Bilingual discourse markers: Evidence from Croatian-English code-switching

This paper examines the presence of English-origin discourse markers in otherwise Croatian speech. Previous examinations of bilingual discourse marking have focused on habitualisation, pragmatic transference, cognitive processes (in terms of reducing the 'mental load' of distinguishing between two systems) or psycholinguistic factors (e.g. 'non-deactivation' of pragmatic forms from one language which speakers habitually speak). However, such studies only address transference of pragmatic norms, whether features and/or forms. They have limited explanatory power where speakers employ discourse forms from one language and equivalent forms from the other language. This study examines the frequency and functionality of English-origin forms compared to Croatian ones in a sample based on speech recordings of 100 Croatian-English bilinguals. Findings here show that English forms generally co-occur with Croatian forms in a statistical sense and only displace Croatian ones where English forms are polyfunctional or perform more functions that their Croatian counterparts. This accounts for the higher statistical frequency of forms such as yeah and so, while forms such as no and you know co-occur to the same degree as their Croatian equivalents. Where a Croatian item such as kao ('like') can fulfil more functions than its English equivalent it is shown to have a higher incidence than like. Thus, linguistic, in particular micro-discourse, features are shown to influence selection of discourse markers. (c) 2006 Elsevier B.V. All rights reserved.