Ten years of research back up the bold new theory advanced by authors Thomason and Kaufman, who rescue the study of contact-induced language change from the neglect it has suffered in recent decades. The authors establish an important new framework for the historical analysis of all degrees of contact-induced language change. "For the integration of contact-induced language change into historical linguistics this book constitutes the greatest breakthrough since Uriel Weinreich's "Languages in Contact of 1953, and I am convinced it will be the touchstone for the further development of the discipline for years to come."--Edgar W. Schneider, "English World-Wide Sarah Grey Thomason is Professor of Linguistics and Terrence Kaufman is Professor of Anthropology and Linguistics at the University of Pittsburgh. Language Contact, Creolization, and Genetic Linguistics By Sarah Grey Thomason and Terrence Kaufman University of California Press Copyright 1992 Sarah Grey Thomason and Terrence Kaufman All right reserved. ISBN: 0520078934 1 Introduction "Es gibt keine Mischsprache" Max Mller (1871.1:86) "Es gibt keine vllig ungemischte Sprache" Hugo Schuchardt (1884:5) For well over a hundred years, mainstream historical linguists have concentrated heavily on system-internal motivations and mechanisms in studying language change. The methodological principles embodied in the powerful Comparative Method include an assumption that virtually all language change arises through intrasystemic causes. Most historical linguists, therefore, would probably still agree with Welmers' view that, in phonology and morphosyntax, external influences "are insignificant when compared with internal change . . . the established principles of comparative and historical linguistics, and all we know about language history and language change, demand that . . . we seek explanations first on the basis of recognized processes of internal change" (1970:45; emphasis ours). Max Mller's claim that mixed languages do not exist reflects this prejudice, both because a mixed language could not arise without extensive foreign influence and because the existence of mixed languages would constitute a potential threat to the integrity of the family tree model of genetic relationship (and hence to the Comparative Method itself). It is surely no accident that it was the first great creolist, Hugo Schuchardt, who provided an extreme counterclaim to Mller's extreme claim about mixed languages. Pidgin and creoles were obviously the prime candidates for mixed-language status. But Schuchardt's interests were by no means confined to pidgins and creoles, and his research on contact-induced language changes of all sorts confirmed his belief in the universality of language mixture. Since 1884, most reactions to Schuchardt's challenge have fallen into three main categories. First, some historical linguists have simply ignored the position Schuchardt represents, along with the massive amount of data that he and many others have adduced in support of the claim that contact-induced language change at all levels of linguistic structure is a pervasive phenomenon. Oksaar, for instance, observes that there are "no clear cases that would permit the generalization of statements that grammatical paradigms, bound morphemes, word order etc. can be subject to interference" (1972:492). Other linguists have accepted the challenge but rejected Schuchardt's conclusion. Arguments along this line are of two types. One is that some linguistic subsystemin practice, either the basic vocabulary or the inflectional morphologyis relatively impervious to foreign influence and is therefore a safe diagnostic tool for classifying a language genetically. That is, that subsystem is taken to be a reliable indicator of the oldest, and thus inherited, linguistic material in the language. The other is the claim that no language is so mixed that it cannot be fit unambiguously into a family tree: it will always be possible to show that the bulk of a language's lexicon and grammatical structures come from the same source. The third major reaction to the proposal that all languages are mixed is an acceptance of the proposition, with all its apparent consequences for genetic linguistics. Arguments for this approach make explicit reference to pidgins and creoles, but they often go beyond the languages traditionally called pidgins and creoles to claim that all languages are in effect creoles, since probably all have undergone considerable foreign interference in the course of their development. In this view, "the traditional Stammbaum theory of linguistic relationships cannot be upheld" (Mhlhusler 1980:34), because (among other things) probably "every system or node on a family tree should have at least two parents" (Bailey 1973:20). We d