"The arbitrary and unrigorous methods that have characterized much of this linguistic paleontology certainly deserve Renfrew's scepticism. . . . Most of the lexemes that can be confidently assigned on the basis of widespread attestation . . . do not tell us much."