Code-switching, the alternating use or mixing of two languages in bilingual communication, is often assumed to be a sign of defective language use, and is therefore likely stigmatized because people incorrectly conclude that speakers who code-switch are limited to producing an amalgam of broken bits of each language. In reality, a general characteristic of mixed utterances is that they are produced fluidly, without hesitation, pauses, or corrections, reflecting a systematic and exquisitely controlled integration of two linguistic systems. This project will collect eye-tracking data to examine code-switching during bilingual language reading and spoken language comprehension. The approach is cross-disciplinary, building on and bridging advances in both linguistic approaches to code-switching and experimental psycholinguistic work on language processing. The proposed experiments are motivated by two central questions: 1) Are there syntactic contexts that facilitate the processing of code-switched language?; and 2) How does the proficiency of speakers in both of their languages modulate language processing during comprehension?
The answer to these questions will contribute to the advancement of knowledge in several ways. With demographic changes in the US and more communication between individuals here and in other countries, the number of people communicating in writing is increasing at an extraordinary rate. Code-switching is ubiquitous in bilingual communities and it is becoming increasingly more present in its written form, especially in the case of Spanish-English speakers. Given the rapid growth of Spanish-English bilingualism in the US, it is increasingly important to improve understanding of how bilinguals process mixed language discourse. Second, the project evaluates proficiency in a second language more rigorously than past studies. Therefore, the results will provide valuable information for assessing second language proficiency, which is crucial in contexts in which proficient bilingual communication is required for successful cross-cultural interactions. In addition, this project will support bilingual graduate student researchers and bilingual undergraduate research assistants from underrepresented populations in the US, including Chicanos and other US Latinos, thus promoting the inclusion and training of individuals from traditionally underrepresented groups in scientific research both at the graduate and undergraduate levels.
Imagine you are walking down the streets of New York or at a mall in Miami. You will hear glimpses of conversations as people walk past you. Some conversations will be entirely in English, others will be entirely in another language (maybe a language you recognize or have learned), but you will also hear people alternating between English and another language. This linguistic behavior is called code-switching. Proficient bilinguals often code-switch in the midst of speaking with other bilinguals, and the linguistic principles that govern the observed code-switches have been the focus of much research. Code-switching performance has been examined primarily from the perspective of bilingual speakers, but there are critical consequences for bilingual listeners because unlike production, which is under the control of the speaker, bilingual listeners often cannot predict whether speakers will code-switch or will remain in one language only. It is easy to imagine why listening to bilinguals code-switch in the middle of a sentence should be taxing for the linguistic system. We know that one strategy listeners use to facilitate comprehension is to predict what the speaker will say next. Almost any English speaker is able to predict the word that fills the blank in the sentence "I drink my coffee with milk and _________." For English-Spanish bilinguals, hearing "azúcar" instead of the predicted "sugar" should cause a momentary feeling of surprise because the word uttered does not match their prediction. What is remarkable is that bilinguals don’t report having any difficulties comprehending code-switched speech of this sort. The central question addressed in the proposed studies was to examine the mechanisms that allow bilingual speakers to comprehend code-switched speech. Although recent studies have examined this question, most have been narrowly focused on switching at the level of single-word processing. A deeper understanding of the comprehension of code-switched language requires investigation of structures involving more than single lexical switches and/or manipulations of the syntactic locus of the switch point. This was the focus of our proposal. Our overarching goal was to systematically examine the variables that modulate processing costs during the comprehension of code-switched sentences. We tested the central hypothesis that bilinguals who are exposed to and engage in codeswitching between their two languages make use of the distributional patterns that arise in production to guide comprehension. We also examined the effect of task while bilinguals comprehend code-switches sentences. The results revealed two significant findings: (1) More frequently produced types of code-switches easier to process by the comprehension system than less frequent types of code-switches. This finding was replicated in all the syntactic sites examined (e.g., auxiliary phrase; complementizer phrase) and provides strong support for experience-based or usage-based models of processing (e.g., MacDonald & Seidenberg, 2006). (2) Different tasks result in distinct patterns of sensitivity during the comprehension of written code-switches. Frequent code-switches did not cause disruption during reading but only when participants were asked to process sentences for comprehension. When participants were asked to judge the acceptability of a sentence containing a code-switch, frequent and infrequent code-switches caused processing difficulty. Theoretically, the results of the studies conducted during this funding period provide critical support of experience-based accounts of language comprehension (e.g., MacDonald & Seidenberg, 2006). According to these models, frequent constructions are more readily activated by appropriate information sources than less common constructions. For example, studies show that comprehension difficulty is influenced by the match between syntactic structure and the frequency with which verbs appear in that structure. Thus, transitive verbs (e.g., believe) most often used by speakers with sentential complements cause less comprehension difficulty when followed by a sentential complement than by a noun phrase (NP) complement (Garnsey, Pearlmutter, Myers, & Lotocky, 1997; Spivey-Knowlton and Sedivy, 1995; Tanenhaus & Trueswell, 1995; Trueswell & Kim, 1998). Based on this evidence, we expected ease of comprehension of code-switched language to reflect production patterns. Of course, while construction frequency may not always predict comprehension preferences (cf. Gibson, Schütze, & Salomon, 1996; Kennison, 2001; Pickering, Traxler, & Crocker, 2000), growing evidence from experience-based studies of sentence processing indicates that frequency of exposure to certain constructions modulates comprehension difficulty. In support of this claim, our findings show that correspondences between comprehension and production extend to code switching. During the funding period, the PI and co-PI were actively engaged in training graduate students and undergraduates from underrepresented minorities. The PI and Co-PI were also involved in activities associated with the Center for Language Science (www.cls.psu.edu), of which Dussias is the Associate Director. Dussias is also involved in the NSF PIRE grant (www.cls.psu.edu/pire) for which she is a Co-PI, along with professors Judith F. Kroll (PI), Janet van Hell and Ping Li (Co-PIs). These contexts provided a rich set of opportunities for training students and for professional development.
