Duality of Patterning: Responding to Armstrong & Stokoe

Summer 1987 DUALITY OF PATTERNING: RESPONDING TO ARMSTRONG & STOKOE Edwin G. Pulleyblank In response to the comments of Armstrong and Stokoe on "The Meaning of Duality of Patterning & Its Importance in Language Evolution" (Sign Language Studies 51), it is evident that I have not succeeded in making clear, at least to my two critics, the fundamental difference that I see between the inherent duality of patterning found in all human spoken languages and such organizational principles as can be discerned in other symbolic systems that, in my view, lack this characteristic. I should like to comment briefly on one or two of the points that were raised: Armstrong defines the problem of demonstrating "that a communication system, say a signed language, has the property of duality" as requiring, firstly, "selection by signers of a relatively small number of the handshapes, facial expressions, etc. that could constitute minimally contrasting pairs. This stock of elements should be relatively fixed (i.e. introduction of new elements should be gradual)" (SLS 51, 122f). I do not know quite what is meant by "relatively" here. I would venture to say that, in principle, at any one synchronic stage, the phonemic inventory of a spoken language is, quite simply, fixed. In practice slippage does occur. Speech communities are never completely uniform, and the clashes that this produces can interfere with the phonological systems of individual speakers. The passing of the phonological system from one generation to the next is another constantly recurring source of potential variation. The most common type of language change involves, however, not the addition of phonemic elements but the loss of distinctions througy mergers; e.g. the loss of distinction between wh and w, which has already affected major dialects of English and will no doubt spread to Copyright 1987 by Linstok Press, Inc. See note inside Front cover. ISSN 0302-1475 SLS 55 Pulleyblank : 176 the rest of the language community eventually -- or the merger of syllabic [n] with syllabic [m], which has occurred quite recently in Cantonese as spoken in Hong Kong. Change in pronunciation of a phoneme -in a specific environment may also occur without loss of distinction, as when [k] became palatalized to [tf] before high front vowels in words like child in Old English. Such changes may eventually lead to the splitting off of a new phoneme from an old. Simple addition of new phonemes, as opposed to the splitting of old ones through conditioned sound change, is not "relatively gradual" but very rare and exceptional. In his remarks on Armstrong's 1983 paper, Hewes claimed that a language can add to its stock of phonemes through borrowing, but as I showed with reference to the example he cited from Middle English, even the possibility of this is severely circumscribed by the need to fit the new item into the existing phonological system. This is because phonemes, however few in number they may be in a given language, are not simply a "stock of elements," each more or less independent of the other, but a highly integrated system based on combining an even smaller number of distinctive features according to rigid rules. Thus, the sound [ts] in English is a combination of two phonemes, which can occur at the end of a word but not at the beginning. In many other languages (e.g. German, Russian, Italian, and Chinese) it is a single phoneme. Native speakers of English can quite easily learn to make this sound. Nevertheless, when words are borrowed from these languages into English, the stock of English phonemes is not enriched. Instead the foreign affricate is deformed. Tsar is pronounced [zarl. The local place name Tsawwassen, near Vancouver, is pronounced with initial [tJ, not [ts]. In Japanese the situation is even more Summer 1987 SLS 55 Pulleyblank : 177 complicated. The sound [ts] is neither a phoneme nor a cluster of two phonemes. It is the allophone of the phoneme /t/ before the vowel /u/. Hence, the English word tool becomes tsuru when borrowed into Japanese. Armstrong makes a point of the fact that in faceto -face communication one can introduce audible non-speech signals; e.g. a lip buzz resembling the sound of breaking wind...