Does language contact inevitably lead to change? This question underlies longstanding and often contentious debates about the linguistic outcomes of contact, fuelled by a fundamental lack of consensus about methodology and definitions. A frequently used term like code-switching (CS) can refer to a number of different things: which language to use, using more than one language in a conversation, changing language within a single sentence, or using a single word from another language. CS is commonly believed to characterize situations of long-term bilingualism, and the view that CS leads to mutual restructuring of the languages in contact is usually offered to explain the existence of linguistic areas (or Sprachbünde), in which co-existing genetically unrelated languages maintain distinct lexicons but share grammatical structure (e.g. Gumperz & Wilson Reference Gumperz, Wilson and Hymes1971). This view is dubbed the ‘convergence via CS’ hypothesis by Torres Cacoullos and Travis (TCT), whose book constitutes an explicit test of this hypothesis via a detailed analysis of the variable expression of subject pronouns in Spanish/English bilingual speech. However, rather than constituting yet another salvo in ongoing debates, this book lays out a series of methodological criteria that must be satisfied in order to demonstrate the connection between contact and convergence.
The book's underlying assumptions and rationale are laid out in ch. 1, ‘Language contact through the lens of variation’ (1–12). Instead of inferring the existence and direction of contact-induced change from isolated examples of bilingual speech taken out of their discourse context, TCT make use of the ‘orderly heterogeneity’ of language to adduce patterns of linguistic behaviour as a more reliable indication of convergence. Bilingual behaviour must be compared to monolingual benchmarks through the identification of ‘conflict sites’, defined as ‘loci of structural divergence at which a determination of the direction of change can be made’ (5).
Inferring contact-induced change requires satisfying several criteria: detailed knowledge of the sociolinguistic situation in which contact takes place, a representative sample of the bilingual community, and appropriate bilingual speech data (Poplack Reference Poplack and Preston1993). These criteria are addressed for the present study in ch. 2, ‘The community basis of bilingual phenomena’ (13–34), and ch. 3, ‘Good data: Capturing language use’ (35–56). New Mexico constitutes an ideal site for testing convergence, as Spanish has been spoken there since the seventeenth century and has been in contact with English for almost 200 years. The US annexation of New Mexico in 1848 led to pressure for the Spanish-speaking population to shift to English, accompanied by stigmatization of Spanish, particularly the local variety. Data come from the New Mexico Spanish-English Bilingual corpus (NMSEB), a set of sociolinguistic interviews with forty speakers from across New Mexico ranging in age from eighteen to eighty nine, conducted by community members in 2010–2011 and transcribed prosodically. Ch. 4, ‘Characterizing the bilingual speaker’ (57–72), provides a ‘composite sociolinguistic profile’ of the NMSEB speakers.
The variable expression of subject pronouns has been linked to a number of functional and formal explanations, which are explored in ch. 5, ‘Subject pronoun expression: Reconsidering the constraints’ (74–110). Functional explanations involve considerations of contrast or emphasis, subject continuity (same vs. switch reference), the resolution of ambiguity, and the accessibility or discourse recoverability of referents. Formal explanations include structural priming (the use of a variant in one utterance primes or prompts its recurrence in subsequent utterances) and the linking of clauses through coordination or subordination (which is correlated with subject continuity). Subject expression may be triggered by the need to resolve ambiguity based on person-number syncretism in some Spanish tenses, as well as by the different discourse functions of tenses. Verb class effects may be traced to collocations of verbs and subjects in particular discourse contexts.
Arguments for convergence on the basis of subject expression often appeal to rates of use, but overall rates can fluctuate for reasons unrelated to contact. As argued in ch. 6, ‘Cross-language comparisons: Reconsidering language types’ (111–35), a better gauge of the grammatical system is provided by the structured variability internal to each language. Evidence for convergence lies in uncovering similarities across languages in the significance of factors conditioning the variation, the direction of effect within each factor and the relative strength of factors (Tagliamonte Reference Tagliamonte2006; Meyerhoff Reference Meyerhoff2009). Ch. 7, ‘Assessing change and continuity’ (136–59), compares present-day speakers with an earlier corpus of NM Spanish speakers. Different rates of subject expression across time are shown to result from different distributions of particular contexts in each corpus, while a comparison of language-internal constraints shows no diminution of effects across time. Ch. 8, ‘The most intimate contact: The bilinguals’ two languages’ (160–73), tests predictions for convergence by comparing the Spanish and English used by the same speaker. On every point of comparison, the Spanish of the bilingual community aligns with monolingual Spanish and their English aligns with monolingual English.
The effects of CS on subject expression are considered in ch. 9, ‘Code-switching without convergence’ (174–87), and ch. 10, ‘Code-switching and priming’ (188–202). Although ‘convergence via CS’ predicts more subject expression in the vicinity of switches to English, rates remain unchanged whether CS is defined as a single word or multiple words. While there is evidence of cross-linguistic priming in bilingual speech (that is, an expressed subject in one language primes the use of an expressed subject in the other language), this effect is weaker than that of intra-linguistic priming. Thus, CS may affect rates of subject expression by shifting the frequency of the conditioning context, but this is not equivalent to cross-linguistic convergence.
Ch. 11, ‘Bilingualism in its linguistic and social context’ (203–10), summarizes the study and contextualizes its results in the context of debates over the linguistic consequences of language contact. (Two appendices provide further information about the transcription conventions and details about the factors conditioning subject pronoun expression across varieties.) Testing the effects of language contact on convergence by comparing the language-internal conditioning of subject expression in bilingual speech with different benchmarks reveals no evidence to support ‘convergence via CS’.
These results challenge the view that convergence is the inevitable outcome of long-term bilingualism and CS, although we should be cautious in generalizing the results of this study to other situations of language contact. As TCT point out (9), contact linguistics has been handicapped by disparate standards of proof and meagre data. Using the structure of linguistic variation as a tool in measuring grammatical similarity invokes the need for quantitative reasoning, which requires adequate samples of representative data. We should also avoid the temptation of trying to fit all situations of language contact into a common mould. TCT caution against the quest for a typology of language contact that makes deterministic predictions about linguistic outcomes. In situations of long-term bilingualism, speakers are adept enough at both languages to keep them separate even as they switch between them. In other words, ‘the two grammars in contact are interconnected, but they are not mixed’ (210). In contrast, situations of rapid language shift are more likely to involve processes of second language acquisition and approximation. The impossibility of predicting linguistic outcomes in one contact situation based on the outcomes in another situation calls for similar accountable studies of bilingual behaviour in its social context.
