Adam's Tongue: How Humans Made Language, How Language Made Humans Page 5
By this time you may be thinking, Well, how could language have evolved, anyway? How could anything have met all the conditions I’ve described to you? You may even be thinking, Hey, maybe it didn’t evolve. Maybe Intelligent Designers are right after all; maybe it was a magical gift from above, sprung ready-made from Jove’s brow, inexplicable (as some have suggested) by any human brain. Or maybe we’re in something like The Matrix, and the whole thing is a colossal illusion—we don’t really have language, we just think we do.
Stay calm, folks. We have language and sure enough it evolved, despite all the seemingly insuperable obstacles that lay in its path.
But I’ve made one thing pretty plain, you may think. It can’t have evolved, as most biologists would claim, from some prior means of communication, some ACS of the last common ancestor that . . . somehow . . . gradually . . . got modified . . . or something. It must have evolved from . . . well, from something else. What, exactly? Well . . . it’s hard to say . . . but something.
That’s exactly what I thought fifteen, twenty years ago. And until relatively recently, for that matter. After all, I was the guy who produced the continuity paradox: “Language must have evolved out of some prior system, and yet there does not seem to be any such system out of which it could have evolved.”
Then how did it evolve? In my earlier work I talked about mental representation systems—maps of the outside world and everything in it—that grew in the brain, over countless millions of years and thousands of species, till they got detailed enough to divide up the world into word-sized bits, just waiting to be given their lexical labels. Once these bits—prelinguistic concepts—were ready, then in some rather ill-defined way, connected somehow with protohuman foraging strategies, a protolanguage, quite different and separate from the protohuman ACS, just somehow popped out. After which, a handy mutation turned protolanguage into language.
Blame it on the rashness of youth (after all, I was only sixty-four at the time). And it wasn’t that bad for a first try. Language and Species was the first book I know of that tried to work through the whole process of language evolution with some degree of detail and depth. The problem was I didn’t have a good framework. Niche construction theory hadn’t been developed. When I didn’t know something, I filled the gap with what the philosopher Daniel Dennett calls “figment” (as in “figment of the imagination”). And I didn’t do what I’m doing here: working through the precise relationship between human and nonhuman communication systems in what I’m sure is excruciating detail (sorry about that, but any serious study is like athletic training—no pain, no gain).
Reactions to what I had written seemed only to confirm my position, at least as far as the continuity paradox was concerned (biologists wouldn’t swallow the mutation, and they were right not to, of course). I hadn’t expected that I’d stop people from believing in continuity, but what did surprise me was that they went on believing without even trying to refute my arguments; blind faith is far commoner in science than we like to admit. So I was not, in any sense of the word, converted by continuists. I converted myself.
It came through trying to think like a biologist. This is not all that easy for people from other fields. What makes interdisciplinary work so hard is that any academic discipline acts like a straitjacket, forcing you to look only in certain directions, blocking other perspectives from view. It takes a good deal of conscious effort, plus a lot of soaking yourself in other people’s literature, to overcome this state of affairs.
The process was speeded up by my encounter with niche construction theory, which made sense of a lot of things that had baffled me. I started to rethink the continuity paradox. Suppose, just suppose, that one took an engineering perspective and started asking whether there was anything in an ACS you might plausibly change that would make it more languagelike. If there was, the next question was whether that thing might plausibly arise through constructing a particular kind of niche. If it could, the next question was: Was there such a niche in human prehistory?
The rest of this book is about the answers to these questions.
2
THINKING LIKE ENGINEERS
SETTING THE BAR
Let’s pretend we’re language engineers, given the task of providing an alingual species with language.
We have to work on a species that has only a standard, average primate ACS. We have to get that species not to full language—that will have to come a lot later—but to something that points away from an ACS, in a direction that could plausibly lead to language. It doesn’t have to be a giant step. Better by far if it’s a small step, because the smaller the step that has to be made, the more plausible it will be from the standpoint of evolution.
But before we can make a start, we have to know where we’re headed. We have to look at language and what it does that an ACS can’t do.
A lot of people have tried to do this, but they’ve set the bar too high.
They compare an ACS with the kind of language that all of us speak today. And they point to evidence like the fact that language consists of three quite clearly distinct levels. Autonomous levels, they call them, which simply means that although all the levels interact whenever we speak, each abides by its own set of rules, each set being different from the other two sets.
There’s a level of meaningless sounds: phonology. Not one of the sounds we use in language means anything, in and of itself. But they’re not meaningless in the way grunts or coughs or sneezes are meaningless. Put a grunt and a cough and a sneeze together, and what have you got? A head cold? Well, nothing that means anything. Put two or three speech sounds together, and what you’ve got might be a word. Potentially, at least—to find out whether it is or not, you’ve got to go up to level 2.
That’s the level of meaningful sound sequences: morphology. That means words, and those bits of things we tack onto words, all the uns and dises and ings and eds that are meaningful too, but only when attached to a word stem. Now we can provide names for things, or to be more precise, names for classes of things—“dog” doesn’t mean this dog or those dogs, it means just a particular kind of animal. Except for one-word exclamations—“Help!,” “Fire!,” and the like—we still really can’t say anything that means very much. For that, we’ve got to go up again, to level 3.
That’s the level of meaningful utterances: syntax. You may mean, but you can’t mean much until you start to put words together into phrases, clauses, sentences. But once you can do a sentence, you’re home free. There’s literally no limit to what you can produce—paragraphs, pages, essays, books, encyclopedias . . . Given the rules of syntax, you can go on churning out language from here to kingdom come.
Now when you see this degree of complexity (and I’ve only skimmed the surface; each level has its own awe-inspiring convolutions) your only rational reaction is to throw up your hands and say, like the Maine farmer:
“You can’t get there from here!”
Well, that was funny the first time around, but the Maine farmer was full of his own cows’ product. You can get from anywhere to anywhere, with the right map. And the right map for this particular bit of country is only now going to be unfolded.
If you insist on comparing ACSs with language as it is today, you’re just setting yourself up for a fall. There’s a much better model to hand.
PIDGINS FLY TO THE RESCUE
I had the great good luck to come to language evolution from the study of pidgins and creoles.
A pidgin is what people produce when they have to talk to other people but don’t have a common language. If you want to know more, read my book Bastard Tongues. For now, it’s enough to know that you yourself may well have pioneered an incipient pidgin, if on vacation in a place where you didn’t speak the language you’ve struggled to communicate your needs to people and they’ve struggled back to try to make you understand them. The only reason your makeshift efforts didn’t evolve into a full-fledged pidgin is that it was just you and a couple of other
people over a period of days. If it had involved most people in the same community over a period of years, those efforts would have resulted in a true pidgin, as sure as puppies turn into dogs.
Think back to what you did. You used any words of the other person’s language you happened to know, but you didn’t put them together in any systematic way. Why not? You might well say, “I didn’t know how, in that language.” Sure, but what was to stop you from putting them together the way you did in your own language? Partly the fact that they were, in every sense, “foreign words”—they were foreign to you, you had to grope for them, so they popped out one at a time, with big gaps in between while you went looking for the next. Partly because you didn’t know all the words you would have needed even for the simplest sentence. You went with what you’d got; when you couldn’t think of a word, you used one out of your own language, or some other language you might know, and hoped the other fellow might know or guess what you meant. And if that still didn’t work, you pointed or gestured or mimed. You used anything at all that might work.
That’s the nearest you or I or anyone will ever come to feeling what things were like at the dawn of language. Still a long way, since it’s as hard for us to forget that we already have language as it is for a jury to heed the judge’s admonition, “Forget anything you may have read or heard about this case.” But it helps. If you’ve had this particular foreign-travel experience, pause a moment here to savor the memory of it.
Not everyone thinks I’m right about this. Dan Slobin, a psycholinguist (that doesn’t mean a linguist who’s psycho, it means one who studies the relations between language and human psychology) at the University of California, Berkeley, thinks a pidgin is not necessarily a good model for early stages of language. He points to the fact that people who create a pidgin already have at least one full human language, whereas the protohumans who started language clearly didn’t.
Now I have great respect for Dan when he’s on his own turf, which is how children acquire their first language. And he’s right insofar as the difference he brings up is a real one. But to clinch an argument, it’s never enough just to point to a difference. You have to explain why this particular difference makes a difference. Most differences don’t. There are big birds and small birds, birds that fly and birds that don’t, but a bird is a bird and we all know one when we see one.
The same is equally true of any variety of what, for want of a better word, we’ll call protolanguage (not to be confused with protolanguages, which are the hypothesized real-language ancestors of existing language families—Indo-European, for example—and seldom more than five thousand years old). Protolanguage is not true language, but it’s made up of languagelike elements. Since I introduced the notion in my 1990 book Language and Species, it’s been accepted by most researchers in the field that the emergence of language as we know it had to be preceded by something intermediate between true language and an ACS, and (by at least some of those researchers) that things similar to this intermediate can still be seen in the world around us—in pidgins, in the speech of the brain-damaged or of infants, or the productions of apes that have been taught various forms of signing.
In determining whether something qualifies as protolanguage, what matters is not whether you (its speaker) do or don’t have a language already; what matters is whether or not you’re in the same situation, that of having to communicate without a proper language to communicate in. Now the content of protolanguage, what you actually say with it, will vary depending on who’s using it—on whether that’s a pidgin speaker, or a Broca’s aphasic, or a “language-trained” ape, or a child younger than two, or a protohuman at the very dawn of language.
What won’t vary at all will be certain limitations—purely formal, structural limitations—on how you can express that content. Regardless of who or what you are, even of what species you belong to, these limitations will reduce you (if you already have a language) or exalt you (if you don’t yet have one) to short and shapeless and disconnected utterances.
If you’re a human speaking a pidgin, the pieces that the pidgin is built with will be ready-made words from one language or another. If you’re a protohuman pioneering language, they won’t be. If you’re a human speaking a pidgin, bits of the syntax of your own language may pop up here and there, although that’s unlikely in the early stages, when you won’t be fluent enough even to cannibalize your original language. If you’re a protohuman pioneering language, there’ll be no bits there to pop up. But in both cases, there won’t be anything you could call structure. No third level, because you have no rules, and if you have no rules, you have no syntax. No second level, because although you have words, those words have no internal structure and so can’t be broken down into bits like anti-dis-establish-ment-arian-ism (“opposition to the withdrawal of government support from a state-sponsored church”) or given inflections to indicate things like tense or number.
Just one level, where what you see is what you get.
But it’s still not down to the level of an ACS. Protolanguage and language share one big thing that ACSs uniformly lack.
Combinability.
ONLY CONNECT (IF YOU’VE GOT ANYTHING
TO CONNECT WITH)
Languages combine lawfully and protolanguages combine lawlessly. In other words, languages have all kinds of constraints on what you may put together with what; protolanguages don’t. Where things can be put together, languages have rules about which goes first; for instance, adjective before noun in English, noun before adjective in French. (Yes, I know we say “court-martial,” not “martial court,” and the French for “good luck” is bonne chance, not chance bonne, but these are exceptions that go against the grain of the language.) Pidgins and other forms of protolanguage don’t have such rules. You can put anything with anything, in any order, provided that the combination is meaningful in some way. But the bottom line is, you can still combine.
ACSs can’t. So far as we know, yet. And I would say that no matter how long or how hard we look, we’ll never find an ACS that can combine stuff. In a moment we’ll see why.
To find animals that can indeed combine communicative stuff is the Holy Grail for those who believe that ACSs segued seamlessly into language—strong continuists, let’s call them. Such a discovery would mean that those animals had a true precursor of syntax, and syntax is believed by some to be the only uniquely human part of language. Consequently, a true precursor of syntax would be a stunning defeat for the “language is something completely different” crowd. Needless to say, any number of ACSs have been searched and researched for the slightest trace of such a precursor.
The latest candidate will give you some idea of how desperate this search is getting.
Diana monkeys and Campbell’s monkeys are two species of African monkeys that inhabit the same territory. Both species give alarm calls on the approach of predators, and Diana monkeys respond to Campbell’s monkeys’ alarm calls just as readily as to those of their own species. With one difference: sometimes a Campbell’s monkey will preface its call with what is known as a “boom” vocalization, a brief low-pitched sound that occurs, on average, about thirty seconds before the alarm. Boom-preceded calls usually relate to some fairly distant predator or to some unexplained event that might mean danger of some kind. When Diana monkeys hear a boom-prefaced call, they seldom respond, and usually go on with whatever they’re doing.
Klaus Zuberbühler, a researcher at the University of St. Andrews in Scotland who discovered this behavior and tested it experimentally, wisely hedges on whether it really is syntactic. What’s puzzling is why anyone would think it might be. In the first place, something that involves two separate species is not very good evidence for what’s supposed to be happening in just one of them. In the second, a receptive skill—ability to determine the meaning of a sequence—is no guarantee of an equivalent productive skill—ability to actually combine things. But it’s mostly the kind of relationship between the two units,
the boom and the alarm, that makes this a dubious precursor for any kind of syntax.
Zuberbühler says that the booms “act as a modifier” to the alarms. Not so; they don’t modify them, but cancel them out. Do you know of any language with a word that means “Please disregard the next word”? I don’t. Typically in any combination of two language units, be they words, phrases, or clauses, one unit truly “modifies” the other by making its meaning more precise:
The English teacher (not just any old teacher).
Shake well before opening (if you do it afterward, you’ll get it all over your shirt).
Male monkeys mate when they see the typical female swelling (not just any old time, like us).
This is what I was talking about just now when I said that a combination, whether linguistic or protolinguistic, had to be meaningful in some way. This is the way—taking something (a subject) and saying something about it (a predicate). Predication is one of the most basic and fundamental processes in language. Syntax may not have any animal precursors, but predication surely was the precursor of syntax. If units couldn’t first combine on the basis of meaning, they’d never have gotten to where they could combine on a structural basis.
So the next question becomes, if units of language and protolanguage can combine, and units of ACSs can’t, why is this? Is it just happenstance? Are the animals not as smart as us? Or is there a principled reason why they can’t, one that makes looking for syntactic precursors among animals a waste of time?
ANIMAL WORDS?
The search for syntactic precursors mightn’t be such a waste of time if animal calls were, in fact, precursors of words.
That’s the other Holy Grail hard-line continuists seek for: things in animal communication that are precursors of true words. The best candidates so far are the alarm calls of monkeys, especially the most thoroughly researched of these: calls made by the vervet monkey of East Africa. Indeed, the poor vervets must by now be heartily sick of being dragged in every time anyone writes about language evolution.