To Appear, in Code-switching and bilinguals’ grammars. In Evangelina Adamou y Yaron Matras (eds), The Routledge Handbook of Language Contact. London/New York: Routledge. Code-switching and bilinguals’ grammars Rena Torres Cacoullos and Catherine E. Travis ABSTRACT Does code-switching entail grammatical convergence or are speakers who regularly code-switch alternating between separate grammars? Underlying debates on code- switching are the methodological issues of what counts as code-switching, as well as appropriate data and evaluation metrics that prioritize community norms over idiosyncratic instances and robust patterns over isolated cases. This chapter illustrates how bilingual behavior as observed in sociolinguistically constructed corpora of spontaneous speech provides replicable findings. Widely entertained mechanisms of contact-induced change are tested through measures of code-switching presence in comparisons with non-contact benchmarks, pivoted on quantitative diagnostics of grammatical similarity evinced in the linguistic conditioning of variation. 1. INTRODUCTION Code-switching (CS), using two languages in a conversation, continues to amass scholarly attention. Though CS was once viewed, if noticed at all, as a haphazard mess not amenable to analysis, it has been a growing subject over the last 50 years. Attention to CS comes from different linguistic branches and approaches, ranging from psycholinguistic experiments seeking to illuminate language processing, to grammaticality judgments serving to adjudicate on syntactic theories, to recordings of conversations revealing usage in the speech community. CS has been implicated in change in bilinguals’ grammars, a conjecture that can only be tested via proper delimitation of CS and diagnostics of change. To begin to understand the impact of CS, we need to know what it is. We find a plethora of terminologies and taxonomies and an astounding lack of agreement even on how to recognize a code-switch. A major dispute bearing on the identification of CS is whether to characterize it as alternation between two languages, each of which retains grammatical independence, or rather as insertion of one language into the other, which would be the matrix language grammatically. In the latter situation, the relation between the two languages is asymmetrical, one being morpho-syntactically dominant, whereas in the former, both languages are at work and the question becomes that of discovering the structural sites of CS. The Equivalence Constraint states that bilinguals tend to avoid CS at points of word order conflict between the two languages; the notion of equivalence, congruence, or matching between languages features in many approaches to CS (Poplack, 1980, Poplack, 2015) (see also Lipski, 1978, Muysken, 2000, Deuchar, 2005, Muysken, 2015). The use of two languages within a speaker turn is exemplified in (1), from a Spanish- English bilingual corpus. Each line of transcription represents an Intonation Unit (IU)