Solution Language Is Unique to Humans

... I learn'd the language of another world.

Lord Byron, Manfred, Act III, Scene 4

Ludwig Wittgenstein once famously remarked that "if a lion could talk, we would not understand him." It is easy to see the philosopher's reasoning: lions must perceive the world in ways quite alien to us. They possess drives and senses we simply do not share. on the other hand, the statement is all wrong. If a lion spoke English, then presumably English speakers could understand him — but the mind of that lion would no longer be a lion's mind. The lion would no longer be a lion. Humans talk; lions do not.230

Some people argue that humans are unique in being the only species in the history of Earth that has employed language. If language developed in only one species — just one out of the 50 billion species that have ever existed — then we might infer that the likelihood of language developing is small. Perhaps it developed in humans just through dumb luck — a chance assembly of several unlikely physical and cognitive adaptations. We are unique on Earth, and we may be unique in the whole Galaxy: perhaps humans are the only creatures that can talk. And since language opens up so many possibilities — so much of what we do individually and socially would not take place in the absence of language — creatures without language would surely be unable to build radio telescopes. No matter how intelligent those creatures might otherwise be, if they had no language, then we would not hear from them.231

Could this explain the Fermi paradox? Perhaps many planets are home to advanced life, but only here on Earth has a species learned to talk. At first glance it seems to be an outrageous suggestion, but it becomes more plausible upon closer perusal.

Noam Chomsky, one of the most profound thinkers of our age, has done more to elucidate the nature of human language than has anyone else.232 Chomsky argues that language is innate. A child does not learn language; rather, language grows in the child's mind. In other words, a child is genetically programmed with a blueprint — a set of process rules and simple procedures that make the acquisition of language inevitable. All of us have a "language organ" — not something you can cut out with a knife, but rather a set of connections in the brain dedicated to language in the same way that parts of the brain are dedicated to vision. In this view, language acquisition happens to a child in much the same way that body hair suddenly sprouts on a pubescent teenager; it is part of growing up. Language is part of our genetic heritage.

Although Chomsky's ideas have been attacked by both adherents to the Standard Social Science Model (who argue that human practices within a social group are moulded by the culture of the group) and philosophers (who argue with Chomsky on several grounds), his theory seems to be the only way to explain several puzzles regarding language acquisition.

For example, language is an infinite system. If I were to speak this present sentence out loud, then there is an excellent chance that I would be the first person ever to utter these particular words in this particular order; it is a unique combination. One can construct an infinite number of sentences from a finite number of words. In order to cope with this infinite set, the brain must be following rules rather than accessing a store of responses. And when one considers what a child hears when its parents and siblings talk to it — just a sequence of sounds, including meaningless "uh's," "huh's" and "coochy-coo's" interspersing the poorly formed and incomplete sentences we all inevitably utter — it is remarkable that children develop and employ complex grammars so rapidly (all without the benefit of training, and often without feedback on the errors they make). Remarkable, that is, unless children are innately equipped with a language acquisition device (lad) that lets them pluck the relevant syntactic patterns from the gobbledygook assaulting their ears. There is just one LAD, common to all humanity; there is not one device for Albanian, another for Basque, and yet another for Czech. Any child — so long as he or she receives sufficient stimuli to trigger the LAD at the correct age — can learn to speak any language. The stimulus need not even be auditory. If they are exposed to signing at the right age, hearing children of deaf parents can acquire sign language.

The operation of the human LAD may be similar to the innate visual acquisition device (vad) of many animals. Scientists have performed experiments on kittens, blindfolding them immediately after birth. If the blindfold is removed any time before the first 8 weeks, the normal devel opment of the kitten's visual system is resumed, and the adult cat will see normally. If the blindfold is kept on for longer than 8 weeks, the cat will suffer permanent visual impairment. It seems, therefore, that there is a critical period in which the VAD must receive external visual stimuli in order to establish the appropriate neuronal connections in specific pre-wired locations in the kitten's brain. If the connections are not established within this period, the chance of developing a fully functioning visual system is lost. Other parts of the brain cannot act as stand-ins for the visual system. The same effect has been observed to occur in those tragic cases in which linguistic input is withheld from children during the critical period up to puberty: their ability to speak grammatically is severely impaired. The existence of a critical period for language acquisition is not necessarily mysterious: it is presumably simply part of the same genetically controlled maturation process that causes our sucking reflex to disappear, our baby teeth to erupt, and all the other changes that occur to the human body. It makes evolutionary sense for the LAD to switch on early, as that way we have the maximum time to enjoy the considerable benefits of language. It also makes sense for the LAD to switch off when its job is done, since maintaining the device presumably incurs considerable costs in terms of energy requirements.

Although different languages differ in the specifics, there is a universality to language. And it is these universal principles that are innate. When a child develops language, then, the procedure follows an internal, predetermined course. A child who is acquiring Dutch will set the parameters of this predetermined system in one way; the child acquiring English will set the parameters in another way; and the child acquiring French will set the parameters of the system in yet another way. But the underlying principles are the same. To use a software analogy, language acquisition is rather like a macro with arguments — one argument for each language. (Vocabulary, of course, must be learned: if individual words were innate, then a neologism like "pulsar" would have to be assimilated into the gene pool before astronomers could use it! Cultural evolution would move at the same glacial pace as genetic evolution. Certain grammatical constructions must also be learned. For example, there is rule for forming the regular past tense of an English verb — namely, add -ed — but the past tense of irregular verbs must be learned on a case-by-case basis.)

In addition to evidence from linguistics and from the study of language acquisition in children, clinical evidence is at least consistent with the notion that language is innate. In some unfortunate patients, trauma or disease harms particular locations in the brain — locations that appear to be responsible for language processing. The effects can be distressing. For example, patients in which Wernicke's area is damaged find it diffi cult to comprehend the speech around them. More bizarrely, they suffer from Wernicke's aphasia: their speech is rapid, fluent, filled with grammatically correct phrases — yet their speech make no sense. They often substitute one word for another, and they coin new words; when asked to name objects, they give semantically related words or words that distort the sound of the correct word. Transcripts of their speech can make for disturbing reading — like reading the ramblings of a psychotic. On the other hand, patients with damage to Broca's area suffer from Broca's aphasia — speech that is slow, halting and ungrammatical. They can often comprehend the speech going on around them, or at least make informed guesses as to the meaning of speech, thanks to their prior knowledge of the world and the built-in redundancy of speech. (They can understand a sentence like "the cat chased the mouse" because they know cats chase mice.) Patients in which the connection between Wernicke's and Broca's areas is damaged suffer a form of aphasia that renders them incapable of repeating sentences. Even worse is the aphasia affecting patients in which Wernicke's and Broca's areas, and the connection between them, are undamaged but isolated from the rest of the cortex. The patients can repeat what they hear but have no understanding of what they are saying; they never initiate conversation. In yet other cases, damage to specific parts of the brain — often through stroke — causes remarkably specific language problems. Some aphasics can recognize colors but not name them; others cannot name food items, though they know what they like to eat; others cannot name items of clothing but have no trouble dressing themselves. At present, neuro-scientists cannot map the brain and highlight different areas as handling different aspects of language. However, the evidence is that language is localized. And although localization itself does not mean language is innate, it does suggest we have a language organ.233

If we possess an innate language faculty, then the obvious question is: how did we come by such an intricate and complex organ? The answer is equally obvious: it evolved by natural selection of heritable variations. Unless we invoke the involvement of a creator, natural selection is the only known process that can generate such wonderful structures. If our language organ is the result of evolution, though, should we not see traces of it in the apes? After all, we are descendants of apes, aren't we? Well, no, we are not. Humans and apes are linked by a common ancestor that perhaps lived as long ago as 7 million years. It is entirely possible that our LAD evolved some time within the last 7 million years, so that it is not shared with the evolutionary branch leading to modern apes. Indeed, it has been suggested that the minds of early modern humans of about 100,000 years ago contained several separate "modules": a module for language, a module for technical intelligence, a module for social intelligence, a module for natural history, and so on. It may be that these isolated modules began to communicate only 50,000 years ago; and only then could people get together in groups and discuss, for example, the merits of a new tool design for use in hunting. (Perhaps it was only then that human consciousness, as we now understand it, developed. Only then did we become fully human.)

The followers of Chomsky would argue that language is specific to the human species. If you want to understand other animals, study what they do best; but it is pointless studying their language capabilities, since language is a human-specific ability. Pigs do not fly; neither do they talk.

But are we sure we are unique? What about chimpanzees, or dolphins, or dancing bees — do they not communicate in their own way? Perhaps they have innate language abilities too. One of the difficulties in contemplating these questions is our language: we seem compelled to anthropomorphize. Even when describing inanimate objects we anthropomorphize: genes are "selfish," the car is "acting funny," my chess program is "figuring out" the best move to make. There is of course nothing wrong with employing metaphor — assigning intentionality to inanimate objects enables us to convey the appropriate thought quickly — but sometimes we can forget that anthropomorphic statements do not necessarily describe what is really happening. We have to be careful when describing an animal's actions in terms of our own conscious thoughts and motives. When we describe an animal as communicating some word or idea — effectively, when we say it is "talking" — we might be completely wrong.

Here is an example where our first interpretation of events may be wrong. Some types of ground squirrel living in open country suffer two main predators: hawks, relying on speed, attack from the air, while badgers, relying on stealth, attack from the ground. When a squirrel spots a predator it chooses (there is an anthropomorphic usage!) from one of two defensive strategies. If it spots a badger, then the squirrel retreats to the opening of its burrow and maintains an erect posture. A badger, seeing that posture, knows the squirrel has spotted it and thus an attack would be a waste of time and energy. If a squirrel spots a hawk, then it runs like hell for the nearest cover. Squirrels also emit two different alarm sounds. If they spot a badger, then they make a rough chattering sound; if they spot a hawk, then they emit a high-pitched whistle. Other squirrels in the vicinity react when they hear the sounds, retreating to their burrows when they hear the badger alarm or running for cover when they hear the hawk alarm. Our natural inclination is to think squirrels are communicating with each other; that they are saying in effect: "Careful, now, there's a badger around; better head for home" or "Oh-oh, hawk; get out of here!" But are they?

As its actions upon spotting a predator clearly show, any individual squirrel is interested in saving its own skin. Indeed, evolutionary theory tells us that this must be the case: a squirrel could not care less about the fate of other squirrels. But if the squirrel alarm calls carry semantic information — if they are calling out "brock!" or "hawk!" in squirrelese — we meet a paradox. Selection will favor those squirrels who keep quiet, sneak off silently, and let the other suckers get eaten; being a non-caller in a group of callers is selectively advantageous, and the squirrel gets to pass on its genes. Soon, though, you end up with a community of silent squirrels; where does the instinct to cry out arise? The behavior of the squirrels makes sense only if their calls do not convey semantic information. Consider the squirrel's "hawk alarm." First, it is a high-pitched whistle — which, as experiments have shown, hawks find difficult to locate. So the squirrel is revealing nothing to the hawk. Second, being the only one to run for cover makes a squirrel conspicuous; it is much better to be one of a group of squirrels that are scrambling around, because the chances of being singled out by the hawk are reduced. Similarly, squirrels that run for cover when they hear a high-pitched whistle are less likely to be eaten by a hawk than squirrels that stand their ground. So selection will tend to favor squirrels that cry out when they see a hawk, and also those that run for cover when they hear a high-pitched whistle. When humans look at the situation they interpret it as squirrels sharing information. But that is not what is happening. The behavior is simply a trait that is passed on through the generations because it is effective. The squirrels do not even have to be aware of one another for this sort of behavior to evolve. No words; no language; just the forces of evolution. A similar analysis can be applied to the famous case of the vervet monkeys, which have "alarm calls" for eagles, leopards and pythons.

But what about chimpanzees like Washoe and bonobos like Kanzi, which have been taught American Sign Language (asl)? Surely the achievements of these creatures proves that some animals have the capacity for language? Even here we must be careful. The team of scientists who trained Washoe for three years claimed at the end of the program that the chimpanzee could use 68 signs, and even string some of them together in two- and three-word sentences. Herbert Terrace, a scientist who was entranced with the idea of communicating with another species, sought to replicate the experiments. He raised the chimpanzee Nim Chimpsky (the reason for the name should be obvious!) in a highly social setting and taught it a set of ASL signs. Terrace videotaped the signing sessions and, after analyzing the data from these sessions, completed most of a book describing Nim's success in acquiring sign language. Then, when he replayed the tapes in slow motion for a final analysis, Terrace made a discovery:

nearly all of Nim's signs were prompted by its human teachers. Furthermore, the chimp's signs were often imitative of what its teachers had just signed. Nim was never spontaneous with its signing; the signs were made to obtain rewards from its teachers (and even then, it resorted to signs only after more direct methods of obtaining a reward had failed). In short, Nim did not display anything like full language. When scientists scrutinized the publicly available tapes of Washoe it was clear the same thing had happened: the chimp was imitating signs that a trainer had just made. Perhaps the strongest criticism of the Washoe experiment came from the one native ASL signer on the team. He recalled how the scientists would log as a sign every vague movement Washoe made, even though the gesture might resemble no valid ASL sign. The scientists' conclusion was a case of wishful thinking. In a similar case with a gorilla called Koko, its trainer explained away Koko's many mistakes by calling them metaphors and mischievous lies. If you take that approach to data analysis, you can find anything. Even in the case of Kanzi, an animal that undoubtedly displays impressive cognitive abilities, great care must be taken not to over-interpret what it does. No matter how generous we are, we simply cannot argue that Kanzi uses language in anything like the way humans use it.

Using a system of rewards to train captive chimpanzees is one thing, but what chimps do in the wild is something else. There is absolutely no indication that chimps — or indeed any other creatures — use language spontaneously. Many other pieces of evidence suggest that animals do not possess symbolic language. For example, in one recent experiment scientists released a dolphin into one end of a pool containing apparatus that (once the dolphin had figured out how it operated) released food. The investigators timed how long it took the dolphin to understand how the apparatus worked, then transferred the dolphin to the other end of the pool. A barrier prevented the dolphin from swimming back to the apparatus, but it could still see the apparatus and could send signals through the water. The scientists released a second dolphin into the pool near the apparatus. On average, the second dolphin took the same time to operate the apparatus as did the first dolphin. We can conclude from this that the first dolphin was unable to tell the second dolphin how the apparatus worked. Dolphins lack an abstract language. A similar experiment with chimpanzees had the same results: the chimps could not communicate their knowledge.

As a final piece of evidence that our relatives lack an innate capacity for language, consider what happens when scientists excise the areas of a monkey's brain corresponding to Broca's and Wernikce's areas in humans: the monkey's ability to produce or respond to vocal calls is unaffected.

Although the suggestion that only humans possess symbolic language may be controversial, many people (myself included, for what it is worth)

think it is self-evidently the case. Even if we can train certain animals to use words, no animals come close to using language in the abstract, spontaneous, playful, creative way that humans use language. It seems silly to deny the fact. It also seems arrogant and anthropocentric to measure the abilities of animals in terms of our capabilities. Birds can perform feats of navigation that no human can match without aids. Some marine animals can, unlike humans, sense electric currents. Dogs can hear sounds beyond our perception and smell scents to which our noses are dead. Bats use an incredible system of echolocation. Horses have been known to pick up on cues that humans miss completely. And so on. Every species has abilities, forged by evolution, which enable them to scrape a living in a world that cares not whether they survive. This diversity is wonderful, and should be celebrated. Defining other species in terms of how well or how badly they use human traits is to demean those species.

Articulate speech is vitally important to the success of our species. Perhaps it is impossible for any species to develop the ability to travel or communicate over interstellar distances if they lack some equally sophisticated method of communication. And yet, in the case of the evolution of human speech, we seem forced to conclude that articulate speech is the result of a series of chance environmental changes and evolutionary responses; it was just good luck. Consider, for example, what happened to the bodies of our ancestors: a restructuring of the human diaphragm, larynx, lips, nasal passages, oral cavity and tongue, all of which were vital for articulate speech to develop, but none of which occurred in order for speech to develop. The changes to these organs were initially completely unrelated to the capacity for speech; they were small changes that brought immediate selective benefits. At least one of the changes — the positioning of the larynx deep in the throat — seems bizarre. Having a larynx low in the throat provides the tongue with enough room to move and produce a large number of vowel sounds, but any food and drink we swallow has to pass over the trachea: choking to death becomes a distinct possibility. If the tape of life were replayed, perhaps humans would not develop language. The benefits are great, but so are the costs.

On Earth, of the 50 billion species that have existed, only humans possess language. Language enables us not only to think, but to think about the thoughts we have. It enables us to reflect upon our thoughts, to try out new patterns of thought, and to record our thoughts. Language is what makes us human. If we ever visit other worlds, perhaps we will find billions of other species — each well adapted to its particular niche, but none of them with the single adaptive trait we are searching for: language.

0 0

Post a comment