The First Word: The Search for the Origins of Language

by Christine Kenneally

  In that book he continued his examination of language in the abstract, discussing the grammars of languages in a wholly new way. Instead of simply being a catalog of all the words and sounds in a language, with instructions for how to put them together, a grammar, he argued, was really a theory of that language.

  As a theory, a grammar should be judged in the same way all scientific theories are: it should explain as much as it can with as little as possible. It should be simple and elegant. Viewed this way, possible grammars of a language can be compared in the same way that different theories in science are: the successful one more fully explains the phenomena in question in as economical terms as possible.

  Syntactic Structures, for example, contrasted two methods for writing a grammar. The best method, said Chomsky, collapsed all of language into a set of rules. And in much the same way that software generates output in a computer, those rules can generate an entire language. For example, an English sentence can be described as “S goes to NP VP,” meaning that a sentence (S) consists of a noun phrase (NP) and a verb phrase (VP). “NP goes to Det N” means that a noun phrase consists of an “a,” the determiner (Det), and a noun (N).7

  Chomsky also pointed out that the set of language rules could be made smaller and simpler if you included ways to relate certain sentences to each other. “The man read the book” and “The book was read by the man,” for example, have a striking similarity. Instead of having separate rules for each of them, Chomsky suggested that the more complicated second sentence was derived from the first. He called this a transformation.8

  If the phrase structure analysis of “The man read the book” is “S goes to NP1 VP NP2,” then “The book was read by the man” can be represented as “S goes to NP2 VP by NP1.” In this way, the relationship between all the simple active sentences of English and their passive versions can be described by just these two simple structures and the transformational rule that links them.

  Language, in this view, is basically a set of sentences. And the job of a grammar, or theory of language, is to generate all of the language’s allowable sentences (“The cat sat on the mat”; “The plane was rocked by turbulence”) but none of the bad ones (“Cat mat the on sat”; “Turbulence plane by the rocked was”). A grammar generates all possible utterances of a language, Chomsky said, “in the same way that chemical theory generates all possible compounds.”9

  Syntactic Structures got Chomsky some attention, but at the time of publication it wasn’t especially well known. Two years later Chomsky made a much larger splash when he published a review of B. F. Skinner’s Verbal Behavior. The review appeared in what was at the time the premier journal of linguistics, Language. Skinner, a psychologist, was already well known for his theory of behaviorism. In its simplest form, behaviorism says that all animals, humans included, are like machines—if you press their buttons in the right way, they’ll respond automatically. The appearance of emotion or thought is irrelevant, because everything can be reduced to behavior. As long as you know what kind of machine you are dealing with—human, feline, avian—you can control its behavior. Even very complicated behavior can be reduced to a series of depressed buttons.

  At the time, people spoke about Skinner in the terms they would later use to describe Chomsky. In her book Animals in Translation, Temple Grandin wrote about the behaviorist’s influence when she was a college student. “Dr. Skinner was so famous,” she remembered, “just about every college kid in the country had a copy of Beyond Freedom and Dignity on his bookshelf.” Of behaviorism she added, “It’s probably hard for people to imagine [the power] this idea had back then. It was almost a religion. To me—to lots of people—B. F. Skinner was a god. He was a god of psychology.”10

  Chomsky’s review was published two years after Skinner’s book came out, oddly late in the day for a book review, even in academia. Nevertheless, it had an immediate impact. Skinner suggested that language was a simple behavior, a notion Chomsky dismissed as absurd. Skinner was used to dealing with lab rats, but pressing a pellet for food is no analogy for producing language. In order to speak, people use great creativity while obeying many complicated rules.

  Chomsky argued: “A typical example of stimulus control for Skinner would be the response to a piece of music with the utterance Mozart or to a painting with the response Dutch. These responses are asserted to be ‘under the control of extremely subtle properties’ of the physical object or event.” But, argued Chomsky, what if we don’t say “Dutch”? What if we say, “Clashes with the wallpaper, I thought you liked abstract work, Never saw it before, Tilted, Hanging too low, Beautiful, Hideous, Remember our camping trip last summer? or whatever else might come into our minds when looking at a picture”? People are not controlled by some unknown aspect of a painting, he said. Their response comes from inside them and is facilitated by the infinite creativity of language.11

  The key idea in Skinner’s behaviorism—if you push someone or something in the right way, it will respond in a predictable manner—was called stimulus-response. But when it comes to language, Chomsky said, particularly when children learn language for the first time, stimulus-response is not a relevant model. What is fundamentally interesting about language is the incredible speed with which children learn thousands and thousands of words and the many rules that combine them. In fact, there just isn’t enough information in the language children hear in their day-to-day lives for them to divine all the rules that they come to know how to use. Chomsky called this phenomenon “poverty of stimulus.” So how do children learn how to speak if language is so incredibly complicated? They must come to the task somehow prepared, he concluded. They must be born with a mental component that helps them learn language.

  It was as if Chomsky had delivered unto Skinner and behaviorism a knockout punch.12 The review garnered enormous amounts of attention from people in all sorts of disciplines. For many academics, this was the moment at which Chomsky seized their attention and would hold them riveted from then on.

  The young professor was propelled into the limelight, and even though his review was widely criticized as glib, biting, and angry, it was these very qualities that seemed to thrill people. As much a polemic as a review, the article was described as “devastating,” “electric,” and a superb job of “constructive destruction.” Chomsky the linguistic freedom fighter was born.13

  Skinner responded that Chomsky hadn’t understood what he was saying, that in some respects it seemed that Chomsky had intentionally misinterpreted him, but the damage was done. From that point on, the obvious influence of behaviorism seemed to fade.

  It took a few years for the impact of Chomsky’s first book to be felt, but by 1964 Charles Hockett, one of the most eminent linguists of the time, described Syntactic Structures as among the field’s few “major breakthroughs.”14 Howard Maclay wrote: “The extraordinary and traumatic impact of the publication of Syntactic Structures by Noam Chomsky in 1957 can hardly be appreciated by one who did not live through this upheaval.”15 Ray Jackendoff remembers that in 1965, when he began his graduate studies (with Chomsky), “generative linguistics was the toast of the intellectual world.”16 Daniel Dennett, the well-known philosophy professor at Tufts, wrote in Darwin’s Dangerous Idea that he could “vividly remember the shockwave that rolled through philosophy when Chomsky’s work first came to our attention.”17 Looking back, Chris Knight of the University of East London wrote that Chomsky may as well have thrown a bomb.18

  In less than a decade, people were proclaiming a psycholinguistic revolution.19 Many young scholars flocked to MIT to work with Chomsky on his new generative linguistics, and in many other universities researchers began to search for the mental component containing the basic, innate generative rules of language with which children are born.

  Chomsky’s theory was expanded and his reputation solidified with Aspects of the Theory of Syntax, published in 1965. A slim but extremely difficult book, Aspects further explained key Chomskyan concepts like dee
p structure and surface structure, and has since become a classic text.

  All the ideas in Aspects rest on the notion that language can be divided into, on the one hand, everything that goes along with actually speaking in a given situation and, on the other, all that is stable and universal. Chomsky called this the difference between competence and performance. Competence, which includes syntax (a perfect, mathematical system), is the innate basis of language and is the same from speaker to speaker. Performance includes whatever is individual or context-specific in language: the myriad differences in the way we pronounce “ketchup,” the use of gesture, the “ums” and the “ahs.”

  Even though he imagined an idealized speaker and hearer as the subjects of his research, language in the Chomskyan sense had little to do with the fact that it overwhelmingly takes place between people. For the Chomskyan linguist, to study what was interesting about language was to discard any variation, the way any given speaker actually speaks, and to focus instead on the skeleton that remains.

  The role of the language specialist was fundamentally changed by these ideas. Linguists were no longer mere catalogers but scholars who were perfectly positioned to unearth the deepest mysteries of their subject. What mattered about a language was not that it came from a particular region like the plains of the Midwest, the villages of Mexico, or the beaches of Asia but that it came from our heads. With generative linguistics, the terrain that the linguist explored shifted from the corners of the planet to the depths of the human mind. Universal grammar specified every rule for every language, and that controlled a child’s ability to develop the correct rules of syntax of each language. It was believed in the early days that universal grammar, or the language organ, was hardwired into people’s brains. Anyone born with UG, which is to say everyone, was born with the potential to learn any language.

  Even though searching for the universal principles of language was hugely different from the way scholars had previously thought about language, early generative linguistics still divided language in the brain in much the same way that linguists of the 1950s had divided languages in the field. Field linguists wrote a grammar by analyzing its structure, sound, and meaning in separate sections. They also believed that when you were learning a language from scratch and assembling its grammar, you should keep these parts of language completely separate—you should never mix levels.

  Generative linguists began to divide language in the brain in the same way. They looked for evidence of a module that controlled syntax, a module that controlled meaning, and a module that processed sound. It was thought that these modules were independent of one another and that language was produced by a coarse-grained interaction between them. Additionally, the separate systems of language had their own subsystems. For example, the syntactic module was made up of a set of smaller modules, each dealing with a different part of syntax, each autonomous.

  In this model, when someone heard speech, the separate modules divided up the signal. The syntactic module extracted from the sound wave all the information regarding syntax, the intonation module analyzed all the pitch variation, and so on. Once each module had sufficiently analyzed the component for which it was responsible, the brain put them all back together as language. One implication of this theory is that when you heard someone else speak, the grammar part of your brain somehow extracted the grammatical information from the sound waves but ignored any other information in those waves that might help interpret it.

  The workings of the language organ were also thought to be completely separate from other parts of the brain. They were separate from the context of spoken language, and they were also completely different from similar systems, like music. Gesture was peripheral and uninteresting. Moreover, human language was entirely distinct from the communication that takes place between other animals. This model of language was consistent with general theories at the time about how the brain functioned—namely, as a series of separate boxes, each of which computes different parts of the world.

  Critics said the model was merely a new version of phrenology, a nineteenth-century “science” that held that for every tendency in an individual, there was a corresponding spot on the brain that controlled it. The brain would bulge or recede in these areas, depending on how developed a given trait was. (Phrenologists even believed that the skull would echo the shape of the brain, so that a person’s character could be read by the bumps and pits of his or her head. For example, someone with a great deal of self-esteem would have a big bump right at the top and back of her head. Phrenology is now the iconic example of silly science.)

  The Chomskyan deconstruction of language was, on the one hand, counterintuitive. The average person who hadn’t taken a university course in linguistics and been rigorously trained to force these elements of language apart would probably consider context crucial to understanding language. He would count intonation as important, and he would be unlikely to completely separate structure from meaning.

  Yet Chomsky’s approach satisfied another kind of intuition: to divide an object into its essential and incidental parts. With language, generative linguists tried to strip away everything peripheral, anything that could be stripped. The hope was to expose the bare bones, discover what was indivisible, and unearth the core.

  Another key insight that Chomsky brought to language studies was the infinitude of language. While so much of language is rote, consisting of things that you have heard before, you don’t have to go far to find words assembled in a way you’ve never heard them put together. Chomsky described this as the infinite use of finite means, calling it “discrete infinity.”

  With discrete infinity, “Kate read the book that Bill wrote” can be embedded in “Ally saw,” becoming “Ally saw that Kate read the book that Bill wrote.” It can be further embedded into something like “Andrew explained how Ally saw that Kate read the book that Bill wrote,” and so on, ad infinitum.

  Ten years after Syntactic Structures was published and two years after Aspects, most papers presented at the 1967 meeting of the Linguistic Society of America discussed Chomsky’s transformations.20 A few years later Chomsky’s growing reputation within linguistics and philosophy had spread into many other fields. In 1970 a Chomsky monograph was published in the Viking Press Modern Masters series, putting him in the company of Einstein and Freud.

  Of course, Chomsky had detractors at this time as well, and the louder his supporters became, the more his critics grew in number. In 1967 Charles Hockett, who had just three years earlier hailed Chomsky’s genius, called him a “neo-medieval philosopher.” Another prominent linguist, George Trager, described him a year later as “the leader of [a] cult…with evil side-effects.”21

  Chomsky’s skirmish with B.F. Skinner turned out to be merely the first in a long line of infamous, bitter conflicts. The next took place in the late 1960s and early 1970s, when a group of linguists calling themselves generative semanticists argued that separating language from the way it was used was ridiculous.22 This group believed that the most fundamental organizing principle of language was its meaning (semantics), not the way it was structured (syntax), as Chomsky’s transformational theorists believed.

  The generative semanticists defined themselves in opposition to the Chomskyan juggernaut, and as Randy Allen Harris (the main historian of this period) recounted it, that opposition took on all the flavor of the 1960s counterculture—irreverent, exuberant, and combative. Their criticisms of Chomsky extended from the way that he divided up language to his ascetic style. One running joke of the era was inventing a title for the world’s shortest book, like “Problems of the Obese” by Twiggy; a popular candidate among linguists was The Bawdy Humor of Noam Chomsky.23 In turn, the generative semanticists were caricatured as unthinking followers of a fad. Chomsky repeatedly insisted that they didn’t actually understand the theories with which they took issue.

  There is a clear pattern in these different conflicts. Again and again, Chomsky’s critics claimed that h
e chose data to support his theories but then discarded it when it no longer suited, and that he intentionally misinterpreted his adversaries and then launched an attack against his own misunderstanding. People also accused him of abandoning ideas that he once promoted without acknowledging that he had changed his position. Another complaint was about the way Chomsky dealt with counterevidence to his theories, most of which he insisted could be simply disregarded.

  When Chomsky put forth his ideas, he typically dictated the terms with which people could reasonably disagree with him. Academics objected to the fact that he laid out his argument and the rules for argumentation at the same time. For instance, he said, “Counterexamples to a grammatical rule are of interest only if they lead to the construction of a new grammar of even greater generality or if they show some underlying principle is fallacious or misformulated.”24 That is, critics could not simply point out that something didn’t work; they had to come up with a new theory in its place that did.

  As relentless as the expansion of Chomsky’s vision seemed to be, it deflated unexpectedly in the 1970s. Part of the appeal of generative linguistics was the way it rendered sentence analysis into mathematical-looking algorithms. Rules like “S goes to NP VP” gave language study a scientific veneer. It turned out, however, that this was not what actually happened in the brain as it processed language.

  If deep structures really existed, it was reasoned at the time, you’d expect people to take longer to understand the more complicated, transformed structure of a given sentence than its simpler basic form. But when psycholinguists tested this in experiments, it did not pan out: the derived sentence took the same amount of time as the basic sentence.

  Soon the voices that had criticized Chomskyan linguistics from the beginning grew to a din. As researchers found that the notion of an innate language organ was not supported by real-world evidence, they became interested instead in the idea of general foundations for language and thought. Even the popular press ran articles about the Chomskyan revolution and declared it over.25