Ringe Donald A. On Calculating the Factor of Chance in Language Comparison

American Philosophical Society, 1992. — 116 p. — (Transactions of the American Philosophical Society 82.1).The consensus of opinion among mainstream historical linguists is that while all human languages are likely to be genetically related, the remoter relationships cannot be demonstrated by reliable linguistic methods because the languages in question have diverged too much. From time to time this conventional wisdom is challenged by scholars who claim to have demonstrated one or more remote relationships; recent challenges include GREENBERG 1987 and SHEVOROSHKIN 1989. The purpose of this paper is to provide an objective test of the validity of such challenges.
Any demonstration of a relationship between languages depends largely on finding words and grammatical affixes of systematically similar shape and at least roughly equivalent meaning in the languages in question. But if such a demonstration is to be convincing, one must show that the similarities adduced could not have arisen by chance. Unfortunately the possibility of chance resemblances is often dismissed without adequate discussion, apparently because common sense suggests that there is very little likelihood of words in different languages being strikingly similar in both form and meaning as a result of sheer chance.