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

by Bickerton, Derek

  Perhaps because of this apparently looser structure, and what seems like the basic simplicity of thoughts like “Roses are red,” many people assume that we get them for free, so to speak, just by having a brain. And this feeds into another widespread belief, held even by some who accept language as the prime motor of human intelligence, that thoughts are somehow logically prior to sentences, that language arose to express thoughts, that you first have to think something before you can dress it up in words and send it out into the world. Remember, there are not only no images and no words in the brain, there are no thoughts there either—only a continuous cascade of neural activity, of pulse rates spiking, of impulses going every which way.

  For me, the belief that thinking preceded speaking in evolutionary time and precedes it operationally in our daily lives is one of those initially plausible beliefs that when held up to the light and carefully examined can be seen to lack any real foundation in either fact or theory. Indeed, I am going to argue that, until we could talk, we could not even think, “Roses are red.” But since you may well be harder than me to convince, let’s look at what’s actually going on in both human and non-human minds and brains.

  NONHUMAN VERSUS HUMAN MINDS

  Start by thinking of the simplest answer to the question, why are other animals trapped in the prison of the here and now?

  The simplest answer is, that’s all they’ve got.

  They can’t communicate about things beyond the here and now because they can’t direct their thoughts outside the here and now. And the reason they can’t do this is also the reason they can only refer to specific, immediate events. They don’t have abstract concepts, and we do.

  Be warned, this is heretical stuff. As a trio of fellow heretics (Derek Penn and Daniel Povinelli of the University of Louisiana, and Keith Holyoak of the University of California, Los Angeles) pointed out in a recent article, “Ever since Darwin, the dominant tendency in comparative cognitive psychology has been to emphasize the continuity between human and nonhuman minds and to downplay the differences as ‘one of degree and not of kind’ (Darwin 1871).”

  For example, in asking his readers to imagine themselves at the dawn of language, Jim Hurford, one of the best writers on language evolution, suggests with breezy confidence that they should think what it would be like to be without language “but otherwise cognitively pretty much as you are.”

  But surely nobody would make such claims without massive evidence?

  Right; according to Irene Pepperberg, Alex the parrot’s mentor and advocate of a psittacine-inclusive approach to language evolution, “for over 35 years, researchers have been demonstrating through tests both in the field and in the laboratory that the capacities of nonhuman animals to solve complex problems form a continuum with those of humans.” Anyone who goes up against this consensus risks being branded as a last-ditch defender of human supremacy, the special, divine creation of “just us.”

  But is this evidence telling us the plain truth or the truth we’d like to hear—whatever reinforces the current brand of the Darwinian wisdom? Could it perhaps be interpreted in quite a different way?

  Before we review the evidence, let me set things up by looking a little more closely at the Irene Pepperberg quotation above: “the capacities of nonhuman animals to solve complex problems form a continuum with those of humans” (my italics). But solve them where? In the world or in their heads? That’s the crucial question.

  What I’m going to claim is that the capacities of nonhuman animals do indeed form a continuum with those of humans when it comes to solving physical problems in the real world. Such problems are highly likely to be solved if:

  • The animal is highly motivated to solve the problem (that is, if the problem stands in the way of fulfilling some immediate need for food, sex, escape, or anything else that contributes directly to the animal’s fitness).

  • Most if not all the things necessary for solving the problem are in plain view.

  • Anything not in plain view is stored in the animal’s episodic memory (the part of memory that, roughly speaking, stores its owner’s experiences in narrative form) and can be triggered by attempting to solve the problem.

  Problems of this nature may even be solved faster than a small child or an adult with a low IQ could solve them. (If you’ve seen the video in which a New Caledonian crow, given a straight piece of wire to get food out of a glass tube, twists the wire into a hook after only a few seconds of fruitless poking, you’ll know what I mean, and never again use the demeaning expression “birdbrain.”)

  The question is not whether animals can solve problems—they obviously can—but whether they have concepts that they can summon at will and manipulate so as to imagine, and thus subsequently produce, novel behaviors.

  BUT MIGHTN’T SOME ANIMALS HAVE CONCEPTS TOO?

  It’s more than likely that a majority of people from the behavioral sciences will at this point if not before cry out, “But of course animals have concepts like ours!” To deny this means claiming there’s a discontinuity between humans and nonhumans. But there’s already one possible discontinuity, that between language and all other forms of communication. One such discontinuity is bad enough—two, and the whole edifice of Darwinian gradualness begins to shake. Or so it (wrongly) seems.

  Think this through a little further. Unless we are ready to accept that bacteria have concepts (and where on earth would they store them?) there has to have been, somewhere in the course of evolution, this exact same discontinuity between animals that had concepts like ours and animals that didn’t. If we don’t place it between humans and nonhumans, where are we going to put it, and why? Given all the things we can do that other animals can’t, what likelier place to find that discontinuity than between us and them?

  In the final analysis, the issue of whether or not animals have concepts like ours is an empirical one. They either do or they don’t. It should be possible, within the next few years, to determine through advanced brain imaging techniques if the difference between us and the rest of nature really is what I’ve suggested here. In the meantime, we should use any other means available to determine whether any other animals have concepts that they can conjure up at will—that they can start thinking about even if there’s nothing currently happening that would directly or indirectly activate those concepts, just as we do. If it turned out that some did have such concepts, we’d have to start looking all over again for why we’re so much more creative than other species.

  In the meantime, let’s look at some of the best evidence for animal concepts.

  Back in the seventies, Richard Herrnstein (later to gain notoriety by coauthoring a book he never lived to see published, The Bell Curve) carried out a series of experiments with pigeons that psychologists are to this day still trying to explain. He himself was dumbfounded: “How can animals with such powers of classification still seem stupid in some ways?” he asked. Surely, only if their minds work very differently from the way ours work.

  By training pigeons on a small sample set, he first got them to reliably distinguish between pictures that had trees in them and pictures that didn’t. (They did this by pecking when there was a tree in the picture and not pecking when there wasn’t.) Easy, you say; pigeons spend half their lives in trees. So Herrnstein moved them on to pictures of people, and they did equally well. Okay, pigeons have usually seen people; certainly all pigeons in a psych lab have seen people. But fish?

  When Herrnstein trained his pigeons on a small set of fish pictures and then turned them loose on a much wider collection they’d never seen before, they did pretty much as well as they’d done with trees and people. You can be absolutely certain that none of the pigeons had ever seen a fish or even knew what a fish was. But that didn’t stop them from recognizing almost every fish they saw, and knowing when something wasn’t a fish.

  How did they do it? Did they go by general outlines or specific features or what? Whatever they used, what did it imply about how t
heir minds worked? Thank God we don’t have to go there; it’s a quagmire. All we have to ask is, does what they did imply they acquired any kind of general concept of fish?

  I don’t think it does. I think they noted and stored a set of distinct features—doesn’t matter what; for our purposes, anything common to fish—and that when a large enough subset of these features was triggered by the picture, they pecked. I don’t suppose for a moment that they ever thought about fish once the experiment was over—wondered what they were, imagined what eating one might be like. True, in some cases the training effects lasted a year or more, but that just shows pigeons have good memories; it doesn’t show they can form any central notion of fishiness.

  So let’s go out of the lab into the wild. Let’s check out scrub jays.

  Scrub jays are a bird species common in the West. They live by collecting seeds that they cache all over the landscape so that when winter comes they’ll have a steady food supply. In the course of a season they may make hundreds if not thousands of these caches. And they remember them all, fly back to them unerringly. More than that, they remember which caches have more perishable seeds and which have less perishable ones, and they don’t bother going to the first kind in the later months of the winter.

  I bet if you or I with our big nonbird brains were to hide seeds in different places all summer and then find them again all winter, we’d do a lot worse than the scrub jays—probably fail to survive, if our lives depended on it. True, scrub jays don’t have much else to do with their time. But that shouldn’t take away from the fact that, like many other species, they have hyperdeveloped capacities that in us are weak or lacking altogether. It’s no reflection on them that our best trick didn’t just keep us alive, but lucked us into a different universe.

  But because they have one mental capacity that goes way beyond us doesn’t mean that their other mental capacities are equally developed, or even that they are developed at all. Too often people seem to think (or maybe feel emotionally, or at least behave as if) there’s some universal fixed scale of intelligence on which all species can be placed somewhere, some higher than others, us higher than all. Evolution does not work that way. A repeated refrain in this book has been this: a species does what it has to do. If a bird enters the seed-caching niche it will probably get, sooner or later, the kind of skills a scrub jay has. Natural selection will see that it does. The ones who do what they have to do better than others will survive longer, breed more, have offspring that do even better than they did. The niche creates the intelligence—not some generalized cleverness, but whatever specialized intelligence the niche needs.

  Well, birds are birds, you say; why would you expect to find humanlike concepts there? You should be looking at our nearer relatives.

  I would, if in their normal lives they did anything at all to suggest they had concepts. It’s surely significant, if seldom noted, that the evidence cited for animal concepts is seldom drawn from the great apes—at least, not from great ape life in the wild. For that kind of evidence, we have to go back to the monkeys.

  Klaus Zuberbühler and his colleagues from the University of St. Andrews in Scotland did a series of experiments that involved playing recordings of alarm calls for leopards and eagles made by Diana monkeys (named after the goddess, not the princess) and the sounds made by the predators themselves. I can’t improve on Jim Hurford’s summary, so I’ll just quote his words:

  “On hearing first an eagle alarm call, then (after five minutes) the shriek of an eagle, female monkeys showed less signs of alarm (giving fewer repeat calls) than after hearing, for example, an eagle alarm call followed by the growl of a leopard.”

  Hurford takes this to mean that Diana monkeys have concepts like ours for eagles and for leopards. He reasons thus: for the monkeys to behave differently depending on whether a predator sound is expected (eagle) or unexpected (leopard) means that, for at least five minutes, the animals must have kept the concept of “eagle” in their minds, and hence they were shocked when what they heard didn’t fit the threat they’d been warned of.

  Of course, that’s one possible explanation. But there are others at least equally likely. One, it’s only surmise that the eagle and leopard calls mean “eagle” and “leopard” to a monkey. They could just as easily mean “threat from above” and “threat from the ground.” The warning monkey may be reacting not to concepts of eagle or leopard but to sounds that it recognizes as threats coming from the air or the ground, respectively.

  Two, receiving an eagle alarm call puts animals on high alert and primes them to be ready with the appropriate strategy—if you don’t hide immediately, be prepared to dive into some bushes the moment you see or hear anything above you. They will remain ready to execute this strategy for some minutes after the warning, until enough time passes without incident that they feel they’re no longer in danger. A predator sound that confirms the warning will simply keep them alert and/or make them go for the bushes. It’s the strategy that persists through time, not the concept of a hovering eagle.

  But suppose that instead of an eagle shriek they hear something they weren’t expecting: the sound of a terrestrial predator. This throws them completely off balance, because the two escape strategies can both be fatal if used with the wrong predator. In the bushes, where you hide from eagles, leopards can grab you. Up a tree, where you’ll be safer from a leopard, an eagle can easily spot you and pick you off. Small wonder Zuberbühler’s monkeys showed less alarm when a signal was confirmed than when it was contradicted. They were alarmed in the second case because they simply didn’t know which strategy to use.

  Perhaps the best case for animal concepts like ours comes from the behavior of “language”-trained apes. You’ll recall how, when they were first taught manual signs, it took them a long time to “get it”—several hundreds or even thousands of trials, spread over weeks or months, before they realized what the signs represented.

  There are two possible explanations here. If people who think apes think like us are right, apes already had concepts of the right kind. They just didn’t have labels for those concepts. Then kind humans came along and supplied labels. It took a while, but sooner or later came that “aha!” moment and the apes slapped the labels they were given on the concepts they already had—end of story.

  The alternative goes like this. Apes didn’t have concepts. Just like any other nonhuman animal they had categories into which they could sort things so they’d know how to respond to them. Those categories didn’t jell into accessible concepts because they only functioned when the apes saw or heard or smelled or touched or tasted features on which the categories were based. That happened occasionally and unpredictably. The network of neurons that got activated when it did happen only linked up at those moments and quickly faded into oblivion when the features ceased to be perceived. Nothing remained that would tie all those features together.

  Then the apes learned signs for their categories. The signs tied all the category features together and gave them a permanent home. They did so because the presentation of the category features—the features that distinguish, say, bananas from M&Ms—was no longer occasional or unpredictable. The researchers kept shoving bananas and M&Ms in the apes’ faces. The neurons in the circuits activated by these presentations, together with those representing the objects’ names, kept firing and firing. Neurons that fire together, wire together. The circuit was reinforced and anchored by the sign that had just been learned.

  If that’s all it takes to learn and use concepts, how come the trained apes didn’t immediately start thinking like us?

  To a very limited extent, they did. Thirty years ago, David Premack showed that “language”-trained apes could pass cognitive tests that untrained apes failed. But, if I’m right, it took us the best part of two million years to get from where we and the apes started—from no language at all—to where we are today. It’s one thing to have representations of words/concepts stored in different parts of the brain. It�
��s another for those representations to be linked by the afferent and efferent fibers that enable signals to pass in both directions from one to another, something that is vital if enough units are to be linked to make a coherent train of thought. Note that no ape has ever joined more than three signs in a communicative message. It’s likely that they’ve never been able to merge more than three concepts into one coherent thought.

  And there are dimensions beyond the cognitive that may be operative here. We won our language. Apes had theirs handed to them on a plate. We needed language to develop our niche. Apes didn’t need it, never wanted it, need it now only to get rewarded and to keep their human keepers happy. And they’ve been using it for less than a human lifetime. All in all, I think they’ve done pretty well, don’t you?

  We should be able to respect them, without trying to turn them into blurred carbon copies of ourselves.

  CONCEPTS AND THE DIVIDE

  So let’s assume that I’m right and that the presence or absence of human-type concepts is what divides humans from nonhumans. It’s not really a question of problem-solving per se, but rather of the kinds of problems that are solved, the ways in which they are solved, and what kind of mental operations go into solving them.

  For some reason, whenever we think about the evolution of intelligence, we tend to see it in terms of solving problems of ever-increasing complexity. We should be looking at the rarity with which animals solve problems by developing new behaviors, and the frequency—constancy would be a better word—with which we do it.