It was about the first danger that the Société de Linguistique de Paris was chiefly worried when, in 1866, they established a moratorium on all speculations about the origins of language. It is more concern about the second danger, however, that has kept modern linguists from inquiring overmuch into the beginnings of the human behavior they study. The Chomskian orthodoxy to which nearly all academic linguists in the United States subscribe holds that certain basic structures common to all languages are innate in human brains and that these innate structures, by limiting the range of possible languages, help children learn the languages they grow up hearing. Chomskians often call these innate structures, collectively, the "grammar organ," as if, just as we are all born with a liver to filter our blood, we are all born with a special part of our brain designed for learning human language. Here the mystification begins. As the doctor who explained to Molière’s Malade Imaginère that a certain sleeping potion worked in virtue of its "soporific power," linguists have discovered that many interesting questions may be suppressed by one tidy abstraction.

Chomsky himself seems aware enough of the difficulties of attributing any kind of physiological reality to the "grammar organ" to be coy about what he thinks it really is and how he thinks it might have come to be. Faced with the criticism that the enormously complex innate structures which his theories require could never have developed through the slow mechanisms of natural selection, given even the most generous estimates of the age of human language, Chomsky ducks the question thus:

What Chomsky seems to be suggesting is that there may be some as yet undiscovered property of neurons that, once it is discovered, will show that any mass of them the size and rough configuration of the human brain would naturally come to have a "grammar organ." In a way, Terrence Deacon might be said to be Chomsky’s intellectual heir, for he sets out to show that the human adaptation for language-learning does indeed grow naturally out of other features of our brains, but he does so in such a different way from the way that Chomsky imagined—using not some new mystery of microbiology, but a combination of the current sciences of neurology, physiology, evolutionary biology, and information science—that Deacon overturns much of the Chomskian view of language.

Before we understand why this matters, we must explore why Chomsky and his followers are so eager to put off any explanation of the nuts and bolts of the relationship of language to the brain. Back in the 1960’s, when Chomsky first unveiled his linguistic system, it had such "robust results" (as his followers like to call them) that it carried all before it by the sheer force of its explanatory power. Previous treatments of syntax had taken the basic form that should be familiar to any student of languages: parts of speech, cases, genders, plus a few rules for putting these together and, of course, more than a few exceptions to these rules. Chomsky simplified matters by explaining the surface tangle of syntactic rules as manifestations of an underlying formal system common to all speakers of a language, which he called the "deep structure" of a language. By positing a relatively limited number of "deep structures" and appropriate rules for converting them into surface expressions according to context, Chomsky was able to account for the way we actually speak much more elegantly than had any before him. This success created a wave of optimism as linguistics departments were founded all over the U.S. in happy expectation that a few decades of work would identify all the deep structures and all the transformational rules of any language one pleased.

It turned out that, like many initially promising ideas, transformational linguistics ("transformational" because of the transformations that must be carried out on the deep structures to turn them into recognizable utterances) was devilishly hard to perfect. Each refinement of the theory became more complex—more levels of structure, more rules, more transformations—until many began to suspect that Chomsky hadn’t nailed it after all, and attempts at "non-transformational" linguistic systems began to appear. For Chomsky to defend his increasingly baroque theory, it was convenient for him not to consider how intricate the physiological adaptations of the naturally-selected "grammar organ" might plausibly be. By stating that current biology was not advanced enough to explain language, he could create as many new entities as he wanted—Ockham be damned.

In holding tenaciously to his view that there were innate mental structures guiding the acquisition of human language, Chomsky had a psychological as well as a linguistic agenda: he and his followers did much to discredit B.F. Skinner, Jean Piaget, and their "behaviorist" followers, who held that human minds were pure intellects at birth, ready to be trained into any conceivable form. In arguing that this was not true in the case of language—that certain conceivable languages could never be learned by a human child, because they were not among the possibilities allowed by the "grammar organ"—Chomsky meant to reaffirm the limits of human culture and personality more generally.

Some of his followers have taken the philosophical consequences of Chomskian linguistics to extremes. Most notably, Jerry Fodor has argued that the Chomskian method of separating syntax from semantics leads inexorably to a refutation of both moral and epistemological relativism. One of Chomsky’s great insights was that speakers of a language could recognize certain nonsense utterances as well-formed sentences: Chomsky’s own favorite example was "Colorless green ideas sleep furiously." It is one thing, of course, to say that sense and syntactic well-formedness can be analytically separated and another to establish that syntactic parsing is carried on in the brain by a separate physiological mechanism having no access to any of the mind’s semantic knowledge. This latter view is Fodor’s, and he is quite upfront about his reasons for being so attached to it:

The link Fodor sees between what he calls the "modularity" of the mind (the theory that certain processes, like syntactic parsing, are carried out by dedicated "modules" having no access to information stored in other parts of the mind) and epistemological certainty is roughly this: if the information we come to have about how things look and sound and are structured syntactically doesn’t come from information already in our mind, and it must come from somewhere, then that somewhere must be the real external world.

Fodor is not naïve: he knows that mental information might somehow be triggered by outside information and yet be a wildly inaccurate representation of it, but he is convinced that the success we humans have had in environments that natural selection could not have forseen indicates that our ideas are indeed accurate representations:

And so, the debate about whether transformational grammar is a psychologically real description of human language comes to be about not just whether it is possible for humans to live otherwise than the way they do (an important enough question) but about whether the way humans live has any justification external to human nature itself. No wonder linguists have been loath to demystify their "grammar organ" by inquiring into its evolutionary origins.

Terrence Deacon, however, is not satisfied to cabin off language from Darwinism, nor is he afraid to confront what some find spiritually troubling realities. He begins with the question, why is human language unique?

Deacon challenges this argument by questioning whether language is as valuable as most assume. He doesn’t dispute that full-blown human language is quite useful. Rather, he questions whether simple languages—the evolutionary missing links Darwinism suggests we should find—are useful.

Deacon argues that the reason there are no simple languages in nature is that animals using simple languages would be unlikely to survive. If any other behavior were lacking in the animal world, a suggestion that it lacks survival value would hardly raise eyebrows, but we are so used to thinking of language as preeminently useful that understanding simple language’s disadvantages requires some mental gymnastics. Deacon’s explanation of this point is technical and contains some confusing terminology, but it is well worth the difficulties. He begins by distinguishing symbolic learning—learning the abstract associations necessary for language—from associative learning—learning to expect some event to follow a certain sensory stimulus. Deacon’s insight is that symbolic learning is time consuming, disrupts associative learning, and yields little initial fruit. A normal, associatively learning antelope connects certain grass movements with being pursued by a lion. When he sees these movements, he takes off running. A symbolic antelope would have to forget these connections and focus instead on the relationship that his mental impressions of grass, movement, lions, and pursuit bear to each other. [5] Long before he arrives at "grass is moving," the symbolic antelope is likely to have been eaten.

We know from hindsight that a diversion of energies away from what is immediately the most useful epistemological strategy will eventually bear enormous fruit, but Nature, proceeding blindly in her small evolutionary steps, never defers survival advantage. As Deacon puts it:

The uniqueness of human language, then, is explained by the crucial failure of evolution as a designer—its lack of foresight. While it would be convenient for people to have three hands, we are unlikely ever to get them, because the initial mutations necessary to move humanity in that direction would produce monsters. Deacon suggests that the capacity to learn simple languages is the mental equivalent of an incipient third hand. The question becomes not "why is there only one species that uses language?" but "why is there any species that does?"

Deacon’s answer to this question is the most speculative part of his book. He argues that the circumstances in which our ancestors were placed made simple language uniquely valuable. He claims (and I have no way of evaluating this claim) that humans are the only animals that organize themselves into large social groups and at the same time into (more or less) pair-bonded couples. Certainly we are all familiar with the problems that the combination of these two forms of social organization can cause—adultery, violence, and abandonment. This troublesome dual social organization was made necessary, Deacon argues, when changing climactic conditions forced our last direct ape ancestors (perhaps no more intelligent than modern chimpanzees) to become hunters and foragers on the African plains, where a nursing mother could not collect sufficient food without the help of a male, and where males needed each other’s help to trap and butcher prey. Ordinarily, fatherly provisioning is impossible in a large group, because males do not know which offspring are theirs. Deacon argues that the first symbolic systems (perhaps systems of hand gestures) were created to ritualize marriage bonds, so that males could be induced to provide for females and their offspring. As Deacon points out, a marriage is a symbolic link making a promise about future behavior—no number of repetitions of any stimulus/event combination can constitute such a promise. While learning the symbolic system of husband and wife would have been difficult, marriage, he argues, was so important to an early hominid’s success in rearing of offspring that the diversion of attention from traditional associative learning was worthwhile.

Deacon next faces the problem of how humanoid apes were able to go, in relatively short time, from having to take great pains to understand a single symbolic relationship to learning complex languages practically effortlessly by the age of three. Deacon refers repeatedly to what he calls "the co-evolution of language and the brain." He argues that as languages emerged and became more complex, they placed a selection pressure on brains to focus more on symbolic, and less on associative, learning. At the same time, as brains became more biased toward symbolic learning, languages became even more complex, exerting even more selection pressure in favor of symbolic learning. With these strong, mutually reinforcing selection pressures egging each other on, linguistic behavior emerged relatively rapidly.

What is unique in Deacon’s account is that he sees languages, not just organisms, as capable of evolution. Like genes, languages are systems of information which survive to a greater or lesser extent depending on how often they are copied and which change by being imperfectly copied. [6] Because they change more rapidly than genes can, languages evolve more rapidly. According to Deacon, the secret to human beings’ incredible knack for learning language is not in the human mind’s complex adaptations for language learning, but in languages’ complex adaptations for being learned by humans. The physical adaptations necessary for language learning, while large in scale, are relatively simple—large brains, descended larynxes, high-speed auditory processing, etc. Natural selection could not give children specific knowledge of symbolic rules such as "languages in which the subject precedes the verb have prepositions rather than postpositions." This is just not the sort of information genes encode.

Instead, Deacon argues, languages themselves evolved to be "user-friendly," i.e. to be organized in the ways that children, given their mental predispositions, would be likely to guess that they would. Languages that failed to conform to children’s innate learning tendencies, if invented by adults, would never be learned by the next generation, and so would die out. Thus, the linguistic universals that convince Chomskians that certain symbolic rules must be innate in human minds are in fact the products of convergent evolution, produced not by a common origin, but by a common adaptive context. It is difficult and somewhat unsettling to think of languages, as Deacon does, not just as things we use to communicate, but as independent information systems that survive by reproducing themselves in our brains. Languages, in Deacon’s view, are something like mental viruses.

While shifting the problem of complexity from slowly evolving genes to quickly evolving information structures elegantly solves the problem of how languages became complex so quickly, it is not a comfortable theory to contemplate. The idea that our communications, our identity, and our consciousness are shaped by a separately evolving entity that originated for its social rather than its epistemic value can be quite frightening. Even more frightening is the idea that our brains have been formed by selection pressure in favor of a learning bias that is sub-optimal in many situations.

Because, according to Deacon, it is impossible for Nature to hard-wire specific instructions for handling the symbolic relationships of language, our brains have evolved to be biased toward symbolic learning quite generally. While useful for language learning, symbolic thinking carries costs. Deacon compares humans to autistic savants, who apprehend the world through limited categories spectacularly useful in some circumstances and utterly hopeless in others. Whereas an autistic savant might see the whole world in terms of numbers of objects, a normal human sees the whole world in terms of symbols, categories, and narrative, and, as we all know, can become quite disoriented when, as so often happens, the world fails to conform itself to the Procrustean bed we have laid out for it. The price of our fantastic evolutionary success, then, is the persistent neurosis and tendency toward despair that plagues everyone to a greater or lesser degree.

Deacon’s book is highly recommended to any reader who is interested in language, ethology, artificial intelligence, or the philosophy of mind, as it makes important contributions to all these fields. While it is difficult to read in places, The Symbolic Species is worth reading in its entirety for its exceptionally clear, detailed, and creative arguments—enjoyable for their virtuosity even if one is not persuaded by them. Readers will also find fascinating inquiries into subjects I have not touched upon here, such as the counterintuitive way that large brains, while not directly correlated to intelligence, favor symbolic learning, or the way that children’s short attention spans facilitate language learning. Darwinian evolution through natural selection was a difficult theory for biologists to accept—how could something as complex, as well-designed, as important as life, they boggled, be the product of chance? Once accepted, however, Darwinism has proved to have a kind of beauty in addition to its explanatory power. Deacon argues that it is time for linguists and philosophers of mind to accept that the systems that study also had serendipitous origins.

Radically original as it is, Deacon’s account of the origins of language is not as different from the Chomskian one in its fruits as it is in its methods. While Deacon’s view would obviate the need for a grammar organ, it would not require either cultural or epistemological relativism. On the contrary, it affirms certain enduring similarities among humans, and one would suppose (though he does not address the issue) that Deacon would think that cultures, like languages, would have to reflect those similarities if they are to survive. The errors Deacon sees as caused by humans’ symbolic learning bias do not, in his view, saturate our perceptions. If they did, there would be little problem. Instead they nag at it, infect it, and keep it in constant tension with the non-symbolic knowledge that our minds receive more directly through the senses.

Curiously, Deacon’s scientific explanation of the origins of language also has notable similarities with more traditional accounts, such as the ancient myths of Genesis and of Prometheus—a story of origins tending back toward its own origins. Like Hesiod, who tells of Zeus’ creating Pandora to punish men for having acquired the arts, Deacon sees the beginning of human life in a dual origin of civilization and sexual jealousy. [7]Like the ancient Hebrews, Deacon sees the fruit of the tree of knowledge as both sweet and dangerous.

Do these similarities confirm Deacon’s theories, suggesting that he has explained what humans intuitively suspected of their own nature all along? Or do they suggest that the Société de Linguistique de Paris was right in 1866? That all so-called scientific inquiries into the origins of language end up as myth and fantasy?

Whether his speculations ultimately explain the origin of language or not, Deacon’s inquiry is valuable. Not only for the thoughts it stimulates, but because in a co-evolutionary world of minds, cultures, languages, and people, whatever Henry James may say, origins matter. James would no doubt argue that, just as the Bible has spiritual value despite its human origins, language and marriage and social living have value in themselves, whether or not they arose to enforce sexual exclusivity. To a certain extent, he is obviously right: language and human society, like the Bible, have changed dramatically through their histories, and they should be evaluated today according to their present worth. But Deacon shows us how change, even huge change, does not efface origins. Because cultural institutions not only evolve according to their historical or biological conditions, but themselves shape the conditions in which they evolve, the co-evolutionary tracks of biology and culture leave recurring patterns in each—patterns which can help us to understand ourselves.