Congruence and the Switch of Content and Function Words: A Typological Analysis on Southern Min/Mandarin Data

Aims: This study aims to examine the relationship between congruence and code-switching (CS) based on CS data from a typologically identical language pair (i.e., Mandarin and Southern Min) and three research questions: (1) What grammatical categories are more likely to be switched when the two participating languages are typologically identical?’ (2) ‘Does full congruence facilitate CS?’ (3) ‘Is the switch of content and function words determined by CS patterns?’. Methodology: Approximately 50 hr of spontaneous speech produced by three groups of subjects were recorded and analysed. A bilingual clause was the basic unit of analysis, and a linear criterion was used to identify switches. Syntagmatic and paradigmatic congruence between these switches and their equivalents in the other participating languages were examined. Finally, Muysken’s typological model was adopted to examine CS patterns and the switches of function/content words. Data and Analysis: Our data displayed a congruent lexicalisation CS pattern in our Southern Min/Mandarin data; the values of Muysken’s two diagnostic features that are in favour of such a CS pattern, namely [-content word] and [+function word], were examined. Findings: Musyken’s (2000, 2013) typological model provides both linguistic and extra-linguistic explanations of our results, although some modifications may be needed. Moreover, our results suggest that even if the CS pattern of our Southern Min-Mandarin data is congruent lexicalization, CS is not entirely unrestricted but appears to be category-conditioned. Originality: Previous studies on congruence only focus on typologically variant language pairs. The current study examines a typologically identical language pair. Significance: Previous studies were empirically inadequate because their findings were mainly based on analyses of data from typologically variant language pairs. Our analysis reveals that Muysken’s typological model could account for CS data from both typologically similar and variant language pairs and provide a more comprehensive examination of the relationship between congruence and CS phenomena.