What Language Did Neolithic Pots Speak? Colin Renfrew's European Farming-Language-Dispersal Model Challenged
Kristiansen, Kristian, Antiquity
Introduction--the historical context
Colin Renffew's book Archaeology and Language--the puzzle of Indo-European origins (Renfrew 1987, modified in 1999 and 2001) was a bold attempt to reopen the dialogue between two historical disciplines. The debate had more or less shut down since the Second World War, although the Journal of Indo-European Studies and the work of Maria Gimbutas had continued a pre-war archaeological approach linking migrations and the spread of language to the so-called Kurgan waves (Gimbutas 1970). Paving the way for a new perspective, Renfrew studied and discussed a number of central issues such as models for language change and social change and the notion of ethnicity exemplified by the Celts. He then applied a processual theory to the problem, and created a model that linked the spread of the Proto-Indo-European language to the spread of agriculture from Anatolia into Europe, taking it several thousand years back in time.
Some formidable methodological and empirical obstacles stood against such a hypothesis, but Renfrew had the courage to overcome (or ignore) them. Since only attack can lead to victory (at least in military tactics), he opened several flanks: one against the method of glottochronology and linguistic reconstruction (chapter 5), one against the current model for Proto-Indo-European society, especially as represented by Dumezil, (chapter 10), and one against the notion of migrations in later prehistory (chapter 2). In this way Renfrew endeavoured to free himself from what he considered to be restrictive paradigms. This new freedom was then employed to develop a model of great time depth for the origin and spread of Indo-European languages.
What went rather unnoticed, however, was the fact that the new model also meant a partial retreat from those very same theoretical and methodological demands Renfrew himself had advanced. It also introduced an unhappy marriage between language, archaeology and genetic studies whose ingredients can very easily be misused in the hands of nationalist extremists. An understandable enthusiasm for expanding the limits of what can be known (Renffew 2000) seems to have blinded Colin Renfrew to his own otherwise strict demands for theoretical and empirical coherence and consistency. The model has been maintained in spite of serious factual critique of important aspects of it (lucidly summarised in Yoffee 1990, also in Stefanovich 1989). As this direction of research has now been propagated in several recent conferences and monographs (Renfrew & Boyle 2000; Renffew et al. 2000), I find it timely to make a critical re-assessment of its basic theoretical and methodological principles. In doing so I shall expose two key opponents in the discussion: Renfrew (the earlier) versus Renfrew (the later).
Since we should confine ourselves to testable models, I start by proposing that while we cannot track the course of languages archaeologically, we can map the institutions with which they are associated. On this basis I want to show that Renffew's critique of the Bronze Age as providing a context for the adoption of Proto-Indo-European institutions does not now stand up. I argue that these institutions can be recognised in terms of identifiable components, and that they are adopted successively across Eurasia during the second millennium BC. By contrast, there is no new evidence that would equate the Indo-European language with the spread of farming in the early Neolithic.
Social transformations during the third and early second millennium BC
Renfrew argued that prehistoric people would only have migrated during the early Neolithic, and then again in the Iron Age, when textual evidence testified to it. This rather surprising scenario suited Renfrew's interpretation and that still remains the best explanation for it. Only by denying mobility to societies between the fourth-second millennium could he make his model fit, as he did not consider the theoretical possibility of a change of language without migration. …