Phonemics and Phonics in Historical Phonology

Abstract

MY PURPOSE IN THIS ARTICLE is to discuss the interplay of phonemic and phonic data in historical phonology, and to demonstrate by examples chosen from the collections of the Linguistic Atlas the importance of giving adequate attention to phonic data not only in historical phonology but also in synchronic linguistics. To achieve my purpose, I must first outline my view of the field of linguistics, even at the risk of making the tail wag the dog. As I see it, linguistics has two major fields: pure linguistics, whether synchronic, diachronic, areal, or social, which deals with usage without reference to the history of the speaker or the speech community; and historical linguistics, which undertakes to account for usage, or change in usage, by correlating purely linguistic data with the character and the history of the speaker or the speech community. Pure synchronic linguistics (glottotechnics) proceeds on the assumption that in a given society, at a given time, the sounds of the current language constitute a system. Operating on this theory, the descriptivist establishes the functional units of sound (the phonemes) and formulates the system of sounds of the language in question. Focusing his attention upon systematics, he stands the best chance of finding all the systematic features peculiar to it. In this pursuit he sets aside, quite properly, all historical considerations and the culture of the community in which the language is used. The feasibility of abstracting from historical and social realities is, indeed, one of the crowning features of this method of dealing with language. To say that language is essentially systematic at any given time, and therefore amenable to the descriptive technique, is not to admit that it is ever wholly systematic. The findings ofdiachronic and of area linguistics sharply contradict any such assumption. Relics of older usage that no longer fit into the current system and piecemeal innovations not yet systematized are ever present. Moreover, the descriptivist cannot ignore the patent fact that systems do change, and that a change in the system of a language is not conceivable without a temporary breakdown of certain structural features. The proper procedure in synchronic linguistics is, therefore: (i) to identify all the systematic features; and (z) to isolate those features that are no longer, or not yet, systematized. To invent ingenious formulas in order to make irregularities look regular is a futile exercise, to say the least, since it is based upon a misconception of the nature of language, which is a historically created system of communication.