EFL learners' attitudes toward Turkish-English code-mixing

The basic notion that has prompted language attitude research in sociolinguistics is that speech is an important mediator in the way people perceive one another. Recent interest in code-alternation (code-switching and code-mixing) in linguistics, which is seen as a distinctive feature of bilingual speech, has led to studies that deal with attitudinal consequences of this behavior. The research on evaluative reactions to code-switching reveal that attitudes toward distinct languages do not always correspond to the attitudes toward the mixed variety of the same languages. Based on these studies, the assumptions that serve as a basis for this study are: (1) listeners' attitudes toward a given speaker are indicative of their attitude toward the language form, (2) code-mixed speech, a type of code-alternation, is a speech style with distinctive characteristics, and (3) people's attitudes toward the code-mixed variety of two languages may be different from their attitudes toward those languages in their distinct forms. The study investigated the attitudes of EFL learners at different proficiency levels--proficient and non-proficient--toward two types of Turkish-English code-mixing--professional and non-professional. The hypotheses were that there is a significant difference between listeners' attitudes to Turkish-English code-mixed speech in terms of their level of proficiency in English and that the context where code-mixing appears moderates the listeners' subjective evaluation of the speaker. The proficient users of English were expected to be more accepting of code-mixing than the users with limited proficiency. It was also expected that code-mixing in a professional context would be perceived more favorably than code-mixing in a non-professional context. In order to test the hypotheses, an attitude test was administered to a group of undergraduate students selected on the basis of their proficiency level in English. The measure assessed, in quantitative terms, the subjects' responses to two speakers, each representing code-mixed speech occurring in different contexts. Data obtained from the measure were analyzed using analysis of variance, which examined the effect of the two factors in question on attitudes. The results showed that there was not a significant difference between the responses with regard to level of proficiency while context was found to be a highly important variable that influenced the listeners' evaluative reactions toward code-mixing.