Beyond the comparative method?

As we have seen in previous chapters, the criteria for establishing genetic relationships among languages were generally clear, and widely known and applied, with reliance on basic vocabulary, sound correspondences, and patterned grammatical evidence of particular sorts – where the comparative method played a central role. Nevertheless, a number of scholars have recently expressed dissatisfaction with what they perceive to be limitations of the traditional methods. “Since the tried-and-true Neogrammarian comparative method can only reach back a few thousand years before the evidence fades out, something else must be tried,” so declares Johanna Nichols (1996b:267), and recently she and others, recognizing the limitations of the comparative method, have proposed differing ways to see past them. While this goal is an appropriate one, none of the alternative approaches proposed to date has achieved success. In this chapter we assess several of these to show why they do not really reach beyond the limitations of the comparative method.

We began talking together and thinking about the subject matter of this book when we prepared a paper for the Spring Workshop in Reconstruction in 1991, held at the University of Pittsburgh. We later decided to write this book, but were not able to do that until now due to other obligations. With respect to the division of labor, William Poser is primarily responsible for the writing of Chapter 5, part of Chapter 3, and parts of Chapter 4 (especially sections 4.8 and 4011). Lyle Campbell is the principal author of the other chapters and sections of this book.

Introduction

What drives linguistic diversification? Why do languages split up and become families of related languages? What accounts for languages spreading to new territory? Because answers to how and why languages diversify into families of related languages are closely connected with issues of genetic affinity and linguistic classification, in this chapter we survey some of the main approaches which propose answers to these questions. We identify misleading claims and attempt to offer directions towards more adequate answers.

As mentioned in Chapter 1, throughout history there have been numerous theories about what causes languages to diversify, accounts involving, for example, migration, war and conquest, trade, geographical isolation, cessation of communication, social and economic organization (e.g. mounted warriors with expansionist proclivities, militaristic patriarchy), linguistically marked group identity entailing rights to resources, technological advantage (in, for example, food production, herding, navigation, metallurgy, military organization), even divine vengeance for the Tower of Babel caper.

Introduction

There has been considerable confusion recently about how certain languages came to be established as members of the Indo-European family. In this chapter we take up the cases that have been discussed most, setting the record straight and deriving methodological object lessons from these successes. The methods used to establish these as Indo-European languages are quite instructive. Specifically we examine how Hittite, Armenian, and Venetic were shown to belong to the Indo-European family. (For Indo-European more broadly, see Chapters 2, 3, and 4.)

Hittite

Hittite was the language of the ancient Hittite empire in central Anatolia (present-day Turkey), and came to be known from the many clay tablets, written in cuneiform (and the earliest of these in Hieroglyphic Hittite), some from Tel el-Amarna in Egypt, but mostly known from the libraries containing Hittite and other clay documents in Boğaz Köy (today called Boğaz Kale), Turkey. Ultimately these were deciphered. Hittite (and the Anatolian branch generally to which Hittite belongs) proved to be Indo-European, against earlier intuitions. The first substantive claims concerning the Indo-European affiliation of Hittite were made by Knudtzon (1902) (cf. also Bugge 1902 and Torp 1902) in a book devoted to two letters between a Hittite ruler and the king of Egypt found at Tel el-Amarna in Egypt.

In this appendix, we attempt to provide a representative, reasonably comprehensive (but by no means exhaustive) listing of the hypotheses of distant genetic relationships which have been proposed. Limitations of space and practical considerations prevent us from evaluating each of these. Obviously these hypotheses are not all of even quality. Some are plausible and deserve further attention; some would find it difficult to escape ridicule, though they might be entertaining. We make no effort to provide extensive references for these various proposals, but do mention some relevant bibliography in some of the cases.

1. Afroasiatic (formerly Hamito–Semitic) (Greenberg 1963)

2. Ainu–Altaic (Patrie 1982)

3. Ainu and Indo–European (Lindquist 1960; Narumi 2000a; Tailleur 1961; see Hamp 1968)

4. Ainu–Austroasiatic (Vovin 1993)

5. Ainu–Gilyak (cf. Naert 1962)

6. Algonkian–Gulf (Haas 1958b, 1960; Gursky 1966–7, 1968)

7. Algonquian–Gulf and Hokan–Subtiaba (Gursky 1965, 1966–7, 1968)

8. Almosan–Keresiouan: Greenberg combined his Keresiouan (composed of Caddoan [including Adai], Iroquoian, Keresan, and Siouan–Yuchi) and Almosan (Sapir's Algonquian–Wakashan, combining Algic and Mosan) (Greenberg 1987:162–4).

9. Altaic (Georg et al. 1999; Grunzel 1895; Menges 1961, 1975; Miller 1987, 1991; Menovshchikov 1968; Poppe 1960, 1965, 1973, 1974, 1975; Ramstedt 1946–7, 1952, 1957; Shherbak 1966, 1986a, 1986b; Starostin 1991b; Starostin et al. 2003; Schott 1853, 1860; see Doerfer 1966, 1968, 1973, 1985, 1988; R'ona–Tas 1974; Ramstedt 1914–15, 1915–16; Starostin 1986; Unger 1990a; cf. Joki 1975, 1976, 1977, 1980)

10. […]