A formalization of one of the steps in the comparative method in historical linguistics; i.e., recognition of modern derivatives of prehistoric phonemes. The theory assumes that words of a hypothetical prehistoric language should be constructed so as to minimize the total number of phonemes in the language and the total number of statements needed to account for the forms of modern words. It provides an algorithm for a computer program, but even small increases in the amount of data to be considered exhaust present computer techniques. The theory nevertheless provides a basis for more efficient heuristic procedures.
Kay, Martin, The Logic of Cognate Recognition in Historical Linguistics. Santa Monica, CA: RAND Corporation, 1964. https://www.rand.org/pubs/research_memoranda/RM4224.html. Also available in print form.
Kay, Martin, The Logic of Cognate Recognition in Historical Linguistics, RAND Corporation, RM-4224-PR, 1964. As of February 16, 2024: https://www.rand.org/pubs/research_memoranda/RM4224.html