Units in code switching: evidence for multimorphemic elements in the lexicon

Studies of insertional code switching frequently include examples in which combinations of several words from the embedded language occur together in matrix-language clauses. In this paper, a range of such constructions are analyzed as multimorphemic lexical units. Such units are a central, yet neglected, element of grammar, which blur the distinction between syntax and lexicon. The paper identifies several types of unit in bilingual code-switching data, including inserted plural nouns, compound nouns, adjective-noun and verb-object collocations, and idioms, and provides evidence for their widespread involvement in insertional code switching. Independent evidence for unit status comes from their frequently observed semantic noncompositionality and other semantic and morphosyntactic features. Recognition of the lexical status of these chunks may contribute to a unified theory of code switching.