The First Word: The Search for the Origins of Language

by Christine Kenneally

  He ended up writing his Ph.D. thesis (which eventually became the second book ever published by MIT Press) about how the physiology of breathing structures how we speak. Speakers make all sorts of muscular maneuvers in articulating words, and these are carefully controlled to make sure that the air pressure generated by their lungs stays at a steady level as they talk. Lieberman found that these maneuvers are keyed to the length of the sentence we intend to speak, showing that humans anticipate a long sentence before they utter a sound. The more he became engaged with these fundamental physical constraints on human language, the more he moved away from the abstract properties of language and toward all the things that Chomsky had dismissed as epiphenomena.

  The problems of speech synthesis and voice recognition are far from solved today. When Lieberman began to wonder about speech, scientists were just beginning to get a glimmer of how complicated it was, and how enormously difficult it was to get a machine to either produce or understand speech. (One of the big differences between now and then is not that the problems have been solved but that researchers have come to appreciate the magnitude of the task.)

  Once he started investigating the biophysics of speech, Lieberman only became more intrigued. The revelation that really shaped his future career came to him one night in the bath. After finishing his Ph.D., he got a job at the University of Connecticut, and one evening after work he lay in the tub, listening to WGBH. The presenter remarked that apes couldn’t talk, and this struck him as worthy of investigation. Why not?

  Lieberman often traveled to New York to teach at Haskins Laboratories and started spending time at Brooklyn’s Prospect Park Zoo. When he took his tapes of hours and hours of ape vocalizations back to the lab to analyze, he found that apes do not make the full range of human sounds. This, he discovered, was because of the physiology of their tongues.

  The human tongue extends from the larynx, deep in the throat, to just behind the teeth. At points along its length it can change its shape. It can be moved up, down, forward, and back; it can be bunched up or extended, widened or curled. Whenever the tongue changes shape, the whole vocal tract is altered, and each different configuration results in a different sound. In contrast, the tongues of other apes lie mostly in their mouths. As a consequence, they don’t have the facility for generating as many specific sounds.

  Lieberman also realized that even though there weren’t as many sounds in the ape repertoire as in human speech, there were enough for the creatures to make a decent stab at talking. Chimpanzees can make m, b, p, n, d, t, and a number of vowel sounds. For a nonhuman, this is not bad. Few other animals can get close—if you could transplant a human brain into, say, a horse’s head, it would not be able to speak human language, because its mouth and tongue could never make the sounds we do.

  Where we differ from the chimpanzees is that they don’t selectively articulate these sounds and manipulate their sequence, as we do when, for example, we say “pie,” “my,” “buy,” “die,” “tie,” or “nigh.” It is as if they have the same vocal instrument—or at least one that is reasonably similar—but they just don’t use it in the same way.

  If it was not the actual range of sounds produced by our respective vocal tracts that enabled us to speak but prevented apes from doing so, then, thought Lieberman, we must differ in our ability to control those sounds. This realization launched him on a quest to determine the connections between motor control and the higher levels of language. He quickly came to the conclusion that in order to truly understand language, you have to begin with biology, and—he is very fond of quoting Theodosius Dobzhansky, a famous evolutionary biologist who died in 1975—“nothing in biology makes sense except in light of evolution.”

  Lieberman’s first book, The Biology and Evolution of Language, was published in 1984. In it he argued against the popular notion that there was a “linguistic saltation”—that is, no single dramatic event gave birth to human language. The Chomskyan idea of an ideal speaker and hearer confused the origins of language rather than illuminated them he said. Instead, he proposed: “Human syntactic ability, in [my] view, is a product of the Darwinian mechanism of preadaptation, the channeling of a facility that evolved for one function toward a different one.”1 He cited Darwin’s discussion of the evolution of lungs from swim bladders: “The illustration of the swim bladder in fish is a good one, because it shows us clearly the highly important fact that an organ originally constructed for one purpose, namely flotation, may be converted into one for a wholly different purpose—namely respiration.”2

  Lieberman was not arguing, as he was careful to explain, that there was no uniquely human specialization for syntax. Rather, his point was that in the brain there was an overlap between the parts that control bodily movements and the parts that allow us to order thoughts and words in cognition and speech. This physical overlap had come about because of the way we had evolved, he said, first developing the ability to physically move our bodies in space and then, overlaid upon that, developing the ability to move words in abstract patterns.

  All was peace and tranquility before the book, said Lieberman, but after its publication he and Chomsky fell out. For months, they argued back and forth, and then for the next eighteen years there was silence.

  In 1990 Lieberman was invited by Behavioral and Brain Sciences to contribute one of the comments on Pinker and Bloom’s paper. He wrote, “It is refreshing to see Pinker and Bloom adopting some of the major premises of my 1984 book: (a) that human linguistic ability evolved by means of Darwinian processes, (b) that the biological substrate for human linguistic ability is subject to the constraints of biology, in particular variation, and(c) that data from psycholinguistics, anthropology, neurophysiology, and so forth, are germane. However, Pinker and Bloom still carry much of the baggage of the MIT School of Linguistics, in particular that guiding principle ‘Not invented here.’”

  What he meant was that if research hadn’t been done at MIT, then, as far as MIT was concerned, it didn’t really exist. Clearly he was more annoyed than gratified.

  Even though Lieberman, Pinker, and Bloom were all writing about language evolution, and even though they all agreed that any analysis of language needed to take biology seriously, there was at least one fundamental difference in their goals. Pinker and Bloom believed that Darwinian evolution and Chomsky’s universal grammar were compatible, and sought to prove both Darwin and Chomsky right. Lieberman, on the other hand, believed the incongruity between slow evolutionary change and an innate language-specific organ was irresolvable. Pinker and Bloom’s argument that universal grammar should and could take account of genetic variation was not acceptable, he said. In order to explore language evolution, you have to completely abandon the idea that humans are born with some kind of grammar device. It just wasn’t possible for both Darwin and Chomsky to be right.

  What Chomsky had wrong about language, according to Lieberman, fell into a larger category of misunderstanding biology. Throughout history, he argued, the most complicated piece of current technology was often used as an analogy for the human body or brain. For example, in the eighteenth and nineteenth centuries the brain was often thought of as a clock or a timepiece. It was imagined to be a telephone exchange in the early twentieth century. And from the 1950s onward, the brain was seen as a digital computer.3

  These metaphors, Lieberman explained, often take on a life of their own. In the early nineteenth century, for example, physicians likened the body to a steam engine. When early steam engines became hot, they would explode, unless safety valves were used to release the intensely heated pressure inside. By analogy, doctors of the time bled patients who had a fever in the belief that releasing blood would lower the body’s temperature.4

  The human mind-brain implied by Chomsky’s theory of language, Lieberman argued, was fundamentally based on the architecture and processes of a computer. In a computer, the central processing unit is a discrete device that generates output by algorithms. Random-access memory
and hard drives are also modular mechanisms. The Chomskyan brain, similarly, has a localized language organ that generates syntax. Sound, structure, and meaning are constructed separately. And the language organ is separate from other parts of the brain, these parts also being separate from one another.

  According to Lieberman, the analogy between the computer and the brain prevents a true understanding of language. Even though formulas can describe a set of sentences, they don’t have much to do with how language is produced by the brain or how the brain and language evolved. “Syntax is not the touchstone of human language, and evolution is not logical,” declared Lieberman. “Evolution doesn’t give a damn about formal elegance.”5

  When Lieberman began working at Brown University in 1976, he turned his attention to the connection between higher levels of language and the motor system. He started with the basal ganglia. These neural structures, the striatum and the globus pallidus, lie beneath the cortex, the brain’s outermost rind. The basal ganglia are responsible for learning patterns of motor activity—playing tennis, dancing, picking up a cup of tea. They also control the way different physical movements or mental operations are ordered, one dance step after another, and they are crucial in responding to a change in the direction of movement or thought.

  Lieberman compared the basal ganglia of neurologically normal people with patients who had Parkinson’s disease. In Parkinson’s the brain progressively degenerates, and among the first and hardest-hit structures are the basal ganglia. The cortex is generally one of the last parts of the brain to be damaged, but when it is, the patient falls victim to dementia. People suffering from Parkinson’s have tremors and rigidity and repeated patterns of movement. What intrigued Lieberman about these people was that they also had trouble comprehending and producing syntax. In addition to showing their physical symptoms, they tended to produce sentences that were particularly short, with only simple syntax.

  Lieberman carried out a study of Parkinson’s patients in which they were asked to say “one,” “two,” or “three” in order to identify which of three pictures best corresponded to a sentence they had heard. People who are neurologically normal generally make no errors when taking this test, but a number of the Parkinson’s patients with damage to the basal ganglia struggled with sentences with slightly complicated syntax and with long, conjoined sentences of simple syntax.

  In another study many Parkinson’s patients were shown to have trouble if they first heard an active sentence (“The hawk ate the sparrow”) and then were asked a related question in the passive voice (“Who was the sparrow eaten by?”). They also had difficulty when the original sentence was passive and the subsequent question was active. The patients experienced no problems in working out the meaning of sentences; it was just the syntax that tripped them up.

  The fact that damage to a brain area that controlled motor skills also affected syntax was a smoking gun for a biological relationship between language and motor control. The basic idea, Lieberman argued, is that there is a dependent relationship “between the syntax of motor control and the syntax of language.”

  Interestingly, these findings overlapped with some of Steven Pinker’s experimental results. Even though the two researchers began with opposite ideas about language and the mind-brain, they agreed on the subject of the basal ganglia and syntax. “Lieberman long ago predicted that the basal ganglia should have an important role in syntax,” said Pinker. “And I found corroborative data that shows it.” He continued:

  A lot of my work on language uses a comparison between regular and irregular verbs as a way of tapping into the combinatorial, recursive part of language and the memory component of language. In particular, when we use “walked” as the past tense of “walk,” you don’t have to memorize that because you can just crank it out using the rule “add ‘ed’ to a verb.” Whereas if you use “broke” as the past tense of “break,” there you can’t use a rule, because there is no rule. You have “break/ broke,” but you have “take/took” and you have “fake/faked.” So that relies on memory.

  So comparing regular and irregular forms is a way of studying this recursive-combinatorial component in the simplest possible way—sticking an “ed” onto a verb is the smallest operation that anyone would be willing to call combinatorial or recursive grammar. The reason that the irregular is a nice comparison is that it doesn’t involve a recursive-combinatorial component, but it means the same thing. It’s just another way of expressing the past tense at the same length and same complexity.

  We found that patients with Parkinson’s disease have more trouble with regular than with irregular verbs, and they have more trouble with novel verbs. Like, when a new word enters the language, like, “to spam,” everyone knows that the past tense is “spammed.” I don’t think you’d look that up in a dictionary or memorize it, but you can just deduce it from your world of recursive grammar. That’s something that patients with Parkinson’s disease have more trouble with than irregular forms, and that fits into Lieberman’s theory that the basal ganglia are implicated in recursive syntax.

  Lieberman has gone on to explore the basal ganglia in a completely different group of subjects. Starting in 1993, he began to compare the linguistic and motor performance of Parkinson’s patients with that of individuals who were climbing Mount Everest. Both sets of people incur brain damage, specifically to the basal ganglia, though the basic cause is very different. Parkinson’s is a progressive and fatal disease, whereas the basal ganglia damage suffered by climbers on Everest results from the lack of oxygen. In most cases it is temporary. Nevertheless, the climbers exhibit a lot of the same deficits experienced by Parkinson’s patients.

  Lieberman set up a monitoring unit at Everest’s Base Camp, fifty-three hundred meters above sea level. His research team administered baseline cognitive tests to the climbers and took samples of their speech. As the climbers ascended the mountain and stopped at the next four camps, further tests and speech samples were obtained by radio link.

  One of the abilities that Lieberman examined was how the climbers assembled the bits that make up distinctive sounds of speech. For example, when you pronounce b, you must coordinate at least two movements. At some point, you open your lips and release air while simultaneously vibrating the vocal cords deep in your throat. Timing the onset of voicing in speech sounds is yet another complicated motor skill at which every normal speaker is expert, though few are consciously aware of it. It is also another kind of movement sequence that gets affected in Parkinson’s disease.

  For example, the only difference between a b and a p is that you vibrate your vocal cords much sooner for the former than for the latter. With a b, voicing occurs within twenty-five milliseconds of opening your lips; with a p, your vocal cords start vibrating more than twenty-five milliseconds after you open your lips. Because Parkinson’s patients experience a breakdown in the onset timing of voicing in speech sounds, some of their b’s sound like p’s, and vice versa. (The same applies to d and t and to g and k.) This deficit occurs alongside an increase in syntactic errors and a delay in the comprehension of simple sentences.

  Lieberman showed that the higher the climbers went up the mountain, the more trouble they had with the timing of their voicing and the more their comprehension of syntax degraded. The farther up they went, the less oxygen they breathed, and just like Parkinson’s patients, they became less adept at distinctly pronouncing sounds like b and p, and they took longer to understand test sentences.

  It’s clear from this evidence, according to Lieberman, that the basal ganglia are crucial in regulating speech and language, making the motor system one of the starting points for our ability not only to coordinate the larynx and lips in talking but to use abstract syntax to create meaningful and complicated expressions.

  One of the important functions of the basal ganglia is their ability to interrupt certain motor or thought sequences and switch to a different motor or thought sequence. Climbers on Everest become increasingly
inflexible in their thinking as they ascend the mountain—stories about bad decision making in adverse circumstances abound. Accordingly, Lieberman’s climbers showed basic trouble with their thinking.

  One mountaineer monitored by Lieberman scored well at base camp but demonstrated extreme anomalies in his speech and a dramatic decline in thinking as he ascended. The researchers told him that he wasn’t functioning normally and advised him to descend, but he refused, insisting he was fine. When the weather took a turn for the worse and his companions descended, he persevered in going forward. A few days later he fell to his death.

  It was later discovered that at the time of his death, a harness the climber needed to secure himself to fixed ropes was not properly attached. There was nothing wrong with the harness itself; the problem was in how it had been used. In order to secure the harness, a correct sequence of steps had to be carried out. It appears that the lack of oxygen supply to the basal ganglia affected the climber’s ability to follow the basic sequence of clipping and unclipping.

  Basal ganglia motor control is something we have in common with many, many animals. Millions of years ago, an animal that had basal ganglia and a motor system existed, and this creature is the ancestor of many different species alive today, including us. When we deploy syntax, Lieberman argued, we are using the neural bases for a system that evolved a long time ago for reasons other than stringing words together.