Symposium Language Contact and the Dynamics of Language: Theory and Implications. Max Planck Institute for Evolutionary Anthropology. Leipzig. May 10-13, 2007. Processes of creole formation and related contact-induced language change. Donald Winford. The Ohio State University. [email protected] It is to some extent arguable that creole formation has always been at the center of the study of contact languages. In the first place, creoles and pidgins were among the first foci of attention for 19th century historical linguists who challenged orthodox Stammbaum theories of language change. The most important of these scholars was undoubtedly Hugo Schuchardt 1882, 1883(, whose work on pidgins creoles and other outcomes of language contact explicitly challenged the conventional wisdom concerning the rarity of language mixture and affirmed the universality of contact-induced change. Schuchardt and other pioneers of creole studies such as Coelho (1880-6) and Hesseling (1897) explored many of the issues that are still topics of debate today, including the relationship between creoles and first and second language acquisition, and the role of substrate influence and universals in creole formation. Secondly, the processes of creole formation and the typological characteristics of creoles have long been a point of reference and comparison for other outcomes of language contact. During the last four decades of the 20th century, creole linguistics in many ways dominated the emerging field of contact linguistics, and it was commonplace for scholars to extend the term “creolization” to a variety of contact languages that arose under very different sociohistorical circumstances from those that produced creoles. The first major publication devoted to the ‘pidginization and creolization of languages” (Hymes 1971) contained several papers on contact situations that had not hitherto been treated as related to creole formation. For instance, Gumperz & Wilson (1971) described the well-known situation of language contact between Urdu, Kannada and Marathi in Kupwar, India, suggesting that the Kupwar varieties have processes of reduction and convergence 1 suggestive of pidginization and creolization” and that “the present state of the varieties is creole-like, in that one finds grammatical structure and lexical shape pointing to different sources, quite like the stereotype of a pidgin or creole as the words of one language used with the grammar of another” (1971:165). Similarly, Southworth (1971) suggested that, “whether or not Marathi qualifies as a true creole, its present characteristics are probably the result of a prolonged process of mutual adaptation between an Aryan language and a local pidgin-creole (or more likely, a series of pidgin-creoles)” (1971:268). Finally, Bailey & Maroldt (1977) argued that Middle English emerged as a result of processes of creolization involving influence from Old Norse and Norman French on Anglo-Saxon. They note that Middle English displays features typically associated with creoles, such as analyticity, simplification of bound morphology, and the use of Anglo- Saxon words in French functions. They describe creolization in a very broad sense, as follows: By creolisation, the authors wish to indicate gradient mixture of two or more languages; in a narrow sense, a creole is the result of mixing which is substantial enough to result in a new system, a system that is separate from its antecedent parent systems. (1977:21). In short, “creolization” became a cover term for different types of language mixture in a variety of contact situations. This generalization of the term seems at odds with its traditional use to refer to the elaboration of a pidgin, which Hymes (1971:84) referred to as “that complex process of sociolinguistic change comprising expansion in inner form, with convergence, in the context of extension in use,” The extension of the term to various contact situations was, in a sense, a reflection of its indeterminacy and lack of clear definition. There was no clear articulation of the precise nature of the linguistic processes that were shared across the claimed cases of “creolization,” or of the similarities and differences that led to such varied outcomes. Inconsistency in the use of the term among creolists themselves did not help. On the one hand, “creolization” has been used to refer to social processes such as the nativization or the vernacularization of a pidgin or expanded pidgin. On the other, it refers variously to a process of restructuring 2 of a target language by adults in second language acquisition, a process of “second language acquisition by several generations of …adults” (Arends 1995:235), and a process of first language acquisition by children targeting impoverished input. Moreover, there emerged a number of dubious assumptions about the nature of creoles and the process of creole formation. Among these are the view that there are specific linguistic features that are definitive of creoles as a type, and specific processes of change and restructuring peculiar to creole formation. Yet, until recently, neither the linguistic features nor the processes have been carefully explained or documented. The work of the last two decades has certainly filled the gap in our understanding. There is now growing recognition that the kinds of linguistic processes found in creole formation are common to other kinds of restructuring under contact. Such processes include various forms of ‘simplification’, varying degrees of structural convergence, the reanalysis of lexical items of one language in terms of the grammatical categories of another, and various other phenomena that are due to internally and externally motivated change. The more we understand about these phenomena, the closer we get to a unified framework in which to describe them. Henceforth, I focus my attention on those aspects of creole formation that have to do with substratum influence, which I will argue manifests itself in a wide variety of contact situations not traditionally treated as involving the creation of creoles. Creolists now generally degree that creole formation represents a somewhat unusual type of group second language acquisition, and therefore shares much in common with other cases of “natural” second language acquisition in terms of the processes of restructuring and the principles underlying them. Significant progress has been made of late in understanding these principles, particularly as they relate to the role of substrate influence. However, there is still no consensus on any one theoretical framework that might best explain the processes of creole formation. Various frameworks and models have been proposed for the description and analysis of substratum influence. Among them are Siegel’s (1999, 2000) Transfer Model, Lefebvre’s (1998) Relexification model, and Myers-Scotton’s (2002) Convergence model. The variety of terms and labels that have been used to describe the workings of substrate influence testifies to the apparent lack of consensus. I will argue that, in fact, the lack of agreement is only 3 apparent, and that these frameworks are all concerned with essentially the same process, and complement one another in various ways. The transfer model provides good insight into the constraints on the way L1 influence works under contact. The relexification hypothesis provides insight into the linguistic processes involved. The convergence model provides insight into the psycholinguistic processes that underlie the changes that occur. Hence, the three models can be integrated into a coherent picture of one of the primary mechanisms involved in creole formation – the transfer type that van Coetsem (1988) refers to as “imposition.” I will argue that imposition is the transfer type involved in the processes of change that have been referred to variously as relexification, transfer and convergence in Creole formation. Van Coetsem’s framework In order to address this issue, there is need for a consistent general framework within which processes of contact-induced change can be distinguished and categorized. The framework I employ here was introduced by van Coetsem (1988, 2000), who distinguishes between two types of cross-linguistic influence, or what he calls ‘transfer types”, namely, borrowing and imposition. The latter is largely equivalent to terms like ‘interference via shift’, ‘transfer’ ‘indirect diffusion’, and ‘substratum influence’ that appear in the literature. Borrowing and imposition, in this framework, are not seen as ‘mechanisms’ or ‘processes’, but rather as transfer types, or vehicles of contact-induced change. In both cases, there is a source language (SL) and a recipient language (RL). The direction of transfer of linguistic features is always from the SL to the RL, and the agent of transfer can be either the RL or the SL speaker. In the former case, we have borrowing (recipient language agentivity), in the latter, imposition (source language agentivity). Lexical borrowing is a typical example of the former kind of transfer, as illustrated in the many lexical items that have been adopted by English from various languages. Source language agentivity is exemplified in cases of second language acquisition, where learners transfer features of their first or primary language (the source language) into their version of the second (recipient) language. 4 Van Coetsem’s distinction refines the traditional distinction between borrowing and ‘interference’ by defining these types of cross-linguistic influence