Age of onset of exposure in codeswitching

Codeswitching, or the bilingual practice of switching between two languages, is a frequently misunderstood phenomenon in many fields, including education. Given the growing number of bilingual students and English Language Learners in U.S. schools, it is imperative that the field of education be informed by current research in bilingualism and language acquisition, including codeswitching. Codeswitching that occurs within a sentence is subject to specific rules derived from the languages involved in the switching. Furthermore, a codeswitcher's intuitions about the grammatical acceptability of certain switches over others, called grammaticality judgments, provide linguists with a unique window into how the language systems interact. In current codeswitching research, it is sometimes claimed that simultaneous and early sequential bilinguals provide more accurate grammaticality judgments than late sequential bilinguals. Although this claim is largely motivated by Critical Period Hypothesis research, the grammaticality judgments of the three groups of bilinguals have yet to be systematically compared to determine if there is indeed a difference in judgments. This dissertation investigates potential differences in intrasentential codeswitching patterns of simultaneous, early sequential and late sequential Slovak-English bilinguals (N = 39) through a comparison of grammaticality judgments. Analysis of potential differences is grounded in generative approaches to first and second language acquisition. Grammaticality judgments from Slovak-English bilinguals were elicited through a survey of constructed items. Chi square results are analyzed to determine variation in judgments attributable to bilingual group based on age of onset of exposure to English. In addition, a sub-study of data from the Welsh-English Siarad Corpus (http://www.siarad.org.uk/siarad.php) is presented. Normed token means for English and mixed tokens for simultaneous, early sequential, and late sequential bilinguals are compared using ANOVA tests, and variability is discussed in light of relevant theoretical considerations. Results from this study indicate that there are few differences attributable to age of onset of exposure, thus helping to clarify current practices in codeswitching research methodology, particularly in terms of identifying characteristics of participants. The study also addresses issues surrounding the critical period hypothesis and the effect of age of onset of exposure in bilingualism, topics which are both directly relevant to the field of education.