Indo-European

Indo-European reaches from Ireland to Bengal, and its recognition in the late eighteenth century created historical linguistics as a field. William Jones, speaking in Calcutta in 1786, observed that Sanskrit bore to Greek and Latin a resemblance stronger than accident could explain, and proposed that all three had sprung from a common source which perhaps no longer exists. He was not the first to notice the resemblance, but the programme he described is what the nineteenth century then carried out. Regular sound correspondences were worked out, the comparative method was built to formalise them, and by the end of the century a reconstructed ancestor existed on paper.

Today the family holds around 450 languages and more speakers than any other, a fact that owes as much to colonial expansion as to anything older. Its branches include Indo-Iranian, Greek, Italic, Celtic, Germanic, Balto-Slavic, Armenian, Albanian, and the extinct Anatolian and Tocharian groups. Anatolian, recovered from Hittite cuneiform tablets only in the twentieth century, forced a substantial revision of the reconstruction, which is a useful reminder that a reconstructed proto-language is a hypothesis maintained under evidence rather than a discovered artefact.

The speakers of Proto-Indo-European left no texts, no inscriptions and no identifiable temples, and almost everything proposed about their religion comes from comparing the vocabularies of their descendants. The method is simple to describe. Where the same word appears in branches that separated early and had no later contact, and where it appears with the same religious sense, the word and the concept are both likely to be inherited. The most secure result is a sky father whose name is reconstructed from Sanskrit Dyaus Pitar, Greek Zeus Pater and Latin Jupiter, all of them the same phrase.

Beyond that, comparative work supports a set of formulae used in praise poetry, a vocabulary of sacrifice and of libation, and a divine pair of horse twins. The method has real limits and they should be stated. It recovers words rather than beliefs, it cannot date what it recovers, and it has been pushed far past what the evidence supports: the trifunctional scheme associated with Georges Dumezil, which sorts Indo-European society and its gods into three fixed classes, remains influential and remains contested. Reconstructed religion is also the part of this field most often taken up for political purposes, which is a reason to be careful with it rather than to avoid it.