On Human Nature

I’m not recommending that others follow in my footsteps, but I saw a collection of old pieces published under the title The Chomsky-Foucault Debates on Human Nature and I picked it up. The Foucault stuff struck me as tediously outdated—who on earth still cares about the Marxist-Leninist view of something?—but Chomsky struck me as outdated too, just not tediously so.

Maybe I shouldn’t be surprised. The reprinted material was new in the 1970s. But these people are supposed to be talking about human nature, which hasn’t changed much in the past 30 years. Chomsky himself points out there has been “virtually no theoretical progress” in our understanding of human nature since Aristotle’s day, so if we can still read Thucydides for a good portrait of human nature, why can’t we read Chomsky-Foucault as well? It is not so easy to produce a classic text, even when addressing a classic theme.

I’m leaving Foucault for another blog, but Chomsky remains the grand old man of language studies and merits attention.

Chiefly, Chomsky’s view of language now strikes me as too abstract. Sometime back in the 60s when I first read his definition of a language as the set of sentences that can be generated by its syntax, I was intrigued by the novelty of the approach. Also, back in those days I was not so familiar with the traps people get themselves into when they try to understand human activity in terms of machine output. In particular, we don’t learn any of the sensuous, concrete reasons humans act as they do. Chomsky’s definition of a language offers an especially clear example of that failing. It gives us no clue about why anyone generates sentences.

Chomsky’s reply to that sort of complaint is visible in this volume. He states quite plainly that he is interested in studying language at an ideal level, and that this level provides the basis of success for all true science (e.g., physics). Chomsky directly rejects the kind of science (he calls it non-science) that keeps to “phenomena and their arrangement and such things” [p. 34]. I know what he’s talking about, I think. Galileo said that all objects fall at the same rate of speed, even though it was empirically demonstrable that they do not. He analyzed motion into two parts: falling itself, and the air that impedes falling, making feathers fall more slowly than bricks. Newton was then able to abstract that “falling itself” element and give us his law of gravity. Einstein eventually rewrote the law of gravity in a way abstract enough to explain why feathers and bricks fall through a vacuum at the same speed.

Science in this view can be seen as a process of increasing abstraction. Start with phenomena. Analyze them into abstract laws and empirical parts, identifying the laws as reality and the empirical parts as circumstantial ephemera. The point of a hypothesis is to test the laws, not the phenomenon.

Chomsky’s famous distinction between linguistic competence and performance seems an attempt to imitate what Galileo accomplished. On the one side we have the rules of a language, the laws of, say, English, and on the other side we have the various frictions that give us empirical examples that stray from what the laws say we should encounter. Where Chomsky deviates from the history of physics—and where he reveals himself as a kind of mathematician rather than a kind of scientist—is that he proceeds to study only the laws while ignoring the phenomena altogether. For the physicists, the road to the laws runs directly through the phenomena. An empirical approach to syntax would examine the problem speakers have in figuring out the difference between whoever/whomever he saw or if I was/were king. What gets in the way of the rules?

Through painstaking observation and experiment, Galileo and Kepler were able to announce some laws of planetary behavior and of motion. Newton then came along and used those laws to explain the observable phenomena of the night sky. Building on Galileo, Newton asserted that objects in motion move in a straight line at a constant rate. One of Kepler’s laws is that planets move in ellipses. How are we to reconcile the two? Introduce another factor, a gravitational force, that explains how a planet is diverted from its straight line into an elliptical orbit. The accuracy of Newton’s predictions struck many people as miraculous, but the accuracy was possible because it rested on a series of extremely accurate observations.

Chomsky pays no such attention to empirical reality. The puzzle that holds his attention is the capacity of humans to use syntax despite the “really quite small quantity of data, small and rather degenerate in quality, that’s presented to the child.” This observation was enough to show that behaviorist theories of language could not explain what was going on, but it was not enough to provide a scientific alternative. Similarly, Kepler’s observations were enough to discredit the Ptolemaic belief that the planetary motions are circular, but there was still plenty of empirical work to do before we got a law of gravity.

That further work is what strikes me now as missing from Chomsky’s discussion; it is what has kept the theory so abstract. News reports on this blog have shown that other investigators have tied linguistic studies to something more concrete. A language is not merely a set of rules for combining symbols. At its core, speech is an activity in which two people pay joint attention to a topic by trading roles as speaker and listener.  There are ample variations on this theme—one person does all the talking, a lone individual talks/thinks quietly, many more than two people participate, etc.—but they do not subvert the core understanding which is that language is a concrete activity requiring attention to a topic.

Under this definition words are not abstract symbols; they are tools for directing attention. Meaning is not some occult system for transforming noise into symbols, but what you get when attention is directed somewhere. Syntax is a system for keeping attention focused, not a set of abstract rules for combining words.

This view of language allows for a concretized version of Chomsky’s abstract argument:

1. All known languages are reflect certain identical “limitations” on the syntax they use. For example, in all languages, a sentence like “He wore a black hat and a white coat,” means the hat is black and the coat is white. There are no languages in which this kind of grouping could mean the hat is white and the coat is black. We take that for granted, but it would be pretty easy to create a programming language that did it differently.
2. The universality of these limitations despite the very different histories and experiences behind different languages can be explained only if individual speakers contribute a great deal to the syntax they use. In other words, language is not something simply imposed on people by the circumstances in which they were born. Thanks to my environment, I speak English; but thanks to my human nature, I speak.
3. Individuals learn to speak particular languages because they pay attention to the world in similar ways. For example, I perceive a hat as black and a coat a white. Therefore, when I start to speak about these things, I mention the black hat together, not because I know some abstract rule but because when I pay attention to the object I perceive its “hatness” and color together. The universals of language reflect something of the universals of attentive perception; the many cultural differences of language reflect the many different ways people can pay attention to something.
4. The abilities that let us select topics, share attention, and alternate between speaking and listening roles are universals of human nature. They enable us to recognize a common humanity in people whose lives and cultures are very different from our own, and their absence in other primates indicates that the nature is truly human and not merely part of ape  or animal nature.

Points 1 and 2 come directly from Chomsky and (I think) we are in agreement on them. Chomsky’s version of point 3, however, is much more abstract. He talks about having an innate “explicit and detailed schematization” that keeps straight things like black hats and white coats. Rules, not the experience of perception, are at work in Chomsky’s story.

Point 4 strays even further from Chomsky for it proposes humans as agents capable of meaningful behavior. Apt speech reflects an ability to select, to share, and to alternate roles. It’s not clear how much agency Chomsky grants humans. In his political debate he insists on the idea that creative behavior is real, but his version of point five seems to make human nature a function of programming rather than biology: “this mass of schematisms, innate organizing principles, which guides our social and intellectual behavior, that’s what I mean to refer to by the concept of human nature.”

An inquiry into speech origins could not have gone anywhere without Chomsky’s work, but re-reading him at this late date tells me it also would have gone nowhere had it stayed with his abstractions.

Comments

[Jeremy]

This debate (from 1971) is made more interesting when compared with the debate between Chomsky and Jean Piaget from 1975 (published in English in 1980 by Piattelli-Palmarini).

In his book "Structuralism" (1968/1970), Piaget endorsed Foucault's basic position (while also criticising him) on "épistèmes." For this reason, the later debate can be seen as an advance (of sorts) on the earlier one.

[TLTB]

It is not true that chomsky and his followers ignore empirical facts. We are all, in fact, obsessed with them. Where your contention lies is in the nature of the facts we are interested in - we don't care much about why people say things or how they use certain constructions. We only care that they say them (and that they do not say certain alternatives). We study language as a natural object and we do so just as any scientist studies a natural object - by removing it from its natural environment and make certain ideal assumptions. The abstract system of mathematical relations that underlie language are essential to explaining language - no theory of joint attention is going to tell us why in some language resumptive pronouns are required in extraction contexts while in others they are not, or why the that-trace effect is so strong in some languages, but absent in others. It is precisely the lack of any plausible functional or semantic explanation for phenomena like these (and many others) that demands the sort of abstract explanations chomsky and co. are after.

And not all science proceeds just from pure empirical observation. Galileo and Einstein tell us this: Galileo hated Kepler's theory of ellipses precisely because it was mathematically inelegant; the same striving for mathematical elegance pointed Einstein to many of his conclusions. It was only later that they were supported by observation. The same can be said of minimalist syntax today: we are assuming that things are mathematically simple and elegant and then checking to see whether the facts can be explained in such systems. It so happens that actually a lot of things fall into place when these sorts of mathematical assumptions are made. Chomsky has repeatedly stated how surprising this is, since we don't expect such elegance from biologically-based systems.

there is no doubt that a theory of joint attention must play a large role in a theory of language and language evolution. but it cannot be the whole ballgame, and it is clear that abstract mathematical principles play a crucial role in the nature of linguistic structure just as they do in vision and other cognitive abilities.

----------------------------

BLOGGER: I was wondering where TLTB had gone and figured that if this post could not smoke him out, he had moved on to greener pastures.

The vital point hers is in the remark, "We don't care much about why people say things or how they use certain constructions." We all get to be interested in whatever we are interested in, but if you don't care much about the motives or fruit of human behavior you are not going to have much to say about human nature (the supposed subject of the debate).

I also find it odd to defend a break with empiricism by citing Galileo's quarrel with Kepler, since Kepler was right. His discovery of eliptical orbits was the fruit of years of precise observations; Galileo scoffed out of an a priori prejudice.

Finally, strange as it seems, I know a ton about Einsten, and dispite his theoretical focuse he always began with empirical data. He had little taste for mathematical elegance, although the power of mathematics to solve the problem of general relativity much impressed him. Yet even general relativity was based on empirical demands. He only buckled down to the job of revising Newton when he realized his work could be tested by observation during a solar eclipse.