J. Marshall Unger
An internal reconstruction of Hungarian based mostly on the morphophonemic alternations of the modern language yields dramatically different results from those obtained from a purely synchronic approach to the same data. This article presents such a reconstruction and argues that the striking contrast between the forms and processes it entails and those favored by a synchronic analysis demonstrates the need to give performance (parole) at least as much consideration as competence (langue) in the study of language history.