Morpho-syntactic structure of code-switched sentences produced by Albanian bilingual speakers

This paper aims to analyze the structure of code-switched sentences uttered by Albanian bilingual students. It examines the predictions made by two theoretical viewpoints, namely Myers-Scotton’s Matrix Language Frame (MLF) and MacSwan’s approach within the Minimalist Program (MP). The students were observed in informal environments within Prishtina international school settings, such as during recess time, on the playground, etc. Tape recordings of the conversations were used to record this naturalistic data, which was then transcribed. In particular, examples from the Albanian/English pair were analyzed, and the MLF and MP predictions regarding combinatorial possibilities were tested. The findings show that our data can only partially be accounted for by the MLF and MP approaches. Switching is mainly unidirectional, with insertions from English into an Albanian syntactic frame. Determiners, such as demonstrative pronouns, indefinite pronouns, adverbs of quantity, as well as copular verbs, auxiliary verbs, and clitics, come from the Matrix Language (ML), complying with the MLF model. However, violations of the MLF model are encountered as well, such as the non-occurrence of stem/affix switching, the occurrence of bare forms, and Embedded Language (EL) islands not always constituting the maximal projection of the phrase, which are further described and predicted in MP model. The paper provides evidence that neither MLF nor MP are able to adequately account for the possible combinations in the mixed clause. According to what the literature to date indicates, integrating the approaches rather than using each one individually results in a better understanding of the grammar of code-switching.