It is fortunate that we have such a refined and specific innate endowment. Were this not so, each individual would grow into some kind of amoeboid creature, merely reflecting external contingencies, utterly impoverished, and lacking the special structures that make a human existence possible. Naturally, the same innate factors that permit the organism to transcend environmental factors, reaching a remarkable level of complexity of organization that does not 'mirror' the limited environment, rule out many possible courses of development and limit drastically the final states that can be reached in physical growth.
Little is known about how any of this happens, but no one seriously doubts that something of roughly this sort is true. Why? Because of the vast qualitative difference between the impoverished and unstructured environment, on the one hand, and the highly specific and intricate structures that uniformly develop, on the other.
Turning to the human mind, we also find structures of marvellous intricacy developing in a uniform way with limited and unstructured experience. Language is a case in point, but not the only one. Think of the capacity to deal with abstract properties of the number system, common to all humans apart from gross pathology, and, it seems, unique to humans. The essence of this system is the concept of adding one, indefinitely. The capacity is not 'approached' by the ability of some birds to match patterns of n objects for some finite (and small) n, just as human language with its discrete infinity of meaningful expressions constructed by abstract rules crucially involving operations on phrases is not simply 'more' than some finite system of symbols imposed on other organisms — or, for that matter, just as the ability of a bird to fly, though finite, is not simply an extension of the human ability to jump; whatever the evolutionary history may have been, quite different mechanisms are involved. The capacity to deal with the number system or with abstract properties of space — capacities that lie at the core of what we might call the human 'science-forming faculty' — are no doubt unlearned in their essentials, deriving from our biological endowment. One can think of many other examples.
These systems have many of the relevant properties of physical organs. We might think of them as 'mental organs'. Thus the human language faculty might well be regarded on the analogy of the heart or the visual system. It develops in the individual under the triggering effect of experience, but the mature system that grows in the mind (that is, 'learned', to use the standard but misleading term) does not 'mirror' the contingencies of experience, but vastly transcends that experience. True, there are differences among individuals contingent on experience; some know English, others Japanese. Similarly, onset of puberty varies over some range, as does body size, or the ability to pole vault, or the distribution of cells of the visual cortex specialized to respond to lines of particular orientation in the visual field. But the pole vaulter will never fly like a bird (even a chicken), and the human language faculty will never grow anything but one of the possible human languages, a narrowly constrained set.
In brief, our genetic endowment provides for the growth and maturation of special mental organs, the language faculty being one. The development of these systems is essentially uniform among individuals. Two people from the same speech community can converse freely on some topic new to them despite substantially different experience, despite the fact that the sentences they produce and understand bear no direct analogy to anything that they have heard. Their minds contain roughly comparable rule systems of highly specific structure determined in general character by some property of the human species. These rule systems cannot be derived from the data of experience by 'induction', 'abstraction', 'analogy', or 'generalization', in any reasonable sense of these terms, any more than the basic structure of the mammalian visual system is inductively derived from experience.
As in the case of the physical body, we are fortunate to have this rich innate endowment. Otherwise, we would grow into 'mental amoeboids', unlike one another, merely reflecting properties of the impoverished environment, lacking the finely articulated structures that make possible the rich and creative mental life that is characteristic of all humans who are not severely impaired by individual or social pathology. These same innate factors provide the basis for a social existence in common with others whose capacities are not unlike our own despite accidents of individual history. We live in a world of shared understanding that extends far beyond the limited experience that evokes cognitive structures in the mind.
Again, the very same innate factors that provide for the richness and variety of mental life, shared with others comparably endowed, impose severe bounds on what the mind can achieve. Scope and limits are intimately related. Our inability to fly like birds derives from the same innate properties that enable us to become humans rather than amoeboid creatures. Comparably, the fact that there are many imaginable languages that we could not develop through the exercise of the language faculty is a consequence of the innate endowment that made it possible for us to attain our knowledge of English or some other human language. Similarly, the fact that there are no doubt many systems of musical organization that we simply could not comprehend or enjoy should be a source of satisfaction, because it reflects the same innate endowment that enables us to appreciate Bach and Beethoven. And the fact that many possible branches of science lie beyond our cognitive reach should cause no dismay when we realize that it results from that innate science-forming capacity that permits the construction of intelligible explanatory theories on weak and limited evidence in at least some domains of thought. And so on.
This talk of 'mental organs' should not mislead. We can discuss a physical organ — say the visual system — in terms of its abstract properties, knowing little about its physical realization. Nothing more than this is implied when one speaks of the mind as a system of mental organs, or when one studies these organs and their interaction as systems of mental representation and mental computation.
How can we proceed to gain insight into the specific properties of particular mental organs? Consider the case of language. There are three basic questions that arise:
(i) What do we know when we are said to know a language?
(ii) What is the basis for the growth of this knowledge?
(iii) How is this knowledge put to use?
The answer to the first question seems to be that knowledge of a language is mentally represented as a 'grammar' — that is, a finite system of rules and principles that interact to determine ('generate') an infinite class of expressions, each with a phonetic form, meaning, and associated structural properties (for example, an organization into words and phrases). As for (ii), it seems that many of the fundamental properties of these grammars are part of innate endowment, so that the child in effect knows in advance what kind of grammar he must construct and then must determine which of the possible languages is the one to which he is exposed. Study of (iii) leads to the construction of 'performance models' that have access to the grammar — the knowledge of language — represented in the mind.
To illustrate with a simple example, consider the reciprocal construction in English: such sentences as 'The men saw each other', 'I asked them about each other', 'We shot arrows at each other', etc. The rule of grammar governing these constructions specifies that the expression 'each other' requires a plural antecedent. Once the antecedent is found, we apply the dictionary rule of interpretation to fix the meaning: roughly, that each of the men saw the other men (or man), etc. However, it is not always so easy to select an antecedent. Sometimes it lies in a different clause, as in 'The candidates wanted [each other to win]' or 'The candidates believed [each other to be dishonest]'. In these sentences the bracketed expression is a subordinate clause with 'each other' as its subject, just as in 'John wants [Bill to be successful]' 'Bill' is the subject of 'be successful' in the bracketed subordinate clause. The antecedent of 'each other' lies outside the clause. But we cannot always select an antecedent outside of the subordinate clause. Consider 'The candidates believed [each other were dishonest]' (compare 'The candidates believed [their opponents were dishonest]') or 'The candidates wanted [me to vote for each other]' (compare 'The candidates wanted [me to vote for them]').
Such facts as these are known to all speakers of English, and analogues appear to hold in all other languages. The facts are known without experience, let alone training. The child must learn that 'each other' is a reciprocal expression, but nothing more, so it seems. No pedagogic grammar would mention such facts as those described above; the student can be expected to know them without instruction. The principles that determine selection of an antecedent, it seems reasonable to assume, belong to 'universal grammar', that is, to the biological endowment that determines the general structure of the language faculty. From another point of view, these principles form part of a deductive, explanatory theory of human language.
Proceeding in this way, we can attempt to construct grammars that answer question
(i) , a theory of universal grammar that in part answers question
(ii) , and performance models that incorporate grammars, answering question
(iii) . Insofar as it succeeds, this quest provides the theory of a particular mental organ.
Others can be studied in the same way, and in principle we should be able to proceed to the study of the interaction of such systems — a central topic as soon as we turn to the study of word meaning, for example.
This seems to be a reasonable paradigm for the study of mind, one that has achieved a certain measure of success and holds much promise for the future.
(Published 1987)
— A. Noam Chomsky
- Bibliography
- Chomsky, N. (1965). Knowledge of Language: Its Nature, Origin and Use.
- Jackendoff, R. (1985). Semantics and Cognition.
- Lightfoot, D. (1982). The Language Lottery.
