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

by Bickerton, Derek

  To the extent that ants function as predators and scavengers (rather than fungus farmers or aphid herdsmen), their food sources are transient, unpredictable, and scattered. In order to feed all the inhabitants of the nest, the only workable method is a fission-fusion strategy.

  And that’s one thing that all our species of interest share. We’ve seen how chimpanzee groups will split and reform, although in the course of a day’s foraging they will seldom be out of visual or aural range of one another. For them, it’s more a case of, well, since there’s not enough fruit on this tree for everyone, let’s you and me move on to the next one.

  Our ancestors had much stronger reasons for a fission-fusion strategy, since even back in australopithecine days food sources were almost certainly more meager, more widely scattered, and more ephemeral than those in the forests of the apes. Even if they gathered at a common site to sleep, there must have been many times when group members were out of sight and earshot of one another, maybe even miles apart.

  Similarly, bees and ants forage either individually (unlikely in our ancestors, for security reasons) or in small groups. Moreover, and this is something that turns out to be quite rare among animal species—it’s certainly not found in any primate other than us—their prey is often much larger than they are. Like Lilliputians with Gulliver, some ant species can overrun and capture small birds, lizards, and the like by sheer weight of numbers. But those numbers have to be recruited—“If you don’t recruit, you get no loot,” as Johnnie Cochrane might have put it.

  Even where the prey is already dead or inanimate, like a fallen fruit, it is subject to decay and to scavenging from other species (although ants seldom have to face competition from larger and fiercer scavengers in the way human ancestors had to), so it has to be exploited as soon as found. Under these conditions, a recruitment strategy is inescapable.

  Really? Couldn’t ants and bees survive without it? It’s possible that they could. But what’s absolutely certain is that, if there were bee or ant colonies that practiced recruitment and bee or ant colonies that didn’t, those that practiced it would prosper spectacularly at the expense of the others, and whatever genes supported recruitment strategies would spread throughout the population. In consequence, it’s no surprise that such strategies have been adopted by probably all bee and ant species except for the relatively few that don’t live socially.

  Among the ant strategies are a couple of things that look surprisingly similar to two of the major building blocks of language: concatenation and predication.

  We saw in chapter 2 that one thing an ACS couldn’t do was to concatenate: that is, put two units together and thereby make something different in meaning from what those units meant when produced separately. But there’s a species of ant, Camponotus socius, that has a complex recruiting behavior. Suppose one of these ants makes a food discovery. It returns to the nest (unless it first meets other ants), meanwhile laying a chemical trail so it can find its own way back. On meeting nestmates, the lead ant draws their attention by a characteristic shaking of its body, signifying a substantial food find, then turns and runs back along the trail, with the others following. As it runs, it continues to pump out the trailblazing chemical.

  It occurred to Bert Hölldobler, a professor of biology at Arizona State and one of the leading researchers into ant behavior, to see what would happen if he removed the lead ant after its shaking display but before it had taken its recruits anywhere near the food. Nothing happened. The ants quickly lost interest and wandered off in different directions.

  Next question: Why did they stop following the original trail? Because the leader wasn’t there to follow, or because a fresh chemical trace was no longer being laid down? In some other ant species, the leader’s physical presence is crucial. That’s in those species that practice a form of recruitment called “tandem running”: one ant grabs hold of another and literally drags it along the trail toward the food source. So Hölldobler manufactured the trail tracer from a mix of ingredients in the ant’s bladder and its poison glands, removed his leader again, and laid down the trace himself ahead of the ants. Sure enough, they followed it.

  Clearly the followers needed the shaking display followed by a continuous chemical flow from the leader to conclude that dinner lay somewhere down the road. The original light trail laid by the lead ant for its own use wasn’t enough, even after the shaking. This isn’t quite like joining words. The shaking and the chemical trail may be meaningless by themselves, so it’s more like the joining together of in-themselves-meaningless speech sounds that we do to make words. But it’s still concatenation of a kind; a kind that, primitive as it is, is found seldom if at all among the behaviors of other species.

  As for predication, you’ll recall that it’s the basic linguistic act of taking some entity and saying something about it: “Dogs bark,” “Birds fly,” “Water’s wet.” Several species of the genus Leptothorax have developed something that, while far from the kind of predication we find in language, is nearer to it than anything in any other species. Here’s one account:

  When a forager finds a plentiful food source it returns to the nest and regurgitates food to its nest-mates. Then it raises its gaster [the rear section of its stomach] and extrudes its sting bearing a droplet of liquid. This attracts nest-mates to it in the nest. As soon as the first of these nest-mates reaches the caller, the caller runs out of the nest and leads the nest-mate to the food source.

  Note that this particular recruitment strategy involves the act of regurgitation. Many species regurgitate to feed their young, but that can hardly be what’s happening here. Rather, it’s as if the ant is presenting a sample of what’s available—“Hey guys, here’s what you’ll get if you follow me.” As we’ll see in the next chapter, this strategy, explaining what kind of food is available, may be crucial for prehuman recruitment. Without knowing what they were being recruited for, it’s doubtful whether protohumans would have allowed themselves to be recruited at all.

  At the other extreme, bees don’t need to sample the product; they already know what they’ll be getting—pollen or nectar. Ants come somewhere between bees and humans in this respect: they consume a wide variety of foods, and they may be influenced by the type of food regurgitated to go out in greater or lesser force. I don’t know if the experiments have been done yet, but they would be easy enough, using different foods, plus a control where the ant was somehow prevented from regurgitating before it raised its gaster. (If it refused to raise its gaster unless it had first regurgitated, that would be significant too.)

  It may seem overly fanciful to regard the whole sequence—regurgitation plus chemical signaling plus tandem running—as the ant equivalent of “Come, such-and-such food is thataway.” Moreover, the sequence, like the calls and other units of ACSs, is primarily designed to get creatures to do things (manipulation), and does so, just like ACS calls do, by means of a stereotyped set of behaviors that cannot be varied or altered in any meaningful ways.

  But it must still be borne in mind that the recruitment sequences I have described are more complex than any communicative behavior in any other species (barring our own) and involve transfers of information more detailed and specific than any other ACS can perform. Moreover, the information transferred is not information about the here and now—as it is in the case of predator warning calls—but refers to things outside the sensory range of message recipients, just as (most of the time) language does.

  RAVENS FLY IN, BUT, ALAS, DON’T TELL US MUCH

  The question that arises at this point is, are ants and bees the only species that need to practice recruitment?

  In the narrow sense in which we’ve defined it—which involves not just telling others you’ve found food, but bridging a time-and-space gap those others couldn’t have bridged for themselves—recruitment turns out to be surprisingly rare in nature. It’s limited by the fact that candidates must be

  • Social: solitary species don’t have to tell anyone a
nything.

  • Foragers over an extensive range: most species travel relatively short distances to feed.

  • Fission-fusion foragers: many species with extensive ranges still keep their usual social groups intact during foraging; consequently, whatever one knows, all know.

  • Foragers for bulk items: the desired food source(s) must be too large or too well defended (or both) for individuals or small groups to handle.

  The only species I’ve so far found that meet these criteria, apart from ants and bees, are ravens. (Thanks, Tecumseh Fitch, for turning me on to them.)

  You can read fuller details about ravens and how and why they recruit in a fine book, Ravens in Winter, by Bernd Heinrich of the University of Vermont. Here is a very brief summary.

  During winter, a major source of food for ravens comes from the carcasses of animals, many of whom have died of starvation. The locations of such carcasses are, of course, quite unpredictable (as are those of dead caterpillars, dead elephants, and the like). Mature ravens pairbond, and if such a pair locates a carcass they will settle on it and stick around until it’s consumed, driving off any other raven that tries to share it. (Note the competition factor, present in the prehuman megafaunascavenging situation, although not significant for ants or bees.)

  Immature ravens forage solo but gather in trees to roost at night (a typical fission-fusion strategy). If in its foraging an immature raven finds a carcass already preempted by a bonded pair, it doesn’t stand a chance. And, at any given time, all the available carcasses in the area may be guarded by vigilant pairs.

  However, if that lone searcher can recruit helpers, it can drive off the pair and access the carcass. But it can’t do this without some way of telling its roostmates about the bonanza it has found.

  Apparently it has a way. Normally, if no one has found anything, ravens don’t follow one another when they go foraging. After a night of roosting together, they all fly off in different directions. But if one has located a carcass the previous day, several if not all of its fellow roosters will take off and follow it to its destination, fighting the current owners, driving them off, and sharing what’s left of the carcass among themselves.

  How do ravens do this? Nobody knows, yet. And it would be extremely hard to find out. You’d have to climb into the treetops (or better, let long-range infrared cameras do it for you) and then somehow determine, out of all the caws and pecks and wing flaps, which one (or ones) carried the crucial message. But even without knowing the mechanics of it, we can be pretty sure that ravens do have an ACS that has somehow achieved displacement.

  It would surely be helpful to know how a species intermediate in its mental powers between ants and humans does it. And it would be helpful too to find more species that face the same problem as ants, bees, ravens, and human ancestors have had to confront. This is one of those times when that all-too-often-repeated saying “More studies are needed” isn’t just an excuse for not making up our minds. Pending knowledge of other species, we’ll just have to soldier on with the ants and the bees.

  However, some of you may still be reluctant to accept that our ancestors could have started language by behaving like ants. For this reason, in the next section I’m going to act as my own devil’s advocate, lay out all the arguments I can think of against the ant/bee scenario, and then show that all of them can be answered.

  DEVIL’S ADVOCATE

  • The “languages” of ants and bees are evolutionary dead ends; tens of millions of years later they haven’t developed into anything more ambitious, whereas language, something that can hardly have started more than two million years ago, is already a system of immense complexity and seemingly limitless productivity.

  Well, what would you expect when one lot have brains smaller than pinheads while the other lot have brains as big as coconuts? Besides, any communication system will fulfill its owners’ needs and no more than those needs. Ants don’t need to gossip—they don’t even have personal lives to gossip about. They don’t need language for sexual display—most of them can’t even have sex. They don’t need it for Machiavellian strategies to enhance their own power and status—their power and status are irrevocably fixed at birth. So why would they develop their “language” any further?

  • Bee or ant “languages” and human language are apples and oranges—the first are hardwired, the second culturally learned. A closed, hardwired system can’t develop into an open, culturally mediated one.

  No one’s saying it could. The reverse is another matter. The production of the first human utterances can only have been spontaneous, not supported by any dedicated infrastructure. Nowadays, language is produced automatically, and speakers are totally unaware of the means by which it’s produced—just as ants are unaware of the chemical mixes and fixed-action patterns that they use in their recruitment strategies. In the same way, the first attempts by ants to recruit helpers for foraging must have been spontaneous behaviors that only later were refined and perfected and absorbed into the nervous system by a process known as the Baldwin effect. Instinct is just fossilized behavior, regardless of whether we’re talking of ants or humans. Changes in behavior trigger changes in genes at least as often (perhaps far oftener) than genetic changes trigger behavioral changes.

  • Ant and bee “languages” are rigorously restricted to a single semantic area—the gathering of food. If human language began with the same function as ant language, why wouldn’t it have remained as a narrowly restricted mechanism for improving foraging capacities, and never acquired any broader functions?

  When I first encountered this objection, my first reaction was to point out differences in brain size and to claim that the first words were spontaneous inventions; once you had a brain big enough and flexible enough to name just one thing, you could name anything. Then when I reconsidered it, I thought, maybe they’re at least partly right. Maybe for tens or hundreds of thousands of years after the first wordlike signals appeared, the modified ACS remained, just like an ant ACS, mired in the business—a vital one, you have to admit—for which it had been originally developed.

  If this were so, it would resolve two other questions that language evolution studies have to face. One is why, if language started two million years ago, did it take so long to develop to its present degree of complexity? The answer to that would now be, because for an indefinitely long period it wasn’t and perhaps couldn’t have been used for anything but scavenging. How protolanguage might have escaped the scavenging trap and branched out into the world is discussed in chapter 11.

  The other question is why, once some simple protolanguage had gotten started, did culture seemingly stagnate, almost up to the point where our very own species, Homo sapiens sapiens—wisest of the wise—emerged?

  This to me is the more important of these two questions, as well as being one that most studies of human evolution either ignore or fudge. It’s mind-boggling when you think about it. Shortly after the big-mammal-scavenging phase we reached at the end of the last chapter, our ancestors began to produce something called an Acheulean hand ax—a teardrop- or pear-shaped stone object that approaches perfect symmetry. Most paleoanthropologists think that Acheulean hand axes were used as tools, for chopping and/or cutting meat, but some think they were projectiles, others that they were simply the cores left after flake tools had been struck off them, and still others that they were an item of display made to capture the hearts of females. Whether it served any or all of these functions, this Lower Paleolithic Swiss Army knife was produced, virtually unchanged, for at least a million years.

  When I give talks on evolution, I often tell my audience things like, “The new model Ford brought out this year is so good it will probably still be in use a million years from now.” That helps to bring home to them the immensity of the gulf between our ancestors and ourselves. It’s unthinkable that our species would produce the same model car even for a decade, let alone a period five orders of magnitude longer, no matter how good it was. O
ur itch for innovation—even if sometimes the new thing turns out worse than what went before—makes any such possibility ridiculous. Ancestors or not, the hand ax makers must have been a totally different kind of being from us.

  The difference lies, of course, in the mind alone. In our physical being, in our emotions and our drives, I’m sure we’re very close to them. There was never any moment at which you could take a parent and a child and say, “This child was a true human, but the parent wasn’t.” Yet somewhere along the way, our minds changed, and they changed quite quickly, as these things go in evolutionary time.

  But if language, true language as we know it, was what reshaped the human mind, as I’ll argue in chapter 10, there’s a simple and straightforward explanation. Protolanguage did indeed stick, for a long time, at the bee/ant level, or only a little beyond it. What changed it, and how it changed, we’ll look at in greater depth in chapters 11 and 12.