Although language change has now been studied systematically for a period of one hundred years and somewhat less systematically for a good deal longer than that, there is still a considerable amount of disagreement about its nature and motivation. I have for this reason made no attempt at presenting a theoretically unified account but have rather thought it more useful to describe in chronological sequence the three major models that have been proposed in order to account for the phenomena of language change, that is to say the neogrammarian, the structuralist, and the transformational-generative models. It is my hope that this approach, by making explicit the major points of difference which separate these three schools of thought and allowing the reader to assess their respective merits and weaknesses, may perhaps go some small way towards closing the gap which has arisen between traditional comparative philology and modern theoretical linguistics.

I owe a lasting debt of gratitude to my late teacher Hans Krahe who first introduced me to the rigours of neogrammarian methodology. I am equally indebted to both colleagues and students at the School of Oriental and African Studies for exposing me to the no less rigorous methods of structuralism and transformational grammar and for giving me the opportunity of discussing problems of linguistic comparison outside the Indo-European field. R. H. Robins and Eugénie Henderson in particular were kind enough to read through an earlier version of Part I and to make valuable suggestions.