Martin West, who has written what is surely the definitive book on Indo-European language and religion, states his case well: ‘The assumption of a single parent language as the historical source of all the known Indo-European languages . . . is still a hypothesis, not an observable fact, but it is an inescapable hypothesis.’ The Indo-European map links languages together in a group that is distinct from other groups, such as those that include Chinese or Tamil, say. The evidence that the Indo-European languages are related lies primarily in their grammar and vocabulary. Thus ‘foot’ is pada in Sanskrit, pes, pedis in Latin, pied in French, fuss in German, foot in English and so forth, and nouns and verbs behave entirely differently from their Hebrew or Navajo counterparts.
Indo-European linguistics assumes a diffusionist, centrifugal cultural movement: the political centre sends out armies and imposes its rule on neighbouring lands. The paradigm is Latin, which did indeed diffuse outwards to all the lands the Romans conquered, which therefore speak languages that we call Romance. Linguists then constructed, on the Roman model, an earlier family tree diverging from the centre, in this case not Rome but the Caucasus (or somewhere else in Central Asia). West calls the original common territory ‘Eurostan’ and remarks: ‘If it be asked what sea the worshippers of these prehistoric divinities went down to in *nawes and sailed on and foundered in, the likely answer is the northern Black Sea or the Sea of Azov.’ The mythical land of the family home might just as well be thought of as *Indo-Europe, the land east of the asterisk.
more from the LRB here.