Combining the Markedness Model and the Matrix Language Frame Model in Analysing Bilingual Speech

Abstract

This paper responds to calls to end the current trend whereby sociolinguistic and grammatical analyses of code-switching (CS) are divorced. It demonstrates ways in which the two models developed by Carol Myers-Scotton -i.e. the Markedness Model (1993a) and the Matrix Language Frame model (1993b) -may be combined to elicit the sociolinguistic and grammatical factors that jointly explain the CS of members of a given speech community. Data analyzed come from Ewe-English CS, spoken in Ghana.Two interrelated points emerged from our combination of the two models in data analysis. The first is that although codeswitchers may be unconscious of what they are doing, they do have the capacity to monitor the sociolinguistic decisions they make and how those decisions shape the grammatical character of the bilingual expressions they produce. We note that clearer evidence that they exercise this potential can be found in their use of marked CS (CS employed to convey some social or discourse message). The second point is that whether CS is marked or is unmarked (not employed to fulfil much social and discourse intentions beyond indexing mutual solidarity) depends basically on sociolinguistic considerations and not so much on whether the CS is insertional or alternational; it is indeed demonstrated that both insertional and alternational types of CS may be treated alike as either marked or unmarked during even the same conversation. The two points throw light on the future of Ewe and languages like it in third world settings that are in intensive contact with English or French.