The vast geographical spread of the Indo-European (IE) languages is rivalled only by that of the Malayo-­Polynesian ones. Explaining the causes of the enormous landward spread of IE languages requires careful reconstruction of long-lost words from an era when writing was unknown. This can be done thanks to the modern understanding of sound shifts, which makes it possible to bring to life the vocabulary of people living six thousand years ago. Efforts to do so are just as exciting as attempts to revive hairy mammoths of the same time period by manipulating elephant DNA. 

DNA is an important part of the story of the IE languages, as Laura Spinney shows in her lively and fascinating account of the origins and spread of this language group. Tracing the movement of peoples by following that of languages is problematic. In the last century, the eminent geneticist Luigi Cavalli-Sforza optimistically thought that, by tracking the spread of both languages and DNA, he could map the movement of peoples in the remote past. Nowadays, historians tend to be cautious in attributing a single identity to groups of ancient migrants. The Germanic Vandals, for instance, were accompanied on their vast trek from northeastern Europe to Roman Tunisia by Indo-Iranian-speaking Alans who fused with them at some point. Ethnic purity is a myth and we are all mongrels.

Proto-Indo-European (PIE), the language reconstructed by philologists and the common ancestor of IE languages, was itself a hybrid. Asking what came before it is rather like asking what came before the Big Bang: we have no real idea. Quite a few words of PIE are recognisable. One such is *dhugh₂ter,