Impact of Greenberg's Work on Present and Future Research
Luca Cavalli-Sforza
Genetics Dept., Stanford University

I have been very fortunate in learning about Greenberg's taxonomic work at a crucial time in our work on the reconstruction of human genetic evolution. I also have a great debt of gratitude with Merritt Ruhlen, who has patiently helped me to understand it and his own work in Greenberg's footsteps. It is getting more and more clear from the joint consideration of archeological and genetic data that the history of modern humans began with an expansion of an East African population to the rest of Africa around 100,000 years ago, followed by a secondary expansion from a related population, about 50,000 years ago, to Asia, most probably via two major routes: one along the coast of South Asia to South East Asia and from there to Oceania and Japan, and another to central Asia and from there westwards to Europe, and towards the east and northeast, via Siberia to America, the last continent to be settled. Language has certainly played a major role in this expansion, given the importance of good communication in the settlement of new continents. Greenberg realized that human languages may have had a single origin, and gave an example with the universal root TIK. Later work by Bengtson and Ruhlen followed this cue and supported Greenberg's intuition. But the rate of transformation and differentiation of languages is so high that one can sympathize with the caution exercised by the majority of other linguists, who usually prefer to think that the problem is insoluble. As a geneticist, however, I am convinced that the languages we speak today have had a single origin, because the original modern human population from which we all descend was very small and must have spoken a single language. Moreover, the way genetic novelties concerned with communication spread in populations suggests that new biological mechanisms favoring it must have spread in a small population. It is interesting, therefore, that there is a close similarity between what we know of the evolutionary tree of languages and the genetic one. Darwin had clearly foreseen this when he suggested that if we knew the genetic tree of humans we could also predict that of languages. The similarity of the two trees has been extended also to other cultural data, and to the molecular evolution of the oldest bacteria and viruses: wherever there is at least partial vertical transmission (from parents to children) in cultural or epidemiological traits, one is likely to find a similarity of their evolutionary tree with that of genetic traits, which are transmitted strictly vertically. This is because the routes of the expansion of modern humans have been the same or very similar, and similarity must also show for all traits that have a comparably long evolutionary history.

Greenberg apparently asked for suggestions for future research. In my view, it is important to expand on inter-disciplinarity - this is almost a trivial suggestion today, but I think the work done on human genetic evolution shows that this is a fruitful avenue. I also believe that it will be important to improve on the quantification of analytic methods, using more accurate measurements, on various aspects of linguistic evolution. I discussed with Greenberg the possibilities of measuring semantic similarity. I realize this is very difficult, but if one looks at the semantic changes which have accompanied the changes of the root TIK, as Greenberg proved, (to show, to indicate, to point, arm, hand, finger, one), it is necessary to agree that one can obtain much more satisfactory measurements of linguistic evolution, only if one studies jointly and more accurately from a statistical point of view semantic, phonetic and all other changes accompanying it.