Iconicity, Arbitrariness, & Duality of Patterning in Signed & Spoken Language: Perspectives on Language Evolution

Armstrong 51 ICONICITY, ARBITRARINESS, & DUALITY OF PATTERNING IN SIGNED & SPOKEN LANGUAGE: PERSPECTIVES ON LANGUAGE EVOLUTION David F. Armstrong Introduction. One of the hallmarks of human languages is their ability to incorporate signs that have no obvious physical relationship to the objects, actions, ideas, emotions, or other entities to which they refer. Languages also have the capacity to incorporate signs with varying degrees of relatedness to their referent. I use sign here and not word, because I intend for this discussion to include signed languages as well as spoken languages. I will use the general term sign to refer both to the words and morphemes of spoken languages and to the gestural signs of signed languages. The present discussion will be oriented toward elucidating the nature of arbitrariness of relationship between sign and referent in signed and spoken languages and will include observations about the biological bases of this phenomenon. (c) 1983 Linstok Press ISSN 0302-1475 38:51 3.00 SLS 38 Spring 1983 Armstrong 52 Iconicity, arbitrariness, & duality. The most famous and probably still the best statement of the notion of arbitrariness in the relationship between sign and object is Shakespeare's famous line in Romeo and Juliet (II:ii): "A rose by any other name would smell as sweet." There is an almost equally famous rebuttal by Gertrude Stein ("A rose is a rose is a rose..." -- more about that below). The terminology used here to describe the relationship between sign and object is that introduced by C. S. Peirce (citations from Buchler ed., 1955: 102f). Peirce identifies three types of sign: the icon, the index, and the symbol. An icon is a sign for an object by virtue of its physical resemblance to the object itself. An index is a sign by virtue of physical effect upon or contiguity with its object -- pointing is an example of an indexical sign. Finally, a symbol is a sign for an object by virtue of a rule or convention of association -ordinary words should commonly fall in this latter category. Clearly when we say that a word has an arbitrary relationship with its referent, we are saying that it is a symbol. By the same token, some gestural signs of sign languages are often called iconic or indexic. What will be examined here is the extent to which these assertions are correct and the extent to which the relationship among signs and referents are biologically based. Also important here is the concept of duality of patterning: A communicative system has duality of patterning if its meaningful signals (pleremes) are built out of some convenient stock of meaningless but differentiating pieces (cenemes) which are subject to constraints on arrangement partly independent of any such constraints on the pleremes.... Every human language has duality of patterning; the cenematic pattern is the sound system. (Hockett 1978: 275) spring 1983 SLS 38 Armstrong 53 I will argue below that there is some question as to the extent of independence between the phonemic and morphemic levels in spoken languages, but Hockett has made it clear that this concept is flexible by maintaining that the independence need only be partial for duality to be present. If sign languages possess a high degree of iconicity, may they be said to have duality of patterning? Poizner (1981) has addressed this question; and Hockett further says: ... I think Ameslan has duality of patterning. Stokoe's description of 1960 was already pretty convincing (or, I should say, I found it so when I got around to giving it a careful reading). Supporting evidence comes in more recent studies, such as Bellugi and P. A. Siple (1973, cited with just approval by Frishberg, 1975: 697): in tests of short-term memory, errors made by users of Ameslan are as apt to derive from the physical similarities of signs as from their semantic affinities. (op. cit., 276) It is important to note that in the final phrases of the passage above Hockett appears to refer to the appearance in ASL, as apparently in other sign languages (see Umiker-Sebeok & Sebeok 1978: xiii-xxx), that some aspect of a sign bears a fairly obvious, i.e. iconic relationship to its referent. An example...