Subjectivity

SUBJECTIVITY W. P. Lehmann The University of Texas at Austin Historical linguistics has primarily used phonological evidence for genealogical classification. The evidence consists of 'characteristic facts and details', in Meillet's terms. Syntactic evidence can now be used. Subjectivity, i.e. the subjective dominance of the principal verb in a sentence, is cited as an illustration of a 'characteristic fact' in Proto-Indo-European and its dialects.* In this last year of the first half-century of the Linguistic Society, we may be permitted to observe that linguistics has made notable advances toward its fundamental aim : the understanding of language. One of the goals of the founders of the Society has been fulfilled : no longer ' viewed as a subsidiary discipline ' (Bloomfield 1925:1-5), linguistics is now recognized as an independent science. This recognition has been gained in part through remarkable developments in theory, accompanied by intensive investigation of language, which have led in turn to confidence in confronting linguistic problems. The greater assurance of our field has also led to fuller understanding of our possibilities, as (like our colleagues in the natural sciences) we have learned from Heisenberg that the mere activity of a scientist modifies the object of his investigation. Further frank commentary on scientists, such as identifying them either as 'geometers' or 'algorists', as might a physicist (Schild 1962), or as nominalists or realists, as might a philosopher, seems still beyond us. Yet even lacking such sophistication, we are beginning to admit the usefulness of various theoretical approaches. In the development of these we have come closer to the aims of the Foundation Members of the Society than to their practical goals—such as thorough description of the American Indian languages, or of American English—or to obligations regarding the 'public interest' (Bloomfield , ibid.) These goals too are increasingly pursued, with promise of rapid advances , especially if we can involve the talents of our young members and students in meeting them. In the midst ofthese theoretical and practical achievements, it is strange that the sub-branch of our science which made the greatest advances during the previous century should have lagged far behind. Historical linguistics, at least until quite recently, did not advance in theory beyond that of linguists in the 19th-century tradition. New insights were of course achieved concerning particular problems, such as the development of the phonological system of Proto-Indo-European, with the loss oflaryngeals. But virtually all such insights, as well as theoretical discussions in historical linguistics, were limited to phonology, apart from the rear-guard actions that may be expected from an overwhelmed activity. The self-selected limitation has now been overcome: historical linguistics can now concern itself with more central components of language, as I intend to illustrate here. The impasse which historical linguistics, had reached was recognized and de- * This paper represents the Presidential Address delivered at the annual meeting of the Linguistic Society of America, in San Diego, on 28 December 1973. I am grateful to several scholars, and especially to Use Lehiste, for comments on this paper. 622 SUBJECTIVITY623 plored by distinguished linguists such as Louis Hjelmslev (1935) and W. Sidney Allen (1953), and little improvement was expected by either. The basic situation must be clear to students who have taken courses in historical linguistics or dealt with the history of a language: historical linguists were occupying themselves almost exclusively with surface features of language. Conclusions regarding language classification, for example, were based on lexical and phonological evidence. The lexical evidence was taken from the 'common words' of a language, not from specific characteristics of highly structured semantic fields such as kinship terms. The phonological evidence was often minute; thus Celtic was associated with Italic because of the treatment of? ... kw- in three words, along with phonological evidence scarcely less tenuous. Both were aligned with Tocharian on the basis of /•-endings in certain verb forms (see also Cowgill 1970). I am not concerned here with the validity of these observations, but rather with the kinds of evidence on which the conclusions are based. If, as theoreticians agree, the phonological component is a non-central segment of language, then historical linguists have limited their concerns to the periphery...