Trees of life and the babel of tongues

Colin Renfrew celebrates the fact that we are now learning about the origins of language from DNA analysis

It seems odd that we are learning more about the past of the human species today from the analysis of samples taken from living humans than we are from the results of archaeological excavations. Yet such is the power of molecular genetics and so rapid the pace of research that these tiny samples of blood or hair, when subjected to DNA analysis, permit the establishment of relationships between individuals and groups and from these the reconstruction of the evolutionary story of human origins, of the first dispersals of our species out of Africa, and of the foundations of human diversity. Our past is within us, and the genes in every one of us have a story to tell that only now is being deciphered.

In this well-written and readable book, Luigi Luca Cavalli-Sforza, who has spent an energetic and fruitful career as a geneticist, tells us how this is happening. The book is partly autobiographical, since Cavalli-Sforza has made many contributions himself to the developing discipline of what I like to call archaeogenetics - the application of the techniques of molecular genetics to the understanding of human history. It is certainly no dry textbook of molecular biology. Instead it is full of opinions and bright ideas. Indeed some of its conclusions remain open to question, but we can expect to see doubtful issues resolved within a few years as the vast flow of new data now becoming available is brought to bear.

Cavalli-Sforza was born in Genoa in 1922 and educated in Italy before coming to this country to work as a graduate student in Cambridge with one of the century's foremost statisticians, R. A. Fisher. Starting as a geneticist researching bacteria, he shifted his focus to human-population genetics and has worked on African pygmies. In 1971 he moved from Italy to Stanford University, where he served until his retirement as professor of genetics. There he came into contact with one of this century's leading linguists, Joseph Greenberg, an association that has clearly stimulated his interest in the possible relationships between the evolution of genetic diversity in humans and the evolution of linguistic diversity.

Cavalli-Sforza has a remarkable ability for working creatively in the little-charted boundary areas between disciplines. With a series of brilliant insights he has on several occasions developed a truly original idea, often in collaboration with a specialist in one of those intersecting disciplines, sometimes reaching a logical conclusion that carries the disciplines forward to a point not yet supported by available data. This has meant that his work has had its critics, and indeed there have been occasions when the critics may have been right, in that the initial conclusions have had to be modified. But that is precisely what makes his work so exciting, and ultimately so scientific, in the true sense of that word. For science advances by hypotheses that only years later can be adequately evaluated, when abundant further data are available.

Long before DNA sequencing became routinely available for those interested in population history, Cavalli-Sforza developed with a Cambridge colleague, Anthony Edwards, the Darwinian notion of evolutionary trees in a way that could make use of the "classical" genetic data then available, from blood groups, enzymes and so on. They used such data to establish a measure of similarity or distance between the human populations sampled. By computing the "genetic distance" between populations all over the world they were able to develop techniques permitting a classification of these populations in tree form. Under certain assumptions, they showed that the tree that they obtained could be regarded as a genuine evolutionary tree. The data then available, in 1961, did not allow of very secure results. But when laboratory techniques developed to allow DNA sequencing, and these were applied by Rebecca Cann and Allan Wilson in 1987 to mitochondrial DNA, which is passed on in the female line from mother to daughter, the result was the strong suggestion that our species Homo sapiens originated in Africa.

With a Stanford colleague, Albert Ammerman, Cavalli-Sforza suggested in 1973 that the spread of farming from Anatolia to Europe at the beginning of the neolithic period could be modelled mathematically by a process of "demic diffusion", using a formula developed by his old teacher Fisher. This idea has proved highly influential, although controversial, among archaeologists.

Around this time he returned to his interest in gene geography - using maps of frequency distributions of various genes (still based on the "classical", genetic markers obtained from blood group data and such) - and with two Italian colleagues applied quantitative techniques to seek simpler patterning in complex data. The method of "principal components analysis" was applied, yielding a strong spatial patterning on the maps obtained for each of the first four or five principal components.

They went on to apply comparable techniques to the human gene-frequency maps for all the continents of the earth, offering suggestions for the most significant demographic episodes of each. The results were set out in a magisterial volume, The History and Geography of Human Genes by Cavalli-Sforza, P. Menozzi and A. Piazza (1994).

This was the first major volume in the new field of archaeogenetics, and the authors proceeded to cap it with another ingenious initiative. In 1988 they drew on the rather controversial work by Greenberg in the field of linguistic classification. Greenberg had classified the language families of Africa into just four super-families, or "macro-families", and then had proceeded to propose that those of the Americas could similarly be lumped into just three macro-families, while most of those of Europe and western Asia could be classified together in his "Eurasiatic" macro-family.

Greenberg and his colleague Merritt Ruhlen had suggested tentatively that all the world's languages could be placed on a single, embracing, evolutionary tree. From the single ancestral mother tongue these different macro-families would have emerged, and they in turn would be ancestral to the language families and the languages recognised by the consensus of linguists. In a bold move, Cavalli-Sforza and his colleagues set side by side the Greenberg/Ruhlen family tree for languages and their own genetic tree based on their gene studies for the populations of the world. The comparison proved highly controversial. For in the first place the majority of the world's linguists regarded the Greenberg/Ruhlen language tree as at best unproven and at worst misguided. And second it seems likely that the pace of linguistic evolution is much faster than the pace of genetic evolution, so that the proposed equivalence might be illusory.

But it is not necessary in science to be right all of the time; it is sufficient to be original and to be interesting. By entering a strong and controversial hypothesis into play at so early a stage, Cavalli-Sforza and his colleagues have initiated more than simply a debate: it is a whole new field of study. My suspicion is that the hypothesis may not prove to be "right". But it is certainly proving to be fruitful.

This book guides the reader through these various ideas in a lucid and understandable way, and it is not over-burdened with technicalities. In the first chapter, Cavalli-Sforza deals with great clarity and good sense with the concept of "race", and asks the very good question "Why does this compulsion to classify human races exist?". He gives short shrift to racist thought but reasserts the utility of the study of human genetic diversity for various purposes including medical research. He has clear ideas on the processes of cultural transmission, and his analysis brings out clearly how they differ from genetic transmission.

Much of the subject material of this book was covered in 1994 in The Great Human Diasporas , written in collaboration with his son Francesco. For me, however, one of the subject's greatest fascinations is the way that new data are already beginning to call into question some of Cavalli-Sforza's conclusions. For although he does discuss the DNA sequence data obtained from mitochondrial DNA, and mentions the exciting new work only now becoming available on Y-chromosome data (transmitted in the male line), their impact cannot yet be fully assessed. My impression is that these new data are now showing that much of the genetic composition of the population of each continent was already established more than 10,000 years ago during the palaeolithic period, following the initial "out of Africa" expansion of our species something like 60,000 years ago.

Cavalli-Sforza likens his principal components to archaeological strata: "It thus appears that the chronological order of the expansions is approximately reflected in the order of the principal components." But as we have seen, he assigns the first principal component for Europe to the spread of farming some 8,000 years ago. That does not tally well with the DNA sequence evidence (from both mitochondrial DNA and Y-chromosome work) that some of the most significant dispersals took place nearly 10,000 years earlier, with others 20,000 years before that. It is clear that when these new results are assimilated there will be a great deal more to tell about the human story and the origins of human diversity.

That for me is the attraction and the fascination of this readable book, which is the best available introduction to some of these great issues. Some of its conclusions may be questionable, but that in a way is its strength. In it we see the coherence of some important ideas developed by a mind with a remarkable ability for synthesis on a broad scale. Cavalli-Sforza is one of the great founding fathers of this new discipline of archaeogenetics, and in Genes, Peoples and Languages he maps out some of its grand themes.

Lord Renfrew is professor of archaeology and director of the McDonald Institute for archaeological research, University of Cambridge.