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

by Bickerton, Derek

  We do know, and it’s one of the better established facts in archaeology, that it was in the period between 2 million and 1.6 million years ago that three things coincided:

  the appearance of the Acheulean hand ax

  the appearance of tooth marks superimposed on cut marks

  the switch from catchment scavenging (the intensive exploitation of relatively small territories) to territory scavenging (ranging over much wider areas)

  The hand ax was the all-purpose butchering tool, used to hack through bone and sinew once the hide had been cut open by flakes, and also as a projectile to drive off the scavenging competition. The cut marks superimposed by tooth marks are the unarguable signature of high-end scavenging—irrefutable evidence that human ancestors had gotten thar fust with the most. The switch to territorial scavenging could have been made possible only by a switch in resource exploitation—meat taking the place of bone marrow. Although it’s hard, and probably always will be, to find smoking guns, direct evidence that power scavenging became our ancestors’ main activity, the balance of what evidence there is points in that direction.

  For a species that had become dependent on scavenging large carcasses, it would become more and more important to accurately read all the signs that the megafauna left, to determine species identity and relative age, numbers in the group, things that indicated an animal might be injured or sick. Disputes would inevitably arise about how the signs should be interpreted. How old were they? How many animals would have made them? Should we follow group A, small but with one animal in it that might be sickening, or group B, much larger but with all members in apparent health?

  Such disputes would intensify once a potential victim had been identified.

  The subgroups you’d have to recruit would have their own agendas. They too might have a sick animal in view. The task of persuading a subgroup to drop what it was doing and come join you would get harder, not easier. How much nearer death is your animal than their animal?

  Then there was the competition. Were there other potential scavengers around? If so, how many? And perhaps even more crucial, what species did they belong to? Here again, iconic signals go a long way, but precisely since they are iconic, not indexical—not, that is, pointing directly at a particular exemplar of something—they’re easier to convert into symbols as they get to be used in more and more contexts. If you can name major predators and indicate the signs they leave, you can teach your children lessons that may in the future save their lives.

  In this way, developing the high-end scavenging niche would have both created new words and deployed old words in new contexts, further weakening the uncoupling of words from situations, from current occurrence—even from fitness.

  COMPLETING THE TRIPLE UNCOUPLING

  This couldn’t have been a rapid process. The coupling with fitness would have been perhaps the longest to break. Imparting information purely for information’s sake still lay ahead.

  The French scholar Jean-Louis Dessalles has proposed one factor that would have encouraged such a process. Primates are highly competitive, constantly seeking to be one up. Modern humans can achieve one-upmanship by being the first to pass on some piece of new and significant information. Surely, Dessalles argues, this behavior should go back as far as there were ways to give information.

  As a starter for language, this proposal seems at first to fall foul of the condition that fails so many promising explanations of how language began. There’s little doubt that the behavior would have helped to expand language at later stages. However, until language had been up and running quite a while, there simply wouldn’t have been enough words for the proposal to work.

  But it does provide a likely source for the creation of new words once protolanguage had gotten started. Relatively rare yet recurring events that severely impacted a group—flash floods, hurricanes, grass fires—were things whose effects could be mitigated by advance warnings. Moreover, ability to give advance warnings would confer prestige on the giver. Consider flash floods, for example; any sound or gesture that reminded the group that a sudden, heavy fall of rain likely preceded an inundation would have been stored and repeated.

  It’s still about fitness, you say. But suppose that while a group is working on a megacarcass, there’s a sharp shower, and an immature, rather nervous hominid gives the flood call. Its fellows begin to head for higher ground. Quickly, wiser elders couple the flood word with a dismissive gesture, one perhaps that already means “Don’t do that!” So negation is born. “There isn’t going to be any flood, this time” is still a long way ahead, but that’s what’s meant. And a multiplication of words, plus the power to combine them, is all that is lacking now.

  A just-so story, of course. But negation has to come in somewhere. It’s one of the first things small children learn—develop would be a better word, because they don’t negate like mother does, with “don’t”s; they just stick the general negative word “no” in front of whatever it is they don’t want. In the scenario I described, our older and wiser hominids would be using the flood word not to refer to an actual or imminent flood but to assert the current nonexistence and future unlikelihood of a flood. And nonexistence is something no ACS can handle. It can’t handle it because ACSs refer, if they refer at all, to things that actually exist, in the here and now—“mind-independent entities” that have a physical life in the real world. But “no flood” can’t refer to any flood there ever actually was—only to the abstract concept of floods.

  These are just some of the steps that might have been taken to divorce words from any real-world things they referred to and make them the outward forms of true concepts, applicable to any or all members of a given class, whether hypothetical or real. Until the triple uncoupling was completed, all you would have would be just a more sophisticated version of bee “language” or ant “language”—a system dedicated to one major purpose (providing food for the group) but useless outside of that function.

  I didn’t always think this way. I supposed that if you gave the equivalent of bee language to a big-brained animal, that animal would rapidly expand said language to help out all the other functions big-brained animals perform. Somehow that seems very natural to members of our species, a species whose members are saturated in language from a time before our memories begin. But what basis do we have for such a notion?

  Language was an unforeseeable development whose properties ran counter to any behaviors that had happened before. It was an evolutionary anomaly at least as great as the emergence, in a world filled exclusively with single-celled creatures, of the very first multicellular organisms. Greater, in fact; multicellular organisms did no more than multiply what had been there before. Language, on the other hand, was a pure novelty.

  I now suspect that what recruitment brought in its train—a system purpose built, devoted solely to the exigencies of power scavenging—may have persisted for several hundred thousand years or more, while hardly yet deserving the name of protolanguage. A kind of hybrid stage, halfway between ACS and protolanguage—little more developed than the “languages” of bees or ants.

  PIDGINS FLY TO THE RESCUE FOR REAL, THIS TIME

  Before this occurred to me, when I had to explain the long stagnation while still assuming that a relatively rich protolanguage developed relatively early, I naturally had recourse to Dan Dennett’s indispensable invention, figment (no writer on human evolution should ever be without it, or ever is without it).

  My particular piece of figment went like this: without syntax, you can’t put thoughts together any more than you can put words together. In order to create new cultural and technological stuff, you first have to put thoughts together in an orderly and disciplined fashion. In order to do this, you need syntax. Alas, our poor ancestors, without syntax, could not put their thoughts together. Therefore, they were unable to think the kind of thoughts that they needed to think in order to achieve real cultural and technological innovations.

  Thence sp
rang the barbed-weapon scenario, designed to show why, without syntax, we could never have invented barbed weapons. I used it in several talks, and it went something like this:

  Any barbed weapon (dart, arrow, fishhook, harpoon) dates to within the last hundred thousand years of prehistory—in other words is indubitably the work of modern humans. What would you have to think, in order to even get the idea of making a barbed weapon? Surely, early Homo sapiens thought something like this: “When I put a smooth point into an animal, often the animal shakes itself and the point falls out. If the point falls out, the wound closes and no longer bleeds, so that the animal isn’t weakened and may escape. If I made a point in such a way that it would stay in the prey, then it couldn’t be jerked out, the animal would go on bleeding, be weakened, and either fall down or be captured. In the long grass I just walked through, burrs clung to my legs, seeds that have attached to them a little thing that catches in my skin and doesn’t fall out. What a great idea to make points with something like that!”

  When I used this example in talks, nobody ever stood up to state the obvious: What was to stop anyone thinking something like the following?

  “Me throw dart/spear. Point hit animal. Point fall out. Wound close. Animal get away. Suppose point stay. Animal bleed. Animal get weak. Catch animal. Look this seed. Seed stick in skin. Seed get little thing. Little thing stick in. Suppose point get same-kind thing. Maybe point no fall out. Me catch animal. Me kill animal. Me eat animal.”

  If what you just read sounds like some kind of pidgin, fine. Although for a variety of reasons it’s far from a perfect model, a pidgin is the nearest thing to protolanguage that we’re going to find in the modern world. Unfortunately, good pidgins are hard to come by in the modern world. I’m sure that everywhere there are incipient pidgins, wherever people with different languages come into contact with one another, but these seldom if ever develop into the full-blown thing nowadays. English is killing pidgins maybe even quicker than it’s killing established languages. Why try to start a new language if there’s a ready-made one already spreading across the globe like crabgrass?

  I was fortunate enough to arrive in Hawaii before the last of the good old-time pidgins disappeared, and I was sometimes amazed by what it’s possible to do with quite limited resources. Here’s just one example, a philosophic meditation on the vicissitudes of life using virtually no syntax and only twenty-two different words: “Some time good road get, some time all same bend get, angle get, no? Any kind same, all same human life, all same—good road get, angle get, mountain get, no? All, any kind, storm get, nice day get—all same, anybody, me all same, small-time.”

  Granted, this was a modern human fluent in his original native language (Japanese, as it happens) and granted too that you wouldn’t expect such flights from ergaster. I merely wanted to demonstrate that you can do far more with a tiny vocabulary, just stringing words together, than you might at first have thought.

  Just what do we gain by putting this into modern syntactic form?

  “Sometimes [when you’re traveling], you drive on good roads while at other times you meet with obstacles like bends and sharp corners, don’t you? Everything else is like that, human life is just like that—sometimes you find good roads, other times you find things that are like corners or mountains, you experience different kinds of weather, sometimes there are fine days and sometimes storms, aren’t there? Well, life’s just like that for everyone, same as it was for me when I was young.”

  It’s longer, it’s more verbose, prices we pay for eliminating obscurities and ambiguities and giving a smoother flow to things. In terms of the actual sequence of ideas, what’s to choose between the two versions? The first may lack grammatical structure, but it has as much semantic and pragmatic structure as the second, and which of those factors would be most important in thinking? Maybe thinking without syntax is less fluent than thinking with it. But, given the right words and enough of them, couldn’t the predecessors of modern humans have done just a little bit better than a million years with the same old hand ax? Wouldn’t they have done something to break what one paleoanthropologist referred to as “the almost unimaginable monotony” of the Lower Paleolithic, the Old Stone Age?

  The conclusion to which this seems to lead us is that protolanguage might not have reached even the level of an early-stage pidgin until our own species emerged. Or it may simply mean that, for reasons still unclear, the ability to connect words, to construct short and practical messages, emerged long before it became possible to link concepts into coherent trains of thought. That’s just one more of the many questions that we cannot yet answer.

  THINGS YOU CAN DO WITH WORDS

  As I said in chapter 2, I eventually came to agree with Terrence Deacon that it was symbolism rather than syntax that marked the boundary between humans and nonhumans. Originally I’d been skeptical about Terry’s claims; it seemed to me that the one thing apes couldn’t handle, and humans could, was syntax. But suppose it was the case that there was no hope of getting to syntax until you had symbols—and not just symbols that you’d had handed to you on a plate, the way chimps got them, but symbols you’d had to fight for and win over hundreds of thousands of years of glacially slow progress? Once that thought—what you might call the “can’t get there from here” hypothesis—had entered my mind, it began to seem foolish to get down on the chimps for failing at something they could never have been expected to do. Moreover—and this is one of the tests of a good thought—it started me thinking in new and productive ways about what it would really have taken to get from no language even to protolanguage.

  What words would have come first in protolanguage, and would it have made any difference?

  Remember all the things that have been advanced as possible selective pressures for language—child care, toolmaking, gossip, hunting, social maneuvering, and any more you can think of. Now even if, as I’ve proposed here, none of these did in fact select for language, it remains the case that they were all matters of greater or lesser concern for our ancestors, and they were all areas where, if you already had the beginnings of language, more of it would make a difference.

  Now do this thought experiment: make lists of what would seem to you to be the ten most useful words in each of these and other similar areas of interest. Put your lists side by side and see how much overlap there is.

  My prediction—very little, maybe even none. The protolanguage would have to be at least as big as the sum total of these lists. But even then would it have been big enough to be useful?

  To find out, ask yourself how much interesting gossip or useful hunting information you could convey even with all of those ten words, or to what extent you could use them to enhance your status in your social group. Do it the brute force way, just like I would: make all the theoretically possible combinations of these words and see how many of them would actually make sense. (Don’t cheat now—no syntax, and no “if”s, “and”s, and “for”s or any other of those meaningless words that glue the meaningful ones together and allow us to interpret them swiftly, quite automatically, and—most of the time—unambiguously.)

  Then ask what might have motivated the creation of any one of these words. “They wanted to talk about X” won’t cut it. They couldn’t yet know what “talking about X” meant, let alone want to do it. You have to try to think of some practical problem the new word might have solved, or some situation that might have pushed someone to create a particular word—anything at all that might have driven the word-making process.

  Believe me, you’ll be doing research that nobody’s done yet. You see, it’s far easier to talk in a vague general way about how protolanguage would grow and how it might work than to get down to the nitty-gritty, the sheer nuts-and-bolts of how all the plausible-sounding things people are constantly saying about early language might work out in practice. You’d think that people who confidently say language resulted from this pressure or that pressure would at least think about what words the
pressure in question would have had to evoke, and if it’s reasonable to suppose that naive hominids not that far removed from other apes could have or would have invented those particular words. But until this book, nobody, literally nobody, who ever wrote about language has ever suggested what the first words might actually have been, or under what precise circumstances they might have been uttered, least of all how those words might have fitted the particular selective pressure that the author, with such confidence, put forward as the crucial engine for language evolution.

  My guess is that none of our ancestors were actually trying to build a vocabulary, because none of them could have known what they were doing. While recruitment signals might in principle have triggered an “aha!” moment, a realization that now symbols could be linked with things other than big dead herbivores, there’s no compelling reason to believe that this actually happened. To the contrary, when we consider that the hominids in question were a lot closer, in minds and behavior, to apes than to us, it seems hardly likely that anything could have happened in quite so quick and dramatic a fashion. And when we recall what we saw in the previous chapter, that there’s good reason to believe that words did not follow but preceded concepts, the chances of an “aha!” moment sink close to zero.

  What is needed here is the opening of a new field of inquiry—the study of hypothetical early vocabularies and the communicative consequences of different word choices in forming vocabularies of a hundred words on down.