Introduction

Scholars appear to agree that a successful demonstration of linguistic kinship depends on adequate methods. Unfortunately, there is disagreement and confusion concerning what these methods are, and hence discussions of methodology frequently assume a central role in considerations of possible remote relationships. Given this state of affairs, it is important to appraise the various methodological principles, criteria, and rules of thumb, as well as pitfalls, relevant to investigating distant genetic relationships. That is the goal of this chapter. We provide guidelines for both framing and testing proposals of distant linguistic kinship, and we point out the frequent errors that need to be avoided. (In Chapter 9 we evaluate several of the more prominent hypothesized distant genetic relationships on the basis of the methods surveyed here.)

In practice, the successful methods for establishing distant linguistic affinity have not been different from those used to validate any family relationship, whether close or distant. The comparative method has always been the primary tool for establishing these relationships. Because the methods for investigating potential distant genetic relationships are not essentially different from those utilized to work out the history and classification of more closely related languages, this has resulted in a continuum from established and non-controversial families (e.g. Austronesian, Bantu, Indo-European, Finno-Ugric, Mayan), to more distant but solidly demonstrated relationships (e.g. Uralic, Siouan-Catawban, Benue-Congo), to plausible but inconclusive proposals (e.g. Indo-Uralic, Proto-Australian, Macro-Mayan, Niger-Congo), to doubtful but not implausible ones (e.g. Altaic, Austro-Tai, Eskimo–Uralic, Nilo-Saharan), and on to virtually impossible proposals (e.g. Basque–Na-Dene, Indo–Pacific, Mayan–Turkic, Miwok–Uralic, Niger–Saharan, and so on).