Adam's Tongue: How Humans Made Language, How Language Made Humans

by Bickerton, Derek

  An overarching concern in studies of language evolution is with whether particular components of the faculty of language evolved specifically for human language and therefore . . . are unique to humans. Logically, the human uniqueness claim must be based on data indicating an absence of the trait in non-human animals, and, to be taken seriously, requires a substantial body of relevant comparative data.

  My jaw had dropped on reading this. Why was it “an overarching concern” to tackle human uniqueness? What on earth had human uniqueness to do with how language evolved? I myself didn’t give a damn about human uniqueness. I just happen to be a member of a species that happens to have language, and I want to know how and why we got it. Every species is unique, by definition, because if it weren’t it would be part of some other species.

  This whole business of human uniqueness is a red herring, dragged in by the culture wars. There are some people who want to prove, for their own ideological reasons, that practically everything about humans is unique and others who want to prove, for their ideological reasons, that virtually nothing about humans is unique. And, as always, there are some wishy-washy people who want a “moderate” compromise between these positions. Hauser falls relatively near, though definitely not at, the “virtually nothing” extreme.

  My reaction is, to hell with the whole thing. Science should have no truck with this kind of stuff. Just answer the real questions, and let the uniqueness chips fall where they may. Because once you start worrying about “human uniqueness,” you inevitably start seeing evolution the wrong way. You start seeing it as how like (or unlike) other species are to humans. You start, you can hardly help starting, to see humans as the standard other organisms are measured by. You’re doing exactly what anti-uniqueness people are supposed not to be doing. You’re making your own species the centerpiece of evolution.

  Whereas of course evolution doesn’t have a centerpiece, or even a center. And even if it did, it would look too obviously self-serving to put ourselves there. After all, all we’re doing is trying to find out what happened and how and why it happened in the period between us and the last common ancestor of chimps and humans.

  What the paper was telling language evolutionists to do was quite unrelated to such concerns, and indeed couldn’t have been more effective if it been specially designed to keep them from pursuing such concerns. It was sending them wild-goose-chasing out into the highways and byways, to conduct all sorts of experiments on all sorts of animals, to determine which of the capacities that contributed to language were found in those animals and which were not. The content of the latter category was, hopefully, zero. Although the authors were careful to include bet-hedging language that could be, and was, pulled out later when they were accused of defining FLN too narrowly, it was clear from the tone of the whole paper that they believed, or at worst hoped, that FLN would turn out to be nothing more than recursion, and that recursion, while apparently absent from the behavior of any other species, was really there in some others, but only used for navigation, or social relationships, or . . . something.

  And if some other species had recursion, how come it hadn’t put recursion together with all its other linguistic precursors and made language, long before we did?

  Well . . . perhaps because “recursion in animals represents a modular system designed for a particular function (e.g. navigation) and impenetrable with respect to other systems. During evolution, [this system] may have become penetrable and domain-general . . . This change from domain-specific to domain-general may have been guided by particular selective pressures, unique to our evolutionary past, or as a consequence (by-product) of other kinds of neural re-organization.”

  Okay, class, explain in simple language how an impenetrable modular system becomes penetrable.

  “Either way, these are testable hypotheses . . .”

  Just how would you test them?

  Six years on, nobody’s yet found recursion in another species. Of course, for all I know, somebody’s finding it right this minute. But I’m not holding my breath.

  CHOMSKY SPELLS IT OUT

  The Hauser-Chomsky-Fitch paper didn’t go uncriticized. For years a debate raged, or rather simmered, in the pages of the journal Cognition between the authors of the Science paper and Steven Pinker and Ray Jackendoff, two authors who had espoused gradualist models of language evolution. Surprisingly, nobody raised any of the issues I’ve discussed in this chapter. The debate focused solely on what was the proper content of FLB and FLN—shouldn’t X, Y, or Z be in FLN, rather than FLB? In other words, it was purely definitional, ignoring, just as the original paper had, all the deeper issues, all the issues most central to language evolution, and worrying about what was unique to humans and what wasn’t, as if this were an issue of substance. Pinker and Jackendoff fought Hauser et al. on the turf the latter had marked out.

  Also lost in the shuffle was any debate on what if anything happened between the dawn of language and its full flowering. You’d think Jackendoff, with the nine stages he’d hypothesized for language evolution, or Pinker, with the one-rule-after-another scenario that he and Paul Bloom had laid out in a 1990 paper, might have complained about this lack. For that matter, Hauser, who had written favorably of the protolanguage concept, might have been expected to bring up the issue of exactly how language, once started, had developed. This shows just how far “Science’s Compass” had succeeded in directing people away from the real issues in language evolution.

  Then it occurred to me that perhaps the omission from that paper of anything between no language and full language had been deliberate, not accidental. Perhaps anything intermediate between no language and full language had to be one of the things that, to keep the Chomsky-Hauser compromise afloat, just had to go.

  Was that possible? Could they actually be claiming that language had sprung, fully formed and complete, straight from Jove’s brow?

  The Science paper never explicitly claimed that. And surely it couldn’t be claiming that, since any such claim would be about as unbiological and unscientific as you could get. Yet if all other bits of language were there in other species, and only one thing, recursion, was then required to give birth to full language, what other interpretation could you put on it?

  Then, at the height of my mystification, like a deus ex machina, Chomsky himself appeared before me to sort it all out.

  Chomsky seldom goes to linguistic meetings, seldom goes anywhere where he can’t devote equal speaking time to language and politics. However, since he was trying to make up for years of lost time and prevarication, he appeared at a meeting on language evolution held at the Stony Brook campus of the State University of New York in the fall of 2005.

  The event was billed as a symposium, one of those Greek thingies where you get to hang with Plato and the boys, drink lots of wine, and explore deep philosophical issues together. Maybe those of us who didn’t know him were hoping Chomsky would stick around long enough to quaff a glass, crack a joke, spin a new theory or two. Of course that was not to be. Chomsky is a busy man; he told me himself he spends six hours a day (or night) just answering correspondence, a hundred or more letters at a go. He came one evening, had dinner, gave a talk, and was gone by noon the next day.

  However, he did leave us with an unusually clear statement of how he thought language had evolved, to which we now turn.

  In some small group from which we all descend, a rewiring of the brain took place yielding the operation of unbounded Merge, applying to concepts of the kind I mentioned . . . The individual so rewired had many advantages: capacities for complex thought, planning, interpretation and so on. The capacity is transmitted to offspring, coming to predominate. At that stage, there would be an advantage to externalization, so the capacity might come to be linked as a secondary process to the sensorimotor system for externalization and interaction, including communication . . . It is not easy to imagine an account of human evolution that does not assume at least this much, in one or another for
m [my italics].

  This needs a little parsing.

  For instance, “externalization” means “to start talking.” All these capacities for complex thought, planning, interpretation, and so on were up and running, and only then did it occur to someone to think, “Hey! Why not use them to chat with?”

  Then, when Chomsky mentions “concepts of the kind I mentioned,” he is referring to an earlier paragraph in his talk, where he stated:

  There do, however, seem to be some critical differences between human conceptual systems and symbolic systems of other animals. Even the simplest words and concepts of human language and thought lack the relation to mind-independent entities that has been reported for animal communication . . . The symbols of human language and thought are quite different . . .

  What Chomsky points out here is indeed a profound difference between humans and other animals, one we’ve looked at several times in previous chapters. In animal communication, as we saw, the kind of “functional reference” you get (if you get any kind of reference at all) is to “mind-independent entities”—directly to actual things in the real world, rather than to our concepts of those things, the way in which words refer. When we use a term like “leopard,” we’re referring (unless we qualify it in some way) not to a particular leopard, least of all to one that’s there before our very eyes, but to leopards in general; the statement “Leopards live in Africa” is not invalidated if a couple of them happen to be living in our local zoo. But when an animal gives a warning call for “leopard,” this refers to one particular leopard, the one that happens to be right here right now. So in this paragraph, Chomsky is focusing on properties of human words, and in the previous paragraph he is saying that human words alone can be merged with others of their kind to form sentences.

  Fine so far; just what I’ve been saying here, and I’ve even explained why it had to be so, why animal signals with their “mind-independent reference” just couldn’t ever, in the nature of things, combine.

  But Chomsky is also stating that human words (and the typically human concepts that underlie those words) are the only kind of unit that recursion can apply to. So when recursion emerged (through a “rewiring” of the brain) these concepts must already have been present. This commits Chomsky to the following sequence of events:

  TIME 1: Animals have concepts that won’t merge.

  TIME 2: Typically human concepts, which will merge, appear.

  TIME 3: The brain gets rewired.

  TIME 4: Merge appears and starts merging typically human concepts.

  TIME 5: Capacities for complex thought, planning, etc. develop.

  TIME 6: People start talking.

  So what Chomsky has done is simply add, to the problem of how language evolved, a second and at least equally horrendous problem—how typically human concepts evolved.

  “At least equally” is probably an underestimate, for the following reason.

  We talk about natural selection selecting from genetic variation. But that’s shorthand for more complex processes; it seriously oversimplifies what’s happening in evolution. Genetic variation is invisible to natural selection. Natural selection can’t pick among genes; it can only pick among physical organisms, selecting those that survive longest and breed most. Did the genes cause some to survive longer and breed more? Well, again, not directly, because what determines long life and successful breeding can only be how physical organisms behave. Sure, a species’ range of possible behaviors is determined by genes; if those genes can’t make wings, flying is out of the question. However, genes don’t determine which of those behaviors animals will produce, and behaviors are the proximate cause of whether an animal survives and breeds or not. Real-world events, the interactions of a physical organism with its environment, are what drive natural selection, not things that happen inside the organism.

  But for Chomsky, language had to evolve inside the organism before it could appear outside the organism.

  Well, that’s okay, you say. Animals that could think and plan for the future, even if they couldn’t talk, would surely outperform and outbreed nonthinkers and nonplanners.

  True enough. But how did they get able to think and plan for the future if they could only do this by first acquiring human-type concepts and then merging them to produce consciously directed, constructive trains of thought? To say animals got gradually smarter through some kind of feedback process simply won’t cut it. It’s not a question of getting smarter. It’s making a jump from “functional reference” to full symbolic reference; from concepts that can’t combine to concepts that can combine. No way can you do that gradually. There are no concepts—in fact there is nothing anywhere—that can a quarter combine, or 65 percent combine. Things can either combine or they can’t.

  As for “the rewiring of the brain,” brains don’t rewire themselves for no reason, or because they just feel like a spot of rewiring. Brains rewire themselves, to the extent they do, because things are happening in the outside world, things that give individuals with rewired brains an advantage over individuals without them.

  So the question becomes, what would have selected for kinds of concepts and kinds of brains different from those that had served all life-forms well since life began?

  Answer: nothing.

  A concept is something in the mind. Once it exists, it can affect behavior. Before it existed, it couldn’t. All that natural selection can see is behavior. So concepts could only have been visible to natural selection once they existed, once they’d begun to affect behavior. But they couldn’t exist until they’d evolved, and they could only evolve if they were selected for.

  So human-type concepts couldn’t have evolved by themselves. They could only have evolved if some other thing had been selected for, something that was visible to natural selection—in other words, some overt behavior that gave an adaptive advantage to those that had it.

  But for Chomsky there’s nothing. All parts of language, everything necessary for its full flowering, just grew, somehow, in the mind, all by themselves.

  When I meet someone whose ideas are different from mine, I always try to think through them, determine why those ideas were chosen, rather than others. To find out where Chomsky’s coming from, we need to look at two aspects of his thinking about language that, between them, virtually force him to accept those ideas. Both aspects, when applied to language today, have proven both useful and (probably) right. It’s when they’re applied to the evolution of language that they create problems.

  HOW CAN WE KNOW WHAT WE THINK UNTIL

  WE SEE WHAT WE SAY?

  One of those aspects concerns the balance in language between thought and communication.

  Chomsky has always emphasized that language is at least as much a system for structuring and thinking about the world as it is a vehicle for communication. I couldn’t agree more. I wrote a whole book saying just that, and exploring the world-creating capacities of language, back in 1990.

  But I overestimated somewhat, and Chomsky overestimates a great deal, how much thinking animals could do without language.

  The fact that language is by now the main engine of thought doesn’t have any implications for its status when it began. That’s the fallacy of first use, the idea that whatever a thing started doing will be what it does mostly nowadays—and vice versa, naturally. It was the fallacy of first use that led Robin Dunbar to propose gossip as the engine of language evolution, just because gossip is what (spoken) language is most used for today. If the fallacy of first use were true of computers, they would have been used first for e-mail and Web browsing, and some of us are old enough to remember how big a fallacy that would be.

  Certainly, language is now the means by which we structure the world of thought, but it would never have gotten off the ground, never developed into what it is today, and certainly never have raised thought to a new power if it hadn’t first entered the real world in the tangible form of communication. As I showed above, only external events
can shape internal events, because only external events are visible to natural selection.

  Chomsky and his followers, who have never liked natural selection, will say that there are other factors in evolution. They will claim that much of evolution results from mysterious, yet-undiscovered laws of form, laws of development, or other similar forces. Even Fibonacci numbers—the strange sequence of figures that seems to control the branching of plants—have been invoked. But this is sheer hand-waving. We know that natural selection works and we know how it works to cause new evolutionary developments. But while things like laws of form may exercise constraints on such developments, no one I know of has suggested that they can cause them, and nobody has the faintest idea, if those laws do work, how they would work to produce the effects of language.

  WITH MY METHOD, YOU CAN LEARN LANGUAGE INSTANTANEOUSLY

  The other aspect of Chomsky’s thinking about language that has had an unfortunate influence on his ideas about evolution is an idealization: the idealization of instantaneity in the acquisition of language.

  Some people who studied how children acquired language were very upset by this. After all, children take at least a couple of years before they become fluent speakers of a language, and since we go on learning new words all our lives, you could even claim the process, far from being instantaneous, is one that never stops.

  But think this through more carefully. What do children start with? Meaningless babbling. What do they all, all normal ones anyway, arrive at? Full control of a human language. Is there any difference between them at that stage? Well, some may talk more than others, some nicer, but there’s no nonsubjective way to determine any difference in how they talk. Whatever happened along the way, whatever funny things they may have said or had said to them, made no difference to the end result. So the process might just as well have been instantaneous. And if they come equipped, in some yet-to-be-defined sense, with some kind of universal grammar—and there’s good reason to believe they do—then if it weren’t for the severe limitations in speech production, focused attention, word-learning, general experience, and the like that all infants labor under, they jolly well might acquire language instantaneously.