Modern Linguistics

Abstract

T NHE demonstration in I786 by Sir William Jones that Sanskrit offered a close relationship to the Classical, Germanic, and Celtic languages ushered in the first major phase of what might be termed modern linguistic studies. The fact that Sanskrit was an extensively documented language, with a highly developed grammatical system and phonological and morphological sub-systems that were at once archaic and transparent, provided the impetus for a century of intensive work in the comparative reconstruction of the parent IndoEuropean language. As research during this period continued, earlier metaphysical and culturally biased attitudes were replaced by a spirit of objective scrutiny whose aim was the discovery of statable correspondences among the forms of different languages and the formulation of generally valid principles of linguistic development. Although there was frequent contention throughout this period from certain scholars who looked upon the demonstrated interlingual correspondences as mere interesting artifacts of history, those in the mainstream of research felt compelled to recognize fundamental principles at work. Toward the end of the nineteenth century, the views of the latter group prevailed. Epitomizing the position of the neo-grammarians, as the ascendant group has come to be called, is the enunciated by August Leskien, according to which sound-changes admit of no exceptions. This law may be regarded either as a factual claim about how sound-change operates or as an article of methodological faith. Taken at face value, it maintains that if a given sound, say p, changes in a language to, let us say, f, then each and every occurrence of p in that language undergoes that particular change. This law was formulated on the basis of numerous observations showing that related words in cognate languages (or, as a special case, in two stages of the same language) displayed regular phonological correspondences. If in related words language X1 showed p and language X2 showed f (and assuming there was evidence to establish X1 as reflecting an earlier stage of development), then the inference was that the inherited p in language X2 had changed tof. On this question there could not very well be any argument. But if, despite many correspondences of this kind, language X2 should (among other possibilities) continue to show p instead of the expected f, then the point