ABSTRACT
Because man is the only creature with a known natural language, we tend to assign a definitional weight to every aspect of human language. Yet it is equally reasonable to suppose that only some properties of human language are definitional and that others are secondary, adaptations to limitations of human information processing. The phoneme is surely such an example. Although it may be unique to man, and therefore one of several ways to distinguish human language from animal communication (Hockett, 1959), it is not a necessary condition for language. No syntactic or semantic distinction depends on whether the primitive level of language is the word or an element below the word. To assign a definitional weight to a property of language solely on the grounds that human language has the property is to adopt an unjustifiably ethnocentric view of language. It is tantamount to assuming that the only possible form of language is the human one.
