In this book, Colin Renfrew presents an original hypothesis regarding the history of the Indo-European languages. A number of interesting ideas are to be found in here, but the central argument is unfortunately deeply flawed. It should be noted that Renfrew himself has modified his views in several important respects since the publication of this book in 1987, but this review addresses the book as published.

Although scholars have long sought to place the origins of the Indo-European language family more concretely in prehistory, the theoretical basis for such an endeavor has rarely been considered since Kossina and Childe in the early decades of the twentieth century. Bringing a modern awareness of the complexities of social process in prehistory, Renfrew discusses the circumstances in which language replacement occurs, and considers how this could occur on a spatial scale ranging from the British Isles to the Indian subcontinent. In brief, Renfrew argues that "if we see a very wide uniformity in language, we should seek a demographic and economic explanation." He then argues that the spread of agriculture at the beginning of the Neolithic in Europe, starting around 7000 BC, is the only time before the rise of empires in historical times that a plausible explanation for widespread linguistic uniformity can be found, and thus that the spread of the Indo-European languages should be associated with the expansion of agricultural populations originally from Anatolia.

One could question whether a unitary cause is needed: John Robb has shown that small-scale random language shifts, repeated over a long period, could spread any language family at the expense of others. And there is a social mechanism for language spread that Renfrew does not specifically consider, when social groups are broken up and the families or individuals incorporated into social groups of another speech community. This process is attested for non-stratified societies such as the Nuer/Dinka and the Iroquois/Huron, so although it does not provide a large-scale unitary explanation, there is no reason it should not be relevant to prehistoric Europe as well.

The attention to the mechanics of language spread is useful, and it certainly shows some of the flaws in previous discussions of Indo-European origins. Renfrew's hypothesis, though, is not well supported by the evidence. First, as Renfrew is well aware, there is still considerable debate on the importance of demographic migration in the spread of agriculture. Perhaps as importantly, the spread of agriculture in Mediterranean Europe was associated with a cultural complex that had little in common with the early agricultural societies of temperate Europe: this is not what would be expected for a single, unitary population spread.

The more important problems, though, are linguistic. Renfrew's archaeological credentials for the discussion of prehistoric Europe are beyond question, but the same cannot be said with regard to historical linguistics. The uncertainties of proto-language vocabulary reconstruction are usefully highlighted, but the discussion of loanwords is an embarrassment. Even a superficial acquaintance with historical linguistics should make it clear that the Latin loanwords found in many Germanic languages are clearly distinguishable from Germanic cognates of Indo-European derivation, because they have not undergone the sound changes described in Verner's law. In fact, the presence of certain sound changes and not others makes it possible to give a relative dating for loanwords such as caseus or pondo fairly specifically to a period when the West Germanic languages had not differentiated significantly from each other, but after they had diverged from the other Germanic languages.

In order to make his argument that the Indo-European languages began to diverge at the beginning of the Neolithic, Renfrew must explain away a variety of lexical evidence suggesting a much later proto-language. The original semantic field of some terms might be incorrectly reconstructed, and some derived words might have been formed independently in separate branches, but some lexical items remain very difficult to explain away. For example, words for "horse" and "copper/bronze" quite clearly reconstruct in Proto-Indo-European, yet should not have been part of the lexicon of Neolithic Anatolia.

Renfrew's repeated criticism of the family-tree model in historical linguistics is not entirely unfounded, but if he had paid as much attention to process in historical linguistics as in archaeology, he might have seen the relevance of it in certain respects. The Indo-European languages notoriously do not form a very tidy family tree: apart from the Anatolian languages, which clearly separated before the others, it is difficult to reconstruct a clear branching sequence. By comparison, the Austronesian languages, which Renfrew presents as an example of a language family similarly spread, do show a very clear branching pattern, in which the earliest divergences occurred nearest the original homeland, with the divergences becoming progressively later as the Austronesian languages spread. That such a pattern cannot be shown for Indo-European argues against any model that would have them spread by a demographic expansion extended over thousands of years.

Despite the problems with the central argument, the book is thought-provoking in a number of ways. Along the way, there are a number of useful, thought-provoking comments. For example, one of the more interesting is the hint regarding the relationship of the Hattic (hattili) language to the Hittite (nesili) language and society: it has long been assumed that Hittite was recently (as of the Middle Bronze Age) intrusive in central Anatolia, and that Hattic was the language of the conquered native population. Renfrew points out that there is no real evidence for which preceded the other, and that much non-Indo-European (presumably Hattic) vocabulary borrowed into Hittite is more what one would expect from a superstrate language (e.g., names of gods and persons) than a substrate language.

It is the willingness to raise such questions where others have simply followed conventional assumptions unthinkingly that gives this book its value. Although the main argument clearly fails on linguistic grounds, the book is nevertheless worth reading, because it forces one to think through theory and evidence more critically. ( )

A controversial new theory on the spread of Indo-European languages. Suggests our language evolved among the first settled farmers of Anatolia around 7000 BC and spread gradually and peacefully from there through agriculture. (Not the traditional theories of mass migrations, conquests on a huge scale.) Interesting arguments. ( )