2017

PIRE fellow Anne Beatty-Martinez presents at the Bilingualism in the Hispanic and Lusophone world conference (BHL).

Exposure-based models of language processing posit that sensitivity to distributional patterns in production can constrain the comprehension system. Codeswitching provides a unique circumstance to elucidate this relationship because codeswitching emerges in some bilingual communities but not in others. This study examines the extent to which individuals’ production choices can predict comprehension difficulty in codeswitching and non-codeswitching bilingual populations. We recruited 22 Spanish-English codeswitchers in the US and 22 non-codeswitchers in Spain who were highly proficient in both languages. We focus on the production and comprehension of mixed noun phrases (mixed NPs; e.g., elM forkM). Evidence from naturalistic corpora suggests that bilinguals exhibit an overall preference for the masculine determiner (el), regardless of the noun’s gender in Spanish. In contrast, switches involving the feminine determiner (la) occur less frequently and are restricted to English nouns that are feminine in Spanish. To illustrate, el-codeswitches such as (elM forkM) and (elM spoonF) are extremely common. To a lesser extent, la-codeswitches involving feminine nouns (laF spoonF) have also been attested in bilingual speech, while those involving masculine nouns (*laF forkM) have not. We predicted that electrophysiological responses would reflect this asymmetry in codeswitchers but not in non-codeswitchers. That is, while codeswitches in general should be unexpected for non-codeswitchers, only rarely-attested switches should induce processing difficulties in codeswitchers. Production was examined in a corpus of unscripted, task-oriented dialogues between two bilingual speakers; comprehension was examined in sentential contexts using event-related potentials where we compared different types of switches (switch vs switch) and effects of switching (switch vs non-switch). The same participants completed the production and comprehension experiments. Results for production show that proportions of noun phrase types (i.e., Spanish, English, or mixed) differed across groups: Codeswitchers produced more mixed NPs than non-codeswitchers and these switches robustly reflected the aforementioned gender asymmetry. In the ERP study, the two groups of bilinguals differed in their processing of codeswitched sentences. While non-codeswitchers were insensitive to the congruency and gender of switched target nouns, codeswitchers demonstrated an asymmetry in how they process different types of switches. Specifically, codeswitchers exhibited an N400 effect to masculine targets in incongruent noun phrases, suggesting greater difficulty in lexical integration. Furthermore, we found that only non-codeswitchers displayed an early positivity (P2) for switch vs no-switch comparisons. Importantly, codeswitchers did not exhibit switch costs in conditions of the sort found in their own utterances during the production study. This is evidenced in the lack of the P2 component in masculine congruent and feminine congruent and incongruent switching conditions. To safeguard against potential language effects in the switch vs no-switch comparisons, we conducted a control experiment in which we compared electrophysiological responses to unilingual translation-equivalent sentences in English and Spanish using the same participants. Because no differences were found due to lexical characteristics of the target words in the control study, we suggest that the P2 switching effect in non-codeswitchers reflects detection of a language change during early monitoring stages of language processing. Overall, the current study further provides evidence that individuals are highly sensitive to the constraints of their language experience, and sheds light on how production and comprehension processes are tightly linked. Our findings demonstrate how switching costs largely depend not only on the type of codeswitch but also the bilingual’s language experience.

PIRE undergraduate fellow Delaney Wilson presented at the Bilingualism in the Hispanic and Lusophone World Conference.

Code-switching, alternating between multiple languages in an utterance as in “the boy ate la pera en la cocina”.1, is a hallmark of bilingualism and shows that bilinguals keep both languages active. While there is a substantial body of research on the interaction between a bilingual's syntactic systems during code-switching, only a few studies have specifically examined the interaction between a bilingual's phonetic systems during the production of code-switched speech. Voice onset time (VOT) is the length of time between the release of a stop consonant and the onset of voicing. Recent studies have shown that Spanish-English bilinguals show significantly longer VOTs while naming unrelated pictures during switched trials2 than during non-switch trials and that their language-switched speech was “more accented”3, indicating that switching induces a difference in phonetic output. Since both of these studies focused on language switching during single picture naming, we employed a sentence creation task to see if VOT lengths are also present in more naturalistic code-switched sentences. Languages differ in the types of VOT they use to implement contrasts, i.e., English is a long-lag language, meaning that the VOTs of voiceless stops in English are longer than the shorter stops found in Spanish. The interaction of the two phonologies potentially leads to a transfer of VOT patterns in code-switched speech. The present study is an acoustic analysis of voiceless stop VOTs (focusing on word-initial /p/ and /t/) in switched and non-switched sentences. The outcomes provide insight into how a meaningful sentence context may affect the phonetic realization of code-switched words embedded within this sentence. A previously used experimental technique4 was adapted to elicit code-switched sentences. Pictures contained two images (the subject and object in the target sentence), each circled in a separate color for a code-switched sentence. The colors indicated in which language each should be referenced. There were four language conditions: a switched sentence from Spanish to English, a switched sentence from English to Spanish, a single-language English sentence, and a single-language Spanish sentence. Target words could appear before or after the language switch. Single-language sentences were presented in separate blocks and were not color-coded. Twenty-four Spanish-English bilinguals were tested at the University of Granada, Spain. The data of 11 participants were analyzed, based on their language proficiency scores for Spanish and English (DELE and MELICET). Forty-eight tokens were selected (2 languages*2 consonants*12 words). A total of 6,336 tokens (48 target words*2 conditions (unilingual, code-switch)*2 positions (before, after switch)*3 repetitions*11 participants) were collected and 5,696 were analyzed. Preliminary results show that English VOTs are longer than Spanish VOTs, which was expected. Pre-switch and initial target positions tend to be longer than their post-switch counterparts, but not significantly so. No significant difference manifested between code-switching and unilingual trials. While this suggests that there is no tangible anticipatory or carryover effect on VOT during code-switched sentences in naturalistic speech, it does show that bilinguals are capable of producing a distinction between the differences in VOT length within languages.

PIRE fellow Manuel Pulido-Azpíroz presents at the 11th International Symposium on Bilingualism (ISB11).

Studies on the processing of second language (L2) collocations (e.g., “set the table”) have reported faster RTs in collocational over unrelated word pairs, but also more efficient processing of L1-L2 congruent collocations (word-by-word equivalents) relative to incongruent ones (e.g., Wolter & Gyllstad, 2011). While previous studies have assumed an advantage of congruent collocations in processing, we explore the alternative hypothesis that differences are due to the effect of L1 interference. Using ERPs, we investigated the effect of implicit L1 activation during a lexical decision task. Spanish-English bilinguals were presented with 656 Verb+Noun sequences in their L2 (80 congruent, 80 incongruent collocations, 80 unrelated, 88 fillers; 328 non-words). To investigate the effect of L1 interference, increased L1 activation was induced through the presence of a cognate noun in 50% of collocations. Results for the bilingual group (N=18) showed congruency-based differences. Repeated measures ANOVAs revealed an effect of congruency in the 340-450 (p<0.05) and 500-600 ms windows (p<0.01). Bonferroni post hoc tests revealed a significantly different negative-going modulation in incongruent collocations peaking at around 550 ms in averaged left frontal electrodes relative to the congruent (p=0.015) and the unrelated conditions (p=0.016). However, congruent collocations that contained cognates also presented a negative peak and were non-significantly different from cognates in incongruent collocations. These results support previous studies that found that when cognates are presented in a sentence facilitation decreases or disappears (e.g., Schwartz & Kroll, 2006). The data suggest that implicit L1 activation during processing of incongruent collocations results in recruitment of inhibition mechanisms within the 500-600ms window.

PIRE fellow Anne Beatty-Martinez presents at the 11th International Symposium on Bilingualism (ISB11).

Exposure-based models of processing posit that sensitivity to production patterns can constrain the comprehension system. Codeswitching serves to elucidate this relationship because codeswitching emerges in some bilingual communities but not in others. We examine whether bilinguals’ production choices predict comprehension performance. We recruited 22 Spanish-English codeswitchers in the US and 22 non-codeswitchers in Spain highly proficient in both languages. We examined the production and comprehension of mixed noun phrases (mixed NPs; elM forkM). Naturalistic corpora illustrate a preference for masculine determiners, regardless of the noun’s gender in Spanish. Conversely, switches involving feminine determiners are restricted to English nouns that are feminine in Spanish. We predicted that non-codeswitchers should be insensitive to this asymmetry whereas codeswitchers should show differential electrophysiological responses to rare (*laF forkM) versus frequent (elF forkM) codeswitches. Production was examined using a map task; comprehension was investigated in sentential contexts using event-related potentials. The same participants completed production and comprehension experiments. Results for production show that codeswitchers produced more mixed NPs than non-codeswitchers (B=2.28, SE=.36, z=6.38, p=.001), and these switches robustly reflected the aforementioned distributional asymmetry. In comprehension, the two groups differed in their processing of mixed NPs. While non-codeswitchers were insensitive to the distributional patterns of different codeswitches, codeswitchers exhibited greater processing difficulty (N400) to rare vis-à-vis frequent codeswitches (F(1,21)=19.77, p=.000, η2p=.49). These findings support our prediction that bilinguals’ comprehension of codeswitches is influenced by the distributional regularities they experience. Furthermore, we found that only non-codeswitchers displayed an early positivity (P2; F(1,21)=10.12, p=.004, η2p=.36) for switch vs no-switch comparisons (elM forkM; elM tenedorM). Importantly, codeswitchers did not exhibit such effects. Overall, our findings shed light on the production-comprehension link, and demonstrate how the correlates of codeswitching largely depend on bilinguals’ experience with distributional codeswitching regularities.

PIRE fellow Anne Beatty-Martinez presents at the 58th Annual Meeting of the Psychonomics Society.

Although bilinguals are typically slower to speak the L2 relative to the L1, changes to the native language have been observed in response to active L2 use. These observations have been attributed to either reduced-lexical access (frequency-lag hypothesis) or to the presence of activation of the nontarget language (competition-based models). Here we characterize these differences as a function of immersion context and language use (e.g., codeswitching). We compared picture naming performance with high- and low-frequency items across three Spanish-English bilingual groups and in their two languages. Overall, bilinguals were slower than monolingual controls. Non-codeswitching bilinguals immersed in their L1 were faster in the L1 than in the L2 and this difference was greater for low-frequency items. Codeswitching bilinguals, regardless of immersion context, were faster in the L2 than in the L1, although this effect disappeared for low-frequency items. We discuss these results with respect to accounts of bilingual language production.

PIRE undergraduate fellow Erika Exton presented at the Annual Meeting of the Psychonomic Society.

Numerous studies on semantic memory observed that monolinguals are faster and more accurate in processing concrete than abstract words. However, many aspects about bilingual semantic processing and lexical-semantic memory are not yet fully understood, including concreteness effects in second language (L2) processing. This study, as part of a larger project on the neurocognitive correlates of semantic processing in younger and older monolinguals, bilinguals, and individuals with aphasia, examined concreteness effects in lexical decision and semantic relationship judgment tasks in the second language in Dutch-English bilinguals. Concreteness effects in L2 lexical decision were small and depended on second language proficiency. Robust concreteness effects were found in the L2 semantic relationship judgment task; moreover, for abstract, but not for concrete words, performance was better for associative pairings than for similarity pairings. Implications for lexical-semantic activation during L2 concrete and abstract word processing and bilingual semantic memory will be discussed.

PIRE undergraduate fellow Delaney Wilson presented at the Annual Meeting of the Psychonomic Society.

Bilinguals have longer voice onset times (VOTs) when naming isolated pictures during switched trials than non-switched trials (Goldrick et al., 2014; Olson, 2013), indicating that switching impacts phonetic output. To understand how this manifests in naturalistic speech, we employed a sentence creation task focusing on word-initial /p/ and /t/ phonemes. Habitual and non-habitual code-switchers produced sentences that switched from Spanish to English, from English to Spanish, and unilingual Spanish and English sentences. Spanish VOTs were shorter than English VOTs in all contexts, for both habitual and non-habitual code-switchers. In habitual code-switchers, English VOTs were shorter in code-switched than in unilingual sentences, suggesting phonological convergence in code-switched sentences. No such difference was found in non-habitual switchers, indicating they maintain phonetic distinctions between both languages in code-switched sentences. The difference between habitual and non-habitual code-switchers suggests that phonological convergence manifests differently with language use.

Experience-based models of language processing posit that sensitivity to distributional patterns in production can constrain the comprehension system (Dell & Chang, 2014; MacDonald, 2013). Codeswitching provides a unique circumstance to elucidate this relationship because codeswitching emerges in some bilingual communities but not in others. Beatty-Martínez & Dussias (2017) examined the extent to which individuals’ production choices predicted comprehension difficulty in codeswitching and non-codeswitching bilingual populations. We recruited 22 Spanish-English codeswitchers in the US and 22 non-codeswitchers in Spain who were highly proficient in both languages. We focused on the comprehension and production of mixed noun phrases (mixed NPs; e.g., elM forkM). Evidence from naturalistic corpora suggests that bilinguals exhibit an overall preference for the masculine determiner (el), regardless of the noun’s gender in Spanish. In contrast, switches involving the feminine determiner (la) occur less frequently and are restricted to English nouns that are feminine in Spanish (Valdés-Kroff, 2016). To illustrate, el-codeswitches such as (elM forkM) and (elM spoonF) are extremely common. To a lesser extent, la-codeswitches involving feminine nouns (laF spoonF) have also been attested in bilingual speech, while those involving masculine nouns (*laF forkM) have not. We predicted that electrophysiological responses would reflect this asymmetry in codeswitchers but not in non-codeswitchers. That is, while codeswitches in general should be unexpected for non-codeswitchers, only rarely-attested switches should induce processing difficulties in codeswitchers. Comprehension was examined in sentential contexts using event-related potentials where we compared different types of switches (switch vs. switch) and effects of switching (switch vs. no-switch); production was examined in a corpus of unscripted, task-oriented dialogues between two bilingual speakers. The same participants completed the comprehension and production experiments. In Experiment 1, the two groups of bilinguals differed in their processing of codeswitched sentences. While non-codeswitchers were insensitive to the congruency and gender of switched target nouns, codeswitchers demonstrated an asymmetry in how they process different types of switches. Specifically, codeswitchers exhibited an N400 effect to masculine targets in incongruent noun phrases, suggesting greater difficulty in lexical integration. Furthermore, we found that only non-codeswitchers displayed a switching effect in the form of an early frontal positivity for switch vs no-switch comparisons. Importantly, codeswitchers did not exhibit switch costs in conditions of the sort found in naturalistic codeswitching. This is evidenced in the lack of a switching effect in masculine congruent and feminine congruent and incongruent switching conditions. To safeguard against potential language effects in the switch vs no-switch comparisons, we conducted a control experiment (Experiment 2) in which we compared electrophysiological responses to unilingual translation-equivalent sentences in English and Spanish using the same participants. Because no differences were found due to lexical characteristics of the target words in Experiment 2, we suggest that the switching effect in non-codeswitchers reflects detection of a language change during early monitoring stages of language processing (Kuipers & Thierry, 2010). In Experiment 3, participants completed a task that elicited naturally-produced codeswitched speech. Results show that proportions of noun phrase types (i.e., Spanish, English, or mixed) differed across groups: Codeswitchers produced more mixed NPs than non-codeswitchers and these switches robustly reflected the aforementioned gender asymmetry found in naturalistic codeswitching. Overall, the current study further provides evidence that individuals are highly sensitive to the constraints of their language experience, and sheds light on how production and comprehension processes are tightly linked. Our findings demonstrate how switching costs largely depend not only on the type of codeswitch but also the bilingual’s language experience.