungarian plurals end in -k, preceded by an epenthetic vowel following vowel harmony after nouns ending in a consonant (Kenesei et al. 1998). Muysken (2000) identifies three ways to incorporate plural nouns into another language: using the plural marker of the language of the noun, using the marker of the matrix language, and using double pluralization with markers from both languages. All these patterns occur when bilinguals insert Dutch nouns into Papiamento sentences: muiz-en ‘mouse-PL’, muis-nan ‘mouse-PL’, muiz-en-nan ‘mouse-PL-PL’ (Muysken et al. 1996). In the current study, the latter two patterns would present additional research questions regarding phonology and prosody, bringing about more environments in which CS norms may emerge, e.g. how does Hungarian vowel harmony work in Dutch nouns containing vowels not present in Hungarian? Three methods are employed to cross-validate the different types of data (cf. Toribio 2018): conversational data, elicited production data through a sentence completion task, and intuitional data through acceptability judgements. Data will be gathered from English-Hungarian, German-Hungarian, and Dutch-Hungarian bilinguals in Hungary, Austria and the Netherlands (estimated 25 participants per language pair per country). Members of the same family, friend group, and (ex)-students of the same heritage school will be targeted in order to uncover community norms. It is expected that the results provide a more fine-grained insight into CS grammars by testing the strength, validity and accuracy of theoretical accounts of CS and in particular by taking into account CS norms that may emerge in bilingual communities.

ABSTRACT. This paper reports on grammatical gender assignment in code-switched noun phrases in heritage speakers of Turkish in the Netherlands. Given that Dutch is a gendered language that distinguishes between common and neuter gender, when a Turkish noun is inserted into a Dutch DP, gender needs to be assigned to it. Several possible strategies have been identified in the literature. In this study we aim to distinguish between the default strategy, by which bilinguals use a single gender value to assign all nouns (e.g., Clyne and Pauwels 2013), and the analogical gender criterion, whereby bilinguals assign the gender of the translation equivalent of the noun (e.g., Jake, Myers-Scotton & Gross 2002). The research questions were:

1) Which code switching strategy is used mostly by Turkish HSs in the Netherlands? 2) Are the results modulated by extralinguistic factors, such as age of onset, dominance in Dutch etc.? 3) Are the results dependent on the type of task

For the first study, 22 HSs of Turkish participated in a director-matcher task (Gullberg, Indefrey &Muysken 2009), which was specifically designed to elicit noun phrases consisting of a determiner, noun and adjective. The task was performed both in a unilingual Dutch mode and in a code-switching mode where Turkish nouns were inserted into Dutch. For the second study, 29 members of a different Turkish community carried out an online auditory judgment task. This group consisted of both first-generation immigrants and second-generation immigrants, or heritage speakers.

For both studies, participants filled out an extensive background questionnaire, containing questions about input, use, age of onset, self-reported proficiency, etc. The questionnaire for the second study also contained questions about the participants’ code-switching habits. The results for the two different studies show different patterns. The production data from study 1 revealed that most participants used common gender as a default. However, the data from the unilingual mode showed that many participants also overextended common gender when producing Dutch DPs (cf. Cornips, 2008). Those participants who were most accurate with Dutch gender in the unilingual mode, also used the analogical gender criterion more often when they code-switched. The judgment data from study 2 do not show any evidence for the analogical gender criterion; only the common gender default is observed, and only in second generation heritage speakers. The first-generation immigrants do not exhibit any clear gender assignment pattern at all, neither in code switching nor in unilingual Dutch mode. The second study also revealed an effect of code-switching frequency such that more frequent code-switchers were more prone to use the common default strategy (cf Królikowska et al. 2019). An additional and unanticipated finding was that, both in judgment and production, heritage speakers preferred indefinite DPs to definite DPs, which potentially indicates cross-linguistic influence from Turkish, which lacks a definite determiner.

Clyne, M., & Pauwels. A. (2013). The Dutch language in Australia. In F. Hinskens & J. Taeldeman (eds.), Language and Space. An International Handbook on Linguistic Variation. Volume 3: Dutch, pp. 858-79. Berlin/Boston: De Gruyter Mouton.

Cornips, Leonie. (2008). Loosing grammatical gender in Dutch: The result of bilingual acquisition and/or an act of identity? International Journal ofBilingualism 12: 105–24.

Jake, J. L., Myers-Scotton. C., Gross. S. (2002). Making a minimalist approach to codeswitching work: Adding the Matrix Language. Bilingualism, Language and Cognition 5: 69–91

Królikowska, Marta Anna, Emma Bierings, Anne L. Beatty-Martínez, Christian Navarro-Torres, Paola E. Dussias, and M. Carmen Parafita Couto. (2019). Gender assignment strategies within the bilingual determiner phrase: Four Spanish-English communities examined. Paper presented at 3rd Conference on Bilingualism in the Hispanic and Lusophone World (BHL), Leiden, The Netherlands, January 9–11

ABSTRACT. This study targets bilingual (Romanian-English) heritage speakers of Romanian (HS), growing up in Toronto, Canada (16 children, 8-11 years old). Although the children learned Romanian as their L1 and use it at home, they are growing up in an English-dominant environment. We know that HS, due to less language exposure and use, are more likely to exhibit language loss or attrition than older bilingual children who had a longer period of monolingualism in their heritage language (Montrul 2008; Flores 2010; Montrul and Bateman, 2020). We also know that the domain of pronominal object clitics is vulnerable to reduced language input and use (Pirvulescu et al. 2014); at the same time, Romanian is a language where these clitics are acquired very early and with minimal errors (e.g. Tomescu and Avram, 2019). In this context, we ask the following questions: 1) do HS of Romanian use object clitics in their production and if yes, are they using the correct form? 2) do HS of Romanian correctly interpret clitic features? 3) How is the use and interpretation of clitic (features) related to the variables Working Memory and Use of Romanian at home? We measured two modalities (comprehension and production) as it is well known that heritage speakers’ language proficiency and accuracy depend on task modality (Pérez-Cortés et al. 2019). We compare results from a Clitic Elicitations Task and a Comprehension Task (picture choice) aimed at testing sensitivity to phi-features (gender and number) in clitics. In the Comprehension Task, half of the items are direct object clitic pronouns, and half are strong pronouns. We administered a Working Memory Test, and a language questionnaire provided information about the amount of input and use. We compare the HS group with a group of multilingual children with L1 Romanian as dominant language from Bucharest, Romania. Tables 1 and 2 show that while Romanian HS perform extremely well, they do perform differently from the Romanian Dominant children; the latter group perform at or near ceiling while Romanian HS children show, comparatively, lower clitic production, with clitic omission and errors of gender being the main divergence. The HS group displays errors in production as well as in comprehension (no statistically significant difference between the two). The variable Use of Romanian at home significantly predicts (correct) clitic production (F(2,13) = 3.79, R2 = .369, p = .017). Both variables Working Memory and Use of Romanian at home significantly predict correct clitic comprehension (F(2,13) = 10.13, R2 = .609, p < .05). For the HS group there is a significant difference, in the comprehension task, between the correct clitic vs strong pronouns responses, children being much more accurate with strong pronouns than with clitics (t(15) = -4.563, p < .001). These results confirm previous results on age effects in heritage language acquisition and at the same time, even though results are similar across tasks, they seem to indicate that errors do arise from computational difficulties and difficulties in accessing linguistic representations due to cognitive load (e.g. Pérez-Cortés et al. 2019).

ABSTRACT. Timing the acquisition of Differential Object Marking in Child Heritage Romanian

Reduced input in a heritage language (HL) during the school age period has been argued to be the possible cause of delayed or incomplete HL acquisition (Montrul 2016, 2000, Polinsky 2018, a.m.o.). The present study looks into the possible effect of input reduction after school entry on the acquisition of differential object marking (DOM) in child heritage Romanian (CHR) in a context in which the societal language is French. Romanian has morphologically marked DOM (the preposition pe and clitic doubling), obligatory with definite pronouns and animate proper names (1), syntactically optional with animate (in)definite descriptive DPs (2). Marking of inanimate objects is generally ungrammatical (3-4), but there are syntactic contexts in which the animacy constraint is overridden and DOM is either allowed (5) or even obligatory (6) with inanimate objects. In order to evaluate if the change in input after school entry affects knowledge of DOM, we used an acceptability judgment task which tested DOM in obligatory, optional and incorrect contexts (see 7). 45 CHSs of Romanian (7-, 9- and 11-year-olds), born to Romanian families living in France, took part in the study. They all began kindergarten at approx. age 4-5. Their responses were compared to those of 45 age-matched Romanian monolinguals living in the homeland. The raw data indicate an overall slight progress from age 7 to age 11 but the change does not reach significance wrt either DOM in obligatory (F(2, 42) = 1.53, p = .23)), optional (F(2,42) = 0.67, p = .51), or with inanimate DPs (F (2, 42) = 1.97, p = .15)). The comparison with age-matched monolinguals living in the homeland revealed a similar acquisition pattern and no significant effect of group at age 7 (F(1,28) = .520, p = .48). The difference becomes significant at age 9, however, when the CHSs accept unmarked animate and marked inanimate objects at a higher rate than the monolinguals (t(28) = -3.37, p = .002)); t(28) = 4.69, p < .001). At age 11, DOM in obligatory contexts is similar in the two groups but only the CHSs continue to accept DOM with inanimate objects and their acceptance rate for DOM in optional contexts remains lower than the one attested with monolinguals (t(28) = -3.21, p = .003)). These findings indicate that the change in input after schooling begins does affect DOM in CHR and identifies age 9 as the likely “peak” of this effect. At age 9, the HSs differ from the monolinguals, but at age 11 the two groups perform alike wrt to obligatory DOM. DOM in optional contexts and with inanimate objects continues to be slightly different in the two groups. The data support the view according to which narrowly syntactic properties are not subject to input amount effects, whereas properties whose acquisition requires the integration of syntax with semantics, discourse constraints, etc., can be affected by amount of input in bilingual acquisition (Tsimpli 2014). They are delayed in L1 but even more delayed in a learning setting with reduced input.

(1) *(Îl) vizitam *(pe) Vasile/el. cl visited PE Vasile he ‘I visited Vasile/him.’

(2) (Îl) vizitam (pe) (un) vecin. cl visited PE neighbour ‘I visited the/a neighbour.’

(3) *(O) vizitam (*pe) Londra. cl visited PE Londra ‘I visited London.’

(4) (L-) am desenat (*pe) pom. cl have drawn PE tree ‘I drew the tree.’

(5) DOM with inanimate DP: an upgrading effect […] le crezi pe topuri/crezi topurile CL believe PE tops ‘you trust tops’

(6) obligatory DOM with demonstratives irrespective of animacy *(L-) am desenat *(pe) acela de acolo. CL have drawn PE that of there ‘I have drawn the one over there.’

(7) Grammaticality Judgment Task: two aliens produce one a sentence with a DOM-ed object and one with an unmarked object. The participant is required to decide “which alien said it better”. The test sentences included:

Obligatory DOM DOM with personal pronouns x 4 DOM with animate proper names x 4 Incorrect DOM DOM with inanimate proper names x 4 DOM with inanimate descriptive DPs x 4 Optional DOM Dom with definite animate descriptive DPs x 4

References

Montrul, Silvina. 2016. The Acquisition of Heritage Languages. Cambridge University Press. Polinsky, Maria. 2018. Heritage Languages and their Speakers. Cambridge University Press. Tsimpli, Ianthi Maria. 2014. Early, late of very late? Timing acquisition and bilingualism. Linguistic Approaches to Bilingualism 4:3. 293-313.

ABSTRACT. Discourse openings and closings are defined as the linguistic material that precedes or follows the core phase of the discourse (see Schegloff 1986, 2007, Schegloff & Sacks 1973 for English; Luke & Pavlidou 2002, Pavlidou 2014 for other languages). Their typical functions include defining textual boundaries that set the core of the text apart from the framing, introducing or closing off a new topic, orienting the addressee to what is coming next, making clear the discourse stance of the speaker, and interpreting and evaluating the material (Berman & Katzenberger 2004, Labov 1972; Tolchinsky et al. 2002). The construction of discourse borders through openings and closings (Schegloff 2007) has been shown to be an area of dynamic difference across languages (e.g., Luke & Pavlidou 2002), but has rarely been studied in situations of language contact (but see Dollnick & Pfaff 2013). We aim to fill this gap in the literature by investigating discourse openings and closings in Turkish heritage speakers in the US with a focus on speakers’ age (adults vs. adolescents), register (formal vs. informal), modality (spoken vs. written texts) and language setting (monolinguals vs. heritage speakers in Germany and US). Our data are based on a corpus analysis of the RUEG corpus (https://hu.berlin/rueg-corpus). The data were elicited using the Language situations set-up (https://hu.berlin/LangSit) yielding different communicative situation with regard to formal and informal registers. Our analysis includes elements produced at discourse openings and closings in the RUEG texts. In total, we analyzed the oral productions of 66 monolingually raised Turkish speakers, 55 heritage speakers of Turkish residing in the US, and 64 monolingually raised English speakers in the US. Our analysis includes a quantitative and qualitative analysis of textual, subjective and intersubjective functions in openings and closings in heritage speakers’ heritage and majority languages in comparison with majority Turkish and majority English. The results of the quantitative analysis show that all groups of speakers use openings and closings in their formal and informal narrations across both spoken and written modalities. There are differences, however, with respect to the distribution of discourse functions in openings and closings. In their majority language (US English) the Turkish heritage speakers pattern a lot with the monolingual US English speakers. In the speakers’ heritage language, however, we see different distribution of discourse functions between heritage speakers and monolingual Turkish speakers. We discuss our results in relation to heritage speakers emerging patterns and possible cross-linguistic influence.

ABSTRACT. There exists a great degree of variability in the documentation of multilingual experience across different instruments. The present paper contributes to the "methods turn" and individual differences focus in (heritage) bilingualism by proposing a comprehensive online questionnaire based on existing questionnaires and the experience of using them to document heritage bilingualism: the Heritage Language Experience (HeLEx) online questionnaire. We validate HeLEx against an extended version of the Language and Social Background Questionnaire designed for heritage speakers (HSs), LSBQ-H, by comparing data elicited with both questionnaires in turn from a group of Turkish HSs (n=174, mean age=32). Our validation focuses on traditional language background variables, including language exposure and use, language proficiency, language dominance, as well as a more novel measure of language entropy. The analyses are based on a subset of key questions from each questionnaire that capture language experience for up to five languages, four modalities, and five social contexts. In a second set of analyses, we explore the impact of different types of response scales, response mechanisms, and manners of variable derivation (including language entropy) on the informativity of the data they can provide, in terms of the scope, granularity and distributional properties of the derived measures. Our results show that both HeLEx and LSBQ-H are successful at detecting the important distributional patterns in the data and reveal a number of advantages of the HeLEx. In the discussion, we consider the impact of methodological choices regarding question phrasing, visual format, response options, and response mechanisms. We emphasize that these choices are not trivial and can affect the derived measures and subsequent analyses on the impact of individual differences on language processing and outcomes.

ABSTRACT. In the field of heritage language bilingualism there has been a focus on studying heritage language grammar. However, according to Polinsky & Scontras, 2020 ,“the time is ripe to subject heritage speakers to the same range of studies common to investigations in traditional psycholinguistics and second language acquisition”, and there has recently been an increasing interest in studying heritage language processing using online methods (Bayram et al., 2020). This study aimed to look at Anticipation and Interference as two domains of sentence processing in Russian heritage speakers. We define Anticipation as pre-activation of the expected input given (sentential) context (Altmann & Kamide, 1999)) and Interference as competition from contextually irrelevant information (Swinney, 1979; Kukona et al., 2014). To give a few examples, people heard sentences like (1) The woman will sign the postcard (Женщина подпишет открытку) or (2) The grandfather will smoke the black pipe (Дедушка выкурит черную трубку), while observing a visual array consisting of four objects (a Visual World Paradigm). By counterbalancing constraining and non-constraining verbs, we used sentences like 1 to examine Anticipation. By manipulating the color of the target and distractor items in the visual display, we used sentences like 2 to examine Interference. Additionally, and going beyond the original design on Kukona et al., 2014, we manipulated grammatical gender of targets and distractors in the stimuli tapping into Interference, i.e. the target and the distractor either matched or mismatched in grammatical gender. This allowed us to examine the role of lexical and grammatical information in guiding sentence processing in heritage speakers. Moreover, some previous work (Nozari et al., 2016) suggests that the ability to inhibit irrelevant information (e.g. inhibiting looks to the black hat when hearing The grandfather will smoke the black pipe) varies as a function of individual differences in inhibitory control. To examine this, we gave out participants a Flanker task tapping into the ability to suppress irrelevant responses. We also collected detailed information on participants’ language use background (LHQ3, Li et al., 2020) as well as asked them to perform vocabulary and grammatical gender knowledge tasks. Having such a comprehensive battery of tasks will allow us to investigate the impact of individual variation in language knowledge, language use and domain general cognitive control on sentence processing in Russian heritage speakers. We are currently in the process of data collection, however since this is an online study (in a sense that we utilized webcam-based eyetracking (see Vos et al., 2022 who tested the efficacy of this method), we are confident that we will be able to collect enough data in time for the conference.

ABSTRACT. How does learning a second language (L2) affect our cognition? Grammatical gender (GG) has been used to demonstrate that arbitrary linguistic categories can influence the ways in which speakers conceptualize inanimate objects (e.g., Samuel et al., 2019). The available evidence, however, seems too scarce and mixed to answer the question of whether and how learning a GG language as an L2 affects bilingual speakers. Whereas some studies reported that, after learning a GG language, native speakers of genderless languages conceptualize objects in a manner that is consistent with the L2 GG (e.g., Kurinski et al., 2016), other studies showed that L2 GG does not affect conceptualization in bilinguals (Lambelet, 2016). This study thus examined the implicit and explicit effects of GG on object conceptualization in native speakers of Turkish (genderless language), French (GG language), and Turkish-French bilinguals. Participants first completed the Implicit Association Test (IAT), in which they were presented with pictures of human faces and objects that differed in their conceptual gender (CG; female or male) and French GG (feminine or masculine). The task instruction did not mention CG nor GG, and participants were asked to classify the faces as either female or male and the objects as either tools or clothing items. Participants also completed the Explicit Gender Attribution Task (EGAT), in which they were asked to choose the gender associated with each object. Finally, individual differences in “sexist” attitudes were assessed using the Gender Role Attitudes Scale (García-Cueto et al. 2015). Turkish speakers’ conceptualization appeared to be affected only by CG: they made more mistakes in IAT when CG mismatched the gender of the face and their EGAT assignments were also consistent with CG (ps < .001). Turkish speakers with relatively high sexist attitudes were affected more by CG in IAT (p = .014). French speakers, in contrast, made more mistakes when either GG (p = .024) or CG (p < .001) mismatched the face gender, and high sexist attitudes meant stronger effects of both GG (p = .007) and CG (p = .032) in IAT. In EGAT, CG (p < .001), but not GG (p = .056), significantly predicted the gender assignments, and the effect of CG was stronger for more sexist French speakers (p < .001), whereas the effect of GG was more pronounced for those who with less sexist attitudes (p = .028). The IAT response patterns of Turkish-French bilinguals were predicted by CG only (p < .001), but both GG and CG predicted their gender assignments in EGAT (ps < .001). Their EGAT responses were predicted more by GG in bilinguals with weaker sexist attitudes and by CG in bilinguals with stronger sexist attitudes (ps < .001). Overall, our findings suggest that (1) GG can intervene with the object conceptualization in native GG language speakers even when GG is irrelevant to the task, but (2) learning a GG language as L2 may only affect the explicit object conceptualization, and the GG effect is not as automatic and rapid for sequential bilinguals.

ABSTRACT. This study investigates Turkish community centers in the United States. These centers are usually used as places of worship, cultural centers, and/or heritage language programs for immigrant children. This study aims to provide a snapshot of Turkish heritage language programs by collecting information from administrators and/or teachers using survey questions about organizational, teacher-related, student-related, and parental issues. Administrators and/or teachers from 12 different centers responded to the online survey. Based on the respondents' comments, this study demonstrates that Turkish heritage language programs overwhelmingly function within the informal cultural context on a voluntary basis and are not integrated into the formal education system. Although the center leaders expend effort to attract kids and parents, enrollments remain low and concentrated among primary school-age children. These centers are overwhelmingly funded by the community, the Turkish state, and private organizations and lack support from U.S. institutions. While some centers prioritize religious education, others give only cultural education. They generally offer few classes, employ teachers with minimal experience, and do not sufficiently collaborate with each other. This study concludes that community centers play a critical role in maintaining heritage languages since they are the only spaces for Turkish children’s professional language development. They provide immigrant families and children with complex social, cultural, and educational experiences. The field needs further study that explores the role of these complex interactions in heritage language proficiency.

ABSTRACT. Only a handful of studies have been concerned with global foreign accent (GFA) in bilingual children [1], and little is known on how GFA develops over time. In this study, we examined the longitudinal development of GFA in bilingual returnees. Returnees are children of immigrant families who spend a significant portion of their formative developmental years (school age) in an L2 majority language (ML) context yet return to their L1 environment as older children or teenagers. During their stay abroad, they are exposed to the ML of the host country and typically acquire this an early L2. At the same time, their L1 becomes a heritage language (HL) which they are only exposed to in the home environment. Yet, upon return to their ‘home country’, this linguistic environment reverses: their L1 HL once again becomes the majority language, whereas the former L2 ML becomes a minority language. Given this trajectory, returnees provide a fruitful avenue to investigate two potential linguistic consequences: heritage language reversal (‘re-development’), and L2 attrition [2], [3]. In this paper, we examined whether Japanese-English returnee bilinguals exhibit signs of such HL reversal and L2 attrition in their speech, using data collected over the course of five years. We recorded 17 returnee children at three times: a few weeks after return to Japan (T1); one year after (T2); and five years after return (T3). Mean age at return was 10.02 (sd = 1.71). Mean age of onset (AoO) to L2 English was 5.15 (sd = 2.59) and mean exposure to L2 English (relative to L1 Japanese) whilst abroad, calculated by [4], was 0.48 (sd = 0.14). Recordings were elicited narratives of a picture book from which we created 10-second samples. These samples, in addition to 17 ‘baseline’ samples of Japanese and English monolingual children, were used in two online accent-rating tasks, in which native speakers of American English and Japanese (each n = 45, and familiar with child speech) rated the degree of GFA of the samples on a 9-point Likert scale. If raters indicated a ‘8’ or ‘9’ (‘very strong foreign accent’), they were additionally asked to indicate what features contributed to their perception of a GFA. Internal consistency for the ratings (Cronbach’s α) was high (0.916 for English and 0.908 for Japanese). A Bayesian model with weakly informative priors investigated the effect of language, time, and two experiential factors of interest (AoO to L2 English and exposure to L2 English) on accent rating. The model suggested that foreign accent in L1 Japanese decreased from T1 to T2, b = -0.31 (-0.48, -0.15) and continued to do so from T2 to T3, b = -0.27 (-0.43, -0.10), as shown in Figure 1. By contrast, foreign accent in L2 English increased from T2 to T3, b = 0.31 (0.13, 0.50). The model also suggested that individuals with a later AoO to L2 English returned to Japan with relatively weak foreign accents in Japanese, but strong foreign accents in English. By contrast, individuals with more exposure to English during their stay abroad returned to Japan with relatively strong foreign accents in Japanese, but relatively weak foreign accents in English. In addition, we found that English raters indicated that both segmental (vowels and consonants) and suprasegmental features (intonation, rhythm, and voice quality) contributed equally to the perception of a strong foreign accent, whereas Japanese raters primarily attributed suprasegmental features to their perception of foreign-accented Japanese (Figure 2). This could suggest that the phonological correlates of perceived global accent may vary cross-linguistically, cf. [5]. Our findings show a swift decrease in foreign accent in the L1 one year after return, and an increase in foreign accent in the L2 five years after return to the majority L1 environment. These may be indicative of HL reversal and L2 attrition in the domain of speech. We discuss how two speech systems develop over time in the bilingual individual in light of existing theories of cross-linguistic influence (CLI) of phonology [6].

ABSTRACT. Inflectional morphology has demonstrated to be particularly vulnerable area of grammar across different languages (e.g., Rothman, 2007). Turkish has an agglutinative morphology; words are composed of a sequence of suffixes, offering grammatical features yet to be investigated in heritage speakers’ online processing of morpho-syntax. We aim to contribute to the understanding of how inflectional morphology in adult heritage speakers of Turkish is affected during online sentence processing by examining the comprehension and production of grammatical evidentiality. Turkish grammaticalizes evidentiality through obligatory verbal inflections. Another important aspect of heritage bilinguals is that they are a heterogenous population. Most studies have considered the linguistic and non-linguistic predictors behind heritage speakers’ individual differences in isolation. We investigate the interplay of individual differences including cognitive variables: chunking ability, working memory and attention and input and usage-based factors: education, language use and exposure.

Our tasks tap into implicit and explicit linguistic knowledge of the feature. The comprehension task using the visual-world eye-tracking paradigm tests heritage speakers’ sensitivity to direct and indirect evidential morphemes by examining moment-to-moment eye-movements while they listened to sentences with grammatical evidentiality. The production task elicits the production of evidential morphemes using animated clips. We specifically address the following research questions: (1) Is there a difference in adult Turkish heritage speakers’ sensitivity to direct and indirect evidential morphology in their comprehension and production of past tense? (2) Is there a difference between on-line and off-line measures of comprehension with regard to heritage speakers’ sensitivity to direct and indirect evidential morphology? (3) Do Turkish heritage speakers’ individual differences (education, language exposure and use, word chunking ability, working memory and attention) affect the processing of evidential morphology?

Regarding the (RQ1), we predict that Turkish heritage speakers would be more sensitive to the direct evidential morpheme than indirect one. As the interface hypothesis suggests that grammatical features at the interface of two linguistic domains, in particular the syntax-pragmatic interface, are more difficult for heritage speakers than features that can be processed at one single level (Sorace & Serratrice, 2009). This is because bilinguals may have access to fewer processing resources and thus may experience difficulty in integrating multiple types of information. In Turkish, indirect evidential is used to express more complex semantics than the direct ones, as it is associated with epistemically modal connotations. Regarding the (RQ2), we expect that Turkish heritage speakers would show higher sensitivity to the target features in the off-line measures than on-line ones. This is because processing in a less dominant language is cognitively more demanding than processing in a dominant language for bilinguals. For example, L2 speakers showed higher sensitivity to complex grammar features in offline than online tasks (Chepyskaia et al. 2021). We also predict that input-related individual differences such as education, language exposure and use alongside cognitive variables attention and chunking ability would contribute to the processing of evidentiality.

ABSTRACT. Heritage language performance can vary based on various linguistic internal and external factors such as, but not limited to, the dominant/societal language, education in both languages, quality and quantity of language use within and outside of the family. In this study, we used sentence and non-word repetition tasks to shed further light into this observed HL variation. 347 adult heritage speakers of Turkish living in Germany took part in the online (web-based on Gorilla) study where they listened to words, non-words and sentences in Turkish and were asked to repeat which was recorded automatically. The responses were rated by two independent evaluators who were blind to the participants' language status following the guidelines prepared by Topbaş et al. (2013). Preliminary descriptive results show that, in the sentence repetition task, one-way ANOVA did not show any overall difference for sentence types depending on their linguistic complexity (All ps < .05). Looking at the individual scores, participants struggled most with two specific sentence types, namely relative clauses and wh-questions. In the non-word task, both the letter count and language-likeness mattered. The participants performed significantly better at language-like non-words than language unlike non-words (p < .001). They also performed better at language like non-words with a Turkish morpheme than language like non-words without a Turkish morpheme (p = .01). Additionally, their performance decreased while the letter count increased. Overall, the less language-likeness and the higher the letter count resulted in lower repetition scores. These results will be discussed in light of the recent discussions on individual variation in HL performance and use in the literature.

ABSTRACT. Common definitions of bilingualism are wide-ranging and inexact, often focusing on a variety of attributes that can be difficult to measure consistently, such as being totally fluent in two languages, growing up with both, using both on a regular basis, and/or having had formal education in both (Grosjean, 2010). Although research over the years has aimed to be more precise and objective in describing different individuals’ bilingualism, there continues to be a need to find more ecologically valid ways to describe bilingualism as the multidimensional life experience that it is (e.g., De Bruin, 2019; Gullifer et al., 2021; López et al., 2021). This need is particularly relevant when working with minoritized and racialized communities such as heritage bilinguals; a failure to explicitly embrace the diverse and dynamic nature of bilingualism can result in the perpetuation of prescriptive and hegemonic views within bilingualism research (Bayram et al., 2021; Flores & Rosa, 2015; Ortega, 2020; Pascual y Cabo & Rothman, 2012). As researchers, we should investigate the use of multi-dimensional measures as proxies to capture the diversity of outcomes along the continuum of (heritage) bilingualism. This study, an extension of Luque et al. (2022) with a different population of heritage Spanish speakers, aims to contribute to this line of work by examining a number of widely used proficiency measures—both objective and self-reported—to better understand how they characterize heritage Spanish speakers. In particular, we examine (1) each measure’s internal reliability, (2) the relationship between measures, and (3) the relationship between proficiency measures and experience factors.

Language background, experience, and proficiency data were collected from 43 heritage Spanish-English bilinguals in the US. Participants completed an extensive bilingual language experience questionnaire, adapted from both the Bilingual Language Profile (Birdsong et al., 2012) and the questionnaire for Spanish-English bilingual speakers from the National Heritage Language Resource Center (Montrul, 2012), which included subjective self-reports of proficiency in all four language skills (speaking, reading, listening, writing). Objective measures included a lexical decision task (Izura et al., 2014), designed to assess vocabulary knowledge and overall proficiency, and a 50-item, multiple-choice written test adapted from a portion of the Diploma de Español como Lengua Extranjera, or DELE (Montrul & Slabakova, 2003), designed to tap into grammar and vocabulary knowledge. Although final data coding and analysis are ongoing, a preliminary correlation analysis between proficiency measures revealed that the objective and subjective measures are positively related to one another, with the two objective measures having the highest (positive) correlation (r(84) = .58). Additional planned analyses include: (a) the examination of internal reliability of each measure, (b) a correlational analysis that should reveal the strength and direction of associations between these measures and different aspects of the participants’ bilingual experience and, (c) a cluster analysis with the aim of understanding how participants with different bilingual language experiences pattern across the different objective and subjective proficiency measures. The results of this study have the potential to shed light on more ecologically valid ways to assess variability in outcomes of heritage bilingual research.

References

Bayram, F., Kubota, M., Luque, A., Pascual y Cabo, D., & Rothman, J. (2021). You can’t fix what is not broken: Contextualizing the imbalance of perceptions about heritage language bilingualism. Frontiers in Education, 6, 628311. https://doi.org/10.3389/feduc.2021.628311

Birdsong, D., Gertken, L. M., & Amengual, M. (2012, January 20). Bilingual Language Profile: An easy-to- use instrument to assess bilingualism. COERLL, University of Texas at Austin. https://sites.la.utexas. edu/bilingual/

De Bruin, A. (2019). Not all bilinguals are the same: A call for more detailed assessments and descriptions of bilingual experiences. Behavioral Sciences, 9(3), 33. https://doi.org/10.3390/bs9030033

Flores, N., & Rosa, J. (2015). Undoing appropriateness: Raciolinguistic ideologies and language diversity in education. Harvard Educational Review, 85(2), 149-171. https://doi.org/10.17763/0017-8055.85.2.149

Grosjean, F. (2010). Bilingual: Life and reality. Harvard University Press.

Gullifer, J. W., Kousaie, S., Gilbert, A. C., Grant, A., Giroud, N., Coulter, K., Klein, D., Baum, S., Phillips, N., & Titone, D. (2021). Bilingual language experience as a multidimensional spectrum: Associations with objective and subjective language proficiency. Applied Psycholinguistics, 42(2), 245-278. https://doi.org/10.1017/S0142716420000521

Izura, C., Cuetos, F., & Brysbaert, M. (2014). Lextale-Esp: A test to rapidly and efficiently assess the Spanish vocabulary size. Psicológica, 35(1), 49-66.

Luque, A., Issa B., Faretta-Stutenberg, M., & Bowden, H. (2022, October 28). Capturing the Diversity in Spanish Heritage Bilingual Language Experiences and Learning Outcomes: Assessing the Assessments. [Conference session]. Bilingualism Forum, The University of Illinois at Chicago, Chicago, IL, United States.

López, B. G., Luque, A., & Piña-Watson, B. (2021). Context, intersectionality, and resilience: Moving toward a more holistic study of bilingualism in cognitive science. Cultural Diversity and Ethnic Minority Psychology. Advance online publication. https://doi.org/10.1037/cdp0000472

Montrul, S. (2012). Bilingual background questionnaire for Spanish/English speakers. National Heritage Language Resource Center. http://www.nhlrc.ucla.edu/data/questionnaires.asp

Montrul, S., & Slabakova, R. (2003). Competence similarities between native and near-native speakers: An investigation of the preterite/imperfect contrast in Spanish. Studies in Second Language Acquisition, 25(3), 351–398.

Ortega, L. (2020). The study of heritage language development from a bilingualism and social justice perspective. Language Learning, 70, 15-53. https://doi.org/10.1111/lang.12347

Pascual y Cabo, D., & Rothman, J. (2012). The (il)logical problem of heritage Speaker bilingualism and incomplete acquisition. Applied Linguistics, 33(4), 450-455. https://doi.org/10.1093/applin/ams037

ABSTRACT. Research suggests that bilingual experience is associated with grey matter changes, such that initial language gains are associated with expansion and language expertise is associated with renormalization. Previous studies on language proficiency development primarily focused on between-subjects, quasi-experimental comparisons of monolinguals and bilinguals. This study proposes a new paradigm to examine language expertise and cortical thickness within heritage bilinguals (n = 215), as well as between bilinguals and monolinguals (n = 145), using data combined from 8 previous MRI studies. In general, results highlight variability within bilinguals, finding relationships between cortical thickness and English proficiency that are relatively consistent within monolinguals, but inconsistent within bilinguals. In all participants, higher levels of proficiency in English—monolinguals’ only language and bilinguals’ second but stronger language—were negatively related to cortical thickness. In bilinguals, higher proficiency in the weaker, albeit first-learned, language was positively related to cortical thickness. Moreover, there was an interaction between language group and English proficiency in predicting cortical thickness, such that the relationship between proficiency and thickness was stronger in monolinguals than in bilinguals. Findings also demonstrate that the regions associated with language expertise differ between bilinguals and monolinguals. Future directions for cognitive-developmental neuroscience research in bilinguals are suggested, particularly the longitudinal examination of cortical changes in relation to bilingual experiences.

ABSTRACT. Monolingual speakers of languages with strict Subject-Verb-Object word order may use verb-semantics (e.g., eat) to predict an upcoming object (e.g., cake) (e.g., Altmann & Kamide, 1999), whereas monolingual speakers of languages with flexible word order, such as Turkish, may form predictions using case-marking cues (e.g., Özge et al., 2019). Little is known about the prediction skills of bilingual speakers of Turkish, and in particular of bilinguals who have acquired or are acquiring their two languages in early childhood (e.g., Karaca et al., 2021). To this end, the current study examined to what extent early bilingual children generate predictions based on case-marking cues in comparison to monolingual children and early bilingual adults in Turkish. In a visual world eye-tracking paradigm, 49 monolingual children (Mage=7.1, SD=0.5), 25 Turkish-Dutch bilingual adults (Mage=26.7, SD=5.2) and 13 Turkish-Dutch bilingual children (Mage=7.4, SD=0.8 - data collection ongoing) listened to Turkish sentences in which case-marking on the first noun (NP1) was manipulated to be either accusative as in (a) and (c) or to be nominative as in (b) and (d), and verb position was manipulated to be either sentence-final as in (a) and (b) or to be sentence-medial as in (c) and (d). The sentences were presented with a picture of the referent of the NP1 (e.g., rabbit), a plausible agent (e.g., fox) and a plausible patient (e.g., carrot) (c.f. Özge et al., 2019).

a)Hızlı tavşanı birazdan şuradaki tilki∅ yiyecek. Speedy rabbitACC soon there foxNOM eat “The fox over there will soon eat the speedy rabbit” b)Hızlı tavşan∅ birazdan şuradaki havucu yiyecek. Speedy rabbit∅ soon there carrotACC eat. “Speedy rabbit will soon eat the carrot over there” c)Hızlı tavşanı birazdan yiyecek şuradaki tilki∅. Speedy rabbitACC soon eat there foxNOM “The fox over there will soon eat the speedy rabbit” d)Hızlı tavşan∅ birazdan yiyecek şuradaki havucu. Speedy rabbit∅ soon eat there carrotACC “Speedy rabbit will soon eat the carrot over there”

The fixations to the agent versus patient image during the predictive time window between the NP1 and NP2 were analyzed using mixed effect logistic regressions. For the verb-medial sentences, the results showed successful prediction effects for all three groups. However, for the verb-final sentences, monolingual children showed a prediction effect and bilingual children did so tentatively, whereas bilingual adults did not at all (see Figure 1). These results suggest that under the influence of non-case-marking languages, early bilingual adults who acquired both languages in childhood may require additional support from other cues (e.g., verb semantics) to generate predictions based on case-marking cues, but early bilingual children may do so to a lesser extent.

ABSTRACT. Recent research calls for the need to investigate individual variables in addition to group comparisons when studying bilingual populations1,2. In heritage language studies, L1 proficiency has been shown to have significant effects3, as has dominance in the L24.

We report results from a larger study investigating the role of relative language use in Spanish/English bilinguals. We compare group analyses to analyses of use variables (historical use, i.e. use over time and current use) in the processing of two types of island violations in L1 Spanish: wh-questions out of temporal adverbials and relative clauses. We use pupillometry which measures changes in pupil diameter in response to linguistic stimuli and is a reliable measure of processing effort5.

Fifty-six Spanish/English fluent bilinguals living in the US (aged 18-45, M = 28.02) listened to aurally presented items in two conditions: grammatical and ungrammatical. Data on L1 Spanish/L2 English fluency, use of L1 over time, and current L1 use were collected in an extensive questionnaire. For the group analysis, participants were classified as Late Bilinguals (LB: n=24) or Heritage Speakers (HS: n=34) based on age of arrival.

Data were analyzed using generalized additive mixed-effects models (GAMMs). For each island type, two models were run: Model 1 compared grammatical (a) and ungrammatical (b) sentences across HS and LB. Model 2 combined participants and looked at the interaction between grammaticality, historical usage, and current usage. For temporal adverbials, LBs as a group showed a significant grammaticality effect with ungrammatical items eliciting larger pupil dilation than grammatical items. HS did not show this effect. Model 2 combines HS and LB and showed current usage to be significant. For both HS and LB, higher use of Spanish at time of testing resulted in a stronger grammaticality effect. Historical usage was significant but did not interact with grammaticality: more use of English over time is associated with larger overall pupil sizes. For Relative Clauses, Model 1 showed a significant grammaticality effect for both LB and HS with a larger effect for LB than HS. Combining participants, Model 2 showed significance for current use: with more Spanish use we see a strong grammaticality effect. Historical use was not significant.

Results are discussed in terms of 1) differences between Heritage Speakers and Late Bilinguals and 2) the effect of relative language use on both bilingual types. While group analyses show mixed results, use variables are consistent indices of increased effort in both HS and LB when processing L1 ungrammatical items. Our results support the claim that individual variables are important factors to consider when studying bilingual populations, and that in addition to proficiency and dominance, historical and current use of a language significantly affects processing of that language. This suggests that neither HS nor LB can be considered homogeneous as a group and points to the need for a more nuanced characterization of these bilinguals than the view that they are cognitively distinct by virtue of age of L2 acquisition.

Sample Stimuli: Temporal adverbial island Context sentence El niño comió el dulce mientras que su tía buscaba la comida. the child eat.PRET.3SG the candy while COMP his aunt search.IMP.3SG the food ‘The child ate the candy while his aunt looked for food.’

Grammatical, no extraction out of temporal adverbial a. ¿Qué niño comió el dulce || mientras que su tía buscaba la comida? what child eat.PRET.3SG the candy while COMP his aunt search.IMP.3SG the food ‘What child ate the candy while his aunt looked for food?’

Ungrammatical, extraction out of temporal adverbial b. *¿Qué tía || el niño comió el dulce mientras que buscaba la comida? what aunt the child eat.PRET.3SG the candy while COMP search.IMP.3SG the food ‘What aunt did the child eat the candy while looked for food?’

Relative clause island Context sentence Paola hizo el gesto que causó la controversia. Paola make.PRET.3SG the joke COMP cause.PRET.3SG the controversy ‘Paola made the joke COMP caused the controversy.’

Grammatical, no extraction out of temporal adverbial a. ¿Qué gesto hizo Paola || que causó la controversia? what joke make.PRET.3SG Paola COMP cause.PRET.3SG the controversy ‘What joke did Paola make that caused the controversy?’

Ungrammatical, extraction out of temporal adverbial b. *¿Qué controversia hizo Paola || el gesto que causó? what controversy make.PRET.3SG Paola the joke COMP cause.PRET.3SG ‘What controversy did Paola make the joke that caused?’

Selected References

1. Grey, S. (2022). Variability in native and non-native language. An ERP study of semantic and grammar processing. Studies in Second Language Acquisition. First View online, published March 2022.

2. Luk, G. and J. Rothman (2022). Experience-based individual differences modulate language, mind and brain outcomes in multilinguals. Brain and Language 228.

3. Bice, K. and J. Kroll. (2021). Grammatical processing in two languages: How individual differences in language experience and cognitive abilitites shape comprehension in heritage bilinguals. Journal of Neurolinguistics 58

4. Gathercole, V. and E. Thomas. (2009). Bilingual first-language development: Dominant language takeover, threatened minority language take-up. Bilingualism: Language and Cognition, 12(2), 213-237.

5. Schmidtke, J. (2018). Pupillometry in linguistic research: An introduction and review for second language researchers. Studies in Second Language Acquisition, 40(3), 529–549

ABSTRACT. Following predictions from the Matrix Language Framework (Myers-Scotton 1993), recent research indicates that intra-sentential code-switching must be regulated by the verbal inflection (Blokzijl et al 2017, Herring et al. 2010, Parafita Couto & Gullberg 2019, Parafita Couto & Stadthagen González, 2019, Ramirez Urbaneja 2020), such that the functional elements of the clause must be in the same language as the verbal inflection (INFL, encompassing several functional categories such as Tense, Aspect, Voice and Agreement.) In (1), switching between D and NP is licensed by INFL, i.e., the determiner must remain in the same language as INFL:

(1) INFL Constraint INFL1 … [DP D1 NP2]

Following (1), Parafita and Stadthagen-González (2019:356) show that (2) is acceptable by heritage Spanish-English bilinguals while (3) is not:

(2) English/Spanish Edgar wanted these zapatos. (3) * Edgar wanted estos shoes.

However, the examples tested in Stadthagen et al (2019) and almost all the examples reported from production data (Blokzjil et al. 2017, Parafita Couto & Gullberg 2019) involve a direct object adjacent to the verb. This may have affected the results, as it is a fact of English grammar that functional categories cliticize to the left, so the determiner of the direct object might have cliticized to the verb. Code-switching between a clitic and its host is a well-known restriction on code-switching (Poplack 1980, MacSwan 1999). This might have led to the difference in acceptability between (2) and (3). As for the Spanish determiner, it is well-known that syllabification in this language crosses word boundaries from left to right and, as Hoot (2012) has argued, prosodic structure is built from left to right. This suggests that the direct object determiner may more easily attach to the verb than to its object. Thus, it seems desirable to test the hypothesis with examples in which the determiner and the verbal inflection are not adjacent. Additionally, (3) involves two consecutive switches in which only a functional category (a determiner) is in the embedded language. Thus, the aim of this study is to confirm (or not) that the INFL constraint is real, beyond the experimental limitations in earlier work. In order to investigate these research questions we carried out one online audio survey (3 point Likert scale) with Dutch- Papiamentu bilinguals (n=43) in The Netherlands. We manipulated the position of the switch and controlled for directionality effects by presenting code-switches in both switching directions (see sample stimuli in next page).

Consistent with recent corpus and experimental literature (cf. Vaughan Evans et al. 2020 for Welsh-English, Balam et al. 20020 for Spanish-English), our preliminary results point to a clear asymmetry regarding directionality effects, reflecting how code-switching is deployed in the community (i.e. Papiamento to Dutch, cf. Parafita Couto & Gullberg, 2019). If we control for directionality, we find that we do find a scale of acceptability, so that the conditions that respect the INFL constraint are preferred. We will discuss how our findings underscore the importance of bilingual language experience in modulating linguistic competence.

Sample stimuli

Papiamentu / Dutch

a. Bo kind snijdt het brood. your child cut the bread b. Bo mucha snijdt het brood. your child cut the bread c. Jouw mucha ta korta e pan. your child PRES cut the bread d. Jouw kind ta korta e pan. your child PRES cut the bread e. Jouw mucha snijdt het brood. your child cut the bread f. Bo kind ta korta e pan. your child PRES cut the bread

References Balam, Osmer, M. Carmen Parafita Couto, & Hans Stadthagen-González. 2020. Bilingual verbs in three Spanish/English code-switching communities. International Journal of Bilingualism 24(5–6), 952–967. Blokzijl, Jeffrey, Margaret Deuchar, & M. Carmen Parafita Couto. 2017. Determiner asymmetry in mixed nominal constructions: The role of grammatical factors in data from Miami and Nicaragua. Languages 2(4), 20. Herring, Jon Russell, Margaret Deuchar, M. Carmen Parafita Couto & Mónica Moro Quintanilla. 2010. ‘I saw the madre’: Evaluating predictions about codeswitched determiner-noun sequences using Spanish-English and Welsh-English data. International Journal of Bilingual Education and Bilingualism 13(5), 553-573. Hoot, B. 2012. Narrow focus on pre-nominal modifiers in Spanish: An Optimality-Theoretic analysis. In K. Geeslin & M. Díaz-Campos (Eds.), Selected proceedings of the 14th Hispanic Linguistics Symposium (pp. 293-307). Cascadilla Proceedings Project. MacSwan, Jeff. 1999. A Minimalist Approach to Intrasentential Code-Switching. New York: Routledge Myers-Scotton, Carol. 1993. Common and uncommon ground: Social and structural factors in codeswitching. Language in society, 475-503. Parafita Couto, M. Carmen & Marianne Gullberg. 2019. Code-switching within the noun phrase: Evidence from three corpora. International Journal of Bilingualism 23(2),695-714. Parafita Couto, M. Carmen & Hans Stadthagen-Gonzalez. 2019. El book or the libro? Insights from acceptability judgements into determiner/noun code-switches. International Journal of Bilingualism 23(1), 349-360. Poplack, S. 1980. Sometimes I'll start a sentence in Spanish y termino en español: Toward a typology of code-switching. Linguistics 18, 581-618. Ramírez Urbaneja, Desirée. 2020. “Tú tienes una little pumpkin?” Mixed noun phrases in Spanish-English bilingual children and adults. International Journal of Bilingualism 24(4), 824-839. Vaughan Evans, Awel, M. Carmen Parafita Couto, Bastien Boutonnet, Noriko Hoshino, Peredur Webb-Davies, Margaret Deuchar, & Guillaume Thierry. 2020. Switchmate! An electrophysiological attempt to adjudicate between competing accounts of adjective-noun code-switching. Frontiers in Psychology 11: 549762.