Pinker's assertion leads naturally to consideration of a fundamental philosophical and scientific issue—when we talk about the "origin and evolution" of language, exactly what is the topic of our discussion? Language is not a trait like eye color or stature, but an aspect of human behavior. The human ability to express and comprehend language and the kind of language that is used depend upon the presence and functioning of a variety of anatomical structures and physiological systems, including the senses of hearing and vision, the upper respiratory system, the hands and arms, and so on. Evolution occurs through changes in the frequencies of the genes that determine, through interaction with the environment, the functioning of such systems. Therefore, we must inevitably define not only language itself but also what is entailed by evolutionary change with respect to its expression. A recent debate between Pinker and the evolutionary biologist Stephen J. Gould reveals some of the most basic problems in this endeavor.
On several occasions, Gould has criticized attempts to "explain" various aspects of human behavior in terms of adaptation. He has been particularly critical of the use of overly clever arguments to show genetic determination and a basis in natural selection for virtually every bizarre human custom. Against these arguments, Gould maintains that a variety of structures possessed by organisms are merely by-products of the presence of other structures, and that they are therefore not present because of the action of natural selection. He terms these "spandrels," after an architectural term for the space between two arches or the space between an arch and the rectangular framework surrounding it. Although the debate is quite general, it is easy to see that language might in some sense be such a spandrel—perhaps it is simply the by-product of an expanding brain and growing cognitive power. Pinker, on the other hand, takes the position that many human behavioral systems, including language, are "modular"—that is, they are analogous to independent, plug-in electronic components and under quite specific genetic control. Pinker also asserts that languages are adaptive in the sense of having evolved under natural selection.
When two such obviously intelligent scholars disagree over matters this fundamental, the prudent student heads quickly for the middle ground. Neither metaphor, the spandrel nor the module, seems particularly apt when applied to language. By any reasonable standard, language clearly is more than a by-product of the evolution of other human traits—it is at the very center of human social activity, arguably the most basic human adaptation. The numerous anatomical and physiological systems currently emerging are best explained as products of coevolution with language. It should seem equally clear, however, that language is similarly not instinctive or modular—it involves a large number of physiological systems, and its processing and production involve much of the brain. Moreover, language shares certain features in common with other aspects of human behavior, and it shows considerable variation both among populations and historically.
At the outset of this book, it is necessary to draw the reader's attention to an issue of terminological usage. The word sign as used in this book generally refers to visible actions of people using what are usually called signed languages, for example, American Sign Language. However, at several points in the book the topic of discussion is the formal study of signs in a generic sense, or semiotics. In this sense, the word sign refers to elements of communication systems, and specifically, to things that stand for or indicate other things. In this sense, the words of spoken languages are signs, as are the signs of signed languages and several other classes of signifying activity such as pointing gestures. In general, the sense in which the word is being used should be clear from its context.