The title of this book and the meeting that gave rise to it imply a fourfold dialogue: between Uralic and Indo-European specialists, and between linguists and archaeologists. Archaeology and linguistics have a complicated relationship. Both disciplines emerged in response to the search for roots and origins that characterised the political and intellectual concerns of late 18th and early 19th-century Europe, and represented alternative methodologies for tracing the origins of "peoples" whose identities were perceived as underlying the emergent pattern of European nationstates.

One outcome was the development of classification systems for ancient artefacts and assemblages, described in terms of "types" and "cultures"; the other was a more elegant procedure for describing regular shifts of pattern in the ways that words were formed and used. The former resulted in what was often a forbidding laundry list of groups named after type sites or distinctive artefacts, which sometimes occupied only tiny areas of the earth's surface but sometimes expanded alarmingly halfway across whole continents (often to disappear in an equally mysterious way); the latter typically produced a set of family trees, both summarising the relationships between recorded languages and suggesting common ancestral entities that might underlie them. Each reflected the nature of the raw material from which they were constructed, but both were intended to illuminate a common reality - the original historical events and processes that gave rise to this diversity of pots and proto-languages.

From time to time (and especially at times of political change) it has seemed important to bring these schemata together and to explore their congruences; at other times (as during the cold war) the two subjects have been happy to exist in relative isolation from each other and to explore alternative agendas. Each period of isolation has allowed the development of fresh insights, potentially useful to such a confrontation: archaeologists have recently (and in some cases perhaps prematurely) come to describe these early groups in terms of their being "farmers" or "pastoralists", rather than just the makers of characteristic pots; while linguists have had an expanding range of analogies to draw on, from comparable studies of language families in Oceania or Africa. Each of these new areas, in turn, has thrown up its own puzzles of the relation between linguistic and archaeological entities, setting the original Indo-European problem in a broader global ±è°ù´Ç²ú±ô¨¦³¾²¹³Ù¾±±ç³Ü±ð .

There remain, however, certain basic difficulties that are inherent in the nature of the different sources of evidence. Early material-culture entities can be given precise dates by tree-ring and radiocarbon chronology, while divergence dates for speech communities are notoriously elastic; but dead pots cannot speak, and even very similar-looking pots may not have spoken the same language. There are many reasons why cultural patterns may have spread, not all of which imply uniformities in speech. Nevertheless, it would be odd if the historical processes underlying these phenomena did not, to some extent, coincide, and the attempt to look for large-scale events that influenced both sorts of patterning is worthwhile. While the attempt to integrate two different sets of evidence in a detailed way in deep prehistory is frustrating and error-prone, the attempt to gain a common historical perspective on the larger questions is particularly timely when a third class of evidence may help to resolve some ambiguities, even though it raises more questions than it provides answers.

ÁñÁ«ÊÓƵ

This third class of evidence is ancient DNA (as opposed to present-day patterns of genetic variability, which suffer from the same drawbacks as comparative linguistics, namely their lack of absolute time-depth), which may indeed resolve the old question of whether people moved rather than just their pots and styles of pot-making, with important implications for models of language change. Such conclusions are for the future, however, and are mercifully absent from this discussion, since it is a fundamental methodological rule of such conversations that each class of evidence must be first evaluated in its own terms and not subjected to a series of ad hoc accommodations to prevailing models in other disciplines. That way lies circularity.

"Archaeology and language" conferences take a number of characteristic forms. One is the "world" model, which tends to feature new research frontiers in Africa and the Pacific (often in conjunction with genetic data), and offers unconventional interpretations of the spread of Indo-European - either at much earlier dates than most Indo-Europeanists accept or from unexpected places. Another is the "Indo-Europeanist" model, which is a more closely knit community of philologists, mythologists and the kinds of archaeologist interested in the domestication of the horse. A third type is closely focused on precise questions such as Greek or Gaulish dialects and the evidence of inscriptions.

ÁñÁ«ÊÓƵ

The present volume falls into none of these categories. Although dealing with a relatively large section of the earth's landmass, it is historically focused rather than comparative; but it deals with two major language families and their interaction, and concerns a potentially long prehistoric period rather than a well-documented historical episode. These features make it refreshing, and the perceptive choice of contributors, together with its readable English, makes it a satisfying volume.

The result is neither seamless (ten purely linguistic papers, eight integrated ones) nor unanimous (do similarities in vocabulary imply contact, or a common "Nostratic" ancestor?); but the differences of opinion usually reflect alternative answers to clearly defined problems, not just incoherent squabbling. There is some elegant infighting among the linguists (J. Koivulehto and E. Helimski: all the more enjoyable for being phrased respectively in contrasting patrician and ironic styles), and some critical cold water from an experienced archaeologist of central Asia (H. P. Francfort); but also the outline of a broadly constructed consensus. Part of this is a new willingness on the part of archaeologists to countenance "migration" as an important category of cultural change. The disrepute into which this concept had fallen was partly a result of its over-easy adoption to explain any kind of stylistic discontinuity, often including those arising from the imperfections of the archaeological record itself. As gaps were filled, it became less easy to invoke simple models of population replacement; and the idea itself lost its attraction by comparison with the kinds of evolutionary description that came to dominate both in the US and the USSR. Contemporary political circumstances, however, together with new kinds of molecular evidence, have combined to revive the idea, which is now being applied in a more critical manner to a fuller and better-dated set of archaeological observations. Although it is still dangerous to align prehistoric entities with any of the genetic groupings identified within living populations, there is the promise of gaining a fuller three-dimensional picture (biological, material-cultural and linguistic), if only over-simple equations can be avoided. The new interest in ethnicity, identity and cultural transmission has brought language back into the focus of archaeological attention and made this a worthwhile encounter.

Some cultural phenomena (and their suggestive linguistic congruences) deserve to be more widely known. One is the advanced school of bronze-casting that characterised the forest belt north of the Eurasian steppe zone in the early 2nd millennium BC. Known as the Seima-Turbino complex, it reached from the Urals to the Altai Mountains and was responsible for transmitting casting techniques as far as China. This is a plausible background for a linguistic community of which Finno-Ugrian forms the western part. Indo-Europeans expanding along the steppe corridor maintained cultural and linguistic exchanges with their northern neighbours. The two histories were interwoven, and several possible weaving patterns are discussed in these papers. (An additional linguistic conclusion: the English language is alive and well and living in Helsinki!) This well-crafted volume is worth seeking out from the ghettos to which librarians may confine it. It represents elegant and interesting scholarship.

Andrew Sherratt is professor of archaeology, University of Oxford.

ÁñÁ«ÊÓƵ