- Which Sounds Change:Descent and Borrowing in the Skou Family1
The process of establishing genetic relationships in the endeavor of historical linguistics is complicated by the fact that some, if not most, languages show no inclination to "behave themselves" by reflecting only material inherited from their protolanguage. Borrowing at many levels is rife in many languages, particularly when geographic separation is slight, and when social contact is frequent. When both language-internal change and language-external change happen during the same time frame, sorting out relations can be complicated. In this article I show that not all sounds behave equally with respect to areal spread. This is presented through a study of sound changes in the Skou languages of New Guinea, and a survey of other reports of diffusing behavior from elsewhere. Recognizing and working with these differences can allow us to sort out relative chronologies and thus historical relations in even complex scenarios of borrowing and change.
This article presents historical argumentation for a small family of languages spoken in the center of the north coast of New Guinea. In addition to presenting some primary data for a family that has previously been under-described, even from a New Guinea perspective, it also attempts to highlight the methodological issues of using the comparative method in a New Guinea context, issues that were raised by Foley (1986) (see 1.1). Unlike Foley's case study of Lower Sepik languages, the languages of the Skou family are spoken in a small and geographically accessible region, and so show a lot more cross-influence than do Foley's group; this could be expected to represent a more complicated picture. This is in fact the case. Accordingly, this article shows in practice the tangle of dialect-like diffusion mixed with genetic inheritance that characterizes most contiguous language families in New Guinea, and establishes a set of principles for operating in this jungle.
1.1 Theoretical Goals.
Sorting a set of isoglosses into those representing diffusion and those representing inheritance is not a new issue in historical linguistics, [End Page 171] and has been the subject of much discussion in the literature since the Family tree (Stammbaum) and Wave model (Wellentheorie) models were proposed in the nineteenth century (Schleicher 1860, Schmidt 1872). While much work has shown that these two models are not in fact mutually exclusive (e.g., Labov 1981, Ross 1988, 1997a), there is still a gulf between them conceptually. It is clear that there are procedural gains to be had with either of these models, but the question of how to sort the data into different sets, one set for analysis by cladistic techniques, one for analysis by diffusion, is not clear.
In common with the more familiar European situation, New Guinea presents similar challenges for cladistic or diffusional analysis. With New Guinea historical linguistics, however, the issues occur on a greater scale. Both socially and geographically, New Guinea is a milieu that lends itself to more intense contact with more diverse neighbors than the large nation-states that characterize Europe, for almost all societies in the region are small scale. To quote Foley (1986:208, 209): "Borrowing has occurred on a massive scale in most European languages, but by a judicious blend of the use of the comparative method and the sifting of written documents, the sorting of cognates from borrowings has been largely accomplished. With Papuan languages, we face a much more daunting assignment. Papuan language families are small and are generally spoken in small areas. The languages are usually contiguous, and have been so for millennia. ... Papuan languages normally exhibit a pattern of enormous cross-influence in all areas; so in no sense can the assumption that the daughter languages develop independently be taken as viable in this context."
In this article, I show how we can, through the careful use of existing methodologies, sort out the inherited, taxon-defining changes from the diffusing—and thus network-defining—changes in the one family, namely, that which contains Skou and the languages of the West Vanimo coast. The smaller Skou family is a convenient point for starting to examine the competing roles of...