Apes, Wolves, Birds, and Humans: Toward a Comparative Foundation for a Functional Theory of Language Evolution

SLS 14 (1977), 21-58 Q William C. Stokoe APES, WOLVES, BIRDS, AND HUMANS: TOWARD A COMPARATIVE FOUNDATION FOR A FUNCTIONAL THEORY OF LANGUAGE EVOLUTION Jane H. Hill A major goal of the modern study of communication is to unite the study of human communication systems, animal communication systems, and other information-carrying systems in nature into a single discipline.' One aspect of such a unified theory is a theory of communication systems as modes of adaptation, integrated in a complex manner into the total mode of adaptation of a group of organisms, and changing under pressure from the mechanisms of evolution . Such a theory has yet to be developed; one of the main reasons it has not been developed is the important difference in orientation which separates most students of human language from most students of communication in other animals. Linguists tend to be "formalists"; animal behaviorists tend to be "functionalists". This prevailing difference in orientation derives in part from fundamental differences in the nature of our access to the evidence. Linguists work on data about which they, as human beings, "know" a great dealalready. Because of this "knowledge", the descriptive units of even the most explicitly behaviorist linguists tend to take human cognitive and perceptual apparatus for granted. For instance, consider the concept of the "phone", the descriptive unit in phonetics. It is well known that the physical nature of human speech is continuous and gradient; sounds shade into one another, and there is a continuous blur from one acoustic pattern into the next. However, humans appear to hear sounds as "discrete" segments of sound, probably as a result of the application of a very complex series of perceptual strategies involving attention to evidence from order and contrast. Linguists developing "linguistic" as opposed to "acoustic" descriptions of language use such discrete segments as basic units. Even Sign Language Studies 14 linguists who use very low-level descriptive units, such as distinctive features, assign these units to segments; when distinctive features occur in more than one adjacent segment, the boundary of their application is a segment boundary. Although phoneticians are fully capable of specifying such facts as "the onset of voicing began gradually at the third millisecond of labial closure, " such gradient statements do not appear in linguistic (as opposed to acoustic or physiological) phonetic description. Linguists do not use gradient units because they do not agree where such gradience is a part of "language", in the sense of langue or linguistic competence, and where it is a feature of physiology or anatomy or of the intrinsic physical properties of sounds. Attempts to specify an empirical basis for the determination of where acoustic and physiological facts end and where the cognitive fact of language begins have been a focus of theoretical discussion in linguistics for many years. Animal behaviorists, on the other hand, do not "know' how animals perceive or what their pragmatic purposes might be. The only cognitive and perceptual units to which they have access are human units, and these are used only with great caution. 3 Therefore, formalisms in animal behavior studies have been far more empirical and "behaviorist" than those of linguistics. The acoustic study of sound, as in sound spectography , is used by linguists, but is considered to be inadequate in itself, because it does not tell us many things that we "know" about the signals of language. However, such acoustic studies are the most important foundation of the description of animal sounds, which are specified in terms of such variables as frequency range, pulsation rate, and harmonic structure alone. Consideration of visual displays, in a tradition dating from Darwin (1872), is usually stated in terms of the musculature involved. When animal behaviorists begin to specify structural units at the level of contrast, they shift from this very specific empirical level to a kind of eye-ball taxonomy that would not be considered particularly well-developed in linguistics, using units like "shriek", "whimper", "pout face", and the like. These units are certainly useful, but the study of the communicative repertoire of an animal species in these terms, while sophisticated, is perhaps more like the study of style in the arts than it is like...