In his book The Language Instinct (1994) Steven Pinker pointed out two fundamental facts about human language that were used by linguist Noam Chomsky to develop his theory about how we learn language. The first is that each one of us is capable of producing brand new sentences never before uttered in the history of the universe. This means that:

[A] language cannot be a repertoire of responses; the brain must contain a recipe or program that can build an unlimited set of sentences out of a finite list of words. That program may be called a mental grammar (not to be confused with pedagogical or stylistic “grammars,” which are just guides to the etiquette of written prose.)

The second fundamental fact is that children develop these complex grammars rapidly and without formal instruction and grow up to give consistent interpretations to novel sentence constructions that they have never before encountered. Therefore, [Chomsky] argued, children must be innately equipped with a plan common to the grammars of all languages, a Universal Grammar, that tells them how to distill the syntactic patters out of speech of their parents. (Pinker, p. 9)

Children have the ability to produce much greater language output than they receive as input but it is not done idiosyncratically. The language they produce follows the same generalized grammatical rules as others. This leads Chomsky to conclude that (quoted in Pinker, p. 10):

The language each person acquires is a rich and complex construction hopelessly underdetermined by the fragmentary evidence available [to the child]. Nevertheless individuals in a speech community have developed essentially the same language. This fact can be explained only on the assumption that these individuals employ highly restrictive principles that guide the construction of grammar.

The more we understand how human language works, the more we begin to realize how different human speech is from the communication systems of other animals.

Language is obviously as different from other animals’ communication systems as the elephant’s truck is different from other animals’ nostrils. Nonhuman communication systems are based on one of three designs: a finite repertory of calls (one for warnings of predators, one for claims of territory, and so on), a continuous analog signal that registers the magnitude of some state (the livelier the dance of the bee, the richer the food source that it is telling its hivemates about), or a series of random variations on a theme (a birdsong repeated with a new twist each time: Charlie Parker with feathers). As we have seen, human language has a very different design. The discrete combinatorial system called “grammar” makes human language infinite (there is no limit to the number of complex words or sentence in a language), digital (this infinity is achieved by rearranging discrete elements in particular orders and combinations, not by varying some signal along a continuum like the mercury in a thermometer), and compositional (each of the finite combinations has a different meaning predictable from the meanings of its parts and the rules and principles arranging them). (Pinker, p. 342)

This difference between human and nonhuman communication is also reflected in the role that different parts of the brain plays in language as opposed to other forms of vocalization.

Even the seat of human language in the brain is special. The vocal calls of primates are controlled not by their cerebral cortex but by phylogenetically older neural structures in the brain stem and limbic systems, structures that are heavily involved in emotion. Human vocalizations other than language, like sobbing, laughing, moaning, and shouting in pain, are also controlled subcortically. Subcortical structures even control the swearing that follows the arrival of a hammer on a thumb, that emerges as an involuntary tic in Tourette’s syndrome, and that can survive as Broca’s aphasic’s only speech. Genuine language . . . is seated in the cerebral cortex, primarily in the left perisylvian region. (Pinker, p. 342)

Rather than view the different forms of communication found in animals as a hierarchy, it is better to view them as adaptations that arose from the necessity to occupy certain evolutionary niches. Chimpanzees did not develop the language ability because they did not need to. Their lifestyles did not require the ability. Humans, on the other hand, even in the hunter-gatherer stage, would have benefited enormously from being able to share kind of detailed information about plants and animals and the like, and thus there could have been an evolutionary pressure that drove the development of language.

Human language was related to the evolution of the physical apparatus that enabled complex sound production along with the associated brain adaptations, though the causal links between them is not fully understood. Did the brain increase in size to cope with rising language ability or did the increasing use of language drive brain development? We really don’t know yet.

The argument against a linguistic hierarchy in animals can be seen in the fact that different aspects of language can be found to be best developed in different animals.

The most receptive trainee for an artificial language with a syntax and semantics has been a parrot; the species with the best claim to recursive structure in its signaling has been the starling; the best vocal imitators are birds and dolphins; and when it comes to reading human intentions, chimps are bested by man’s best friend, Canis familiaris. (Pinker, PS20)

It seems clear that we are unlikely to ever fully communicate with other species the way we do with each other. But the inability of other animals to speak the way we do is no more a sign of their evolutionary backwardness than our nose’s lack of versatility compared to the elephant’s trunk, or our inability to use our hands to fly the way bats can, are signs that we are evolutionarily inferior compared to them

We just occupy different end points on the evolutionary bush.

POST SCRIPT: But isn’t everyone deeply interested in golf?

If you want yet more reasons why TV news is not worth watching . . .

One of the most interesting questions in language is whether animals can talk or at least be taught to talk. Clearly animals can communicate in some rudimentary ways, some more so than others. Some researchers are convinced that animals can talk and have spent considerable efforts to try and do so but with very limited results. In the comments to an earlier post, Greg referred to the efforts by Sue Savage-Rumbaugh (and Duane Rumbaugh) to train the bonobo chimpanzee Kanzi to speak, and Lenen referred to the development of spontaneous language in children who had been kept in a dungeon. There have been other attempts with chimps and gorillas named Washoe, Koko, Lana, and Sarah.

One thing that is clear is that humans seem to have an instinctive ability to create and use language. By instinctive, I mean that evolution has produced in us the kinds of bodies and brains that make learning language easy, especially at a young age. It is argued that all humans are born possessing the neural wiring that contains the rules for a universal grammar. The five thousand different languages that exist today, although seeming to differ widely, all have an underlying grammatical similarity that is suggestive of this fact. For example, this grammar affects things like the subject-verb-object ordering in sentences. In English, we would say “I went home” (subject-verb-object) while in Tamil it would be “I home went” (subject-object-verb).

What is interesting is that of all the grammars that are theoretically possible, only a very limited set is actually found in existence. We do not find, for example, languages where people say “Home went I” (object-verb-subject). What early exposure to language does is turn certain switches on and off in the universal grammar wiring in our brains, so that we end up using the particular form of grammar of the community we grow up in. This suggests that language structures are restricted and not infinitely flexible, indicating a biological limitation.

The instinctive nature of language can be seen in a natural experiment that occurred in Nicaragua. There used to be no sign language at all in that country because the children were isolated from one another. When the Sandinistas took over in 1979, they created schools for the deaf. Their efforts to formally teach the children lip reading and speech failed dismally. But because the deaf children were now thrown together in the school buses and playgrounds, the children spontaneously developed their own sign language that developed and grew more sophisticated and is now officially a language that follows the same underlying grammatical rules as other spoken and sign languages. (Steven Pinker, The Language Instinct, 1994, p. 24)

What about animals? Many of us, especially those of us who have pets, would love to think that animals can communicate. As a result, we are far more credulous than we should be of claims (reported in the media) by researchers that they have taught animals to speak. But others, like linguist Steven Pinker, are highly skeptical. When looked at closely, the more spectacular elements of the claims disappear, leaving just rudimentary communication using symbols. The idea that some chimps can be taught to identify and use some symbols or follow some simple spoken commands does not imply that they possess underlying language abilities comparable to humans. The suggestion that animals use sign ‘language’ mistakenly conflates the sophisticated and complex grammatical structures of American Sign Language and other sign languages with that of a few suggestive gestures.

The belief that animals can, or should be able to, communicate using language seems to stem from two sources. One lies in a mistaken image of evolution as a linear process in which existing life forms can be arranged from lower to higher and more evolved forms. One sees this in posters in which evolution is shown as a sequence: amoebas→ sponges→ jellyfish→ flatworms→ trout→ frogs→ lizards→ dinosaurs→ anteaters→ monkeys→ chimpanzees→ Homo sapiens. (Pinker, p. 352) In this model, humans are the most evolved and it makes sense to think that perhaps chimpanzees have a slightly less evolved linguistic ability than we do but that it can be nudged along with some human help. Some people are also convinced that to think that animals cannot speak is a sign of a deplorable species superiority on our part.

But that linear model of evolution is wrong. Evolution is a branching theory, more like a spreading bush. Starting from some primitive form, it branched out into other forms, and these in turn branched out into yet more forms and so on, until we had a vast number of branches at the periphery. All the species I listed in the previous paragraph are like the tips of the twigs on the canopy of the bush, except that some (like the dinosaurs) are now extinct. Although all existing species have evolved from some earlier and more primitive forms, none of the existing species is more evolved than any other. All existing species have the same evolutionary status. They are merely different.

In the bush image, it is perfectly reasonable to suppose that one branch (species) may possess a unique feature (speech) that is not possessed by the others, just like the elephant possesses a highly useful organ (the trunk) possessed by no other species. All that this signifies is that that feature evolved after that branch separated from the rest of the bush and hence is not shared by others. The fact that nonhuman animals cannot speak despite extensive efforts at tutoring them is not a sign that they are somehow inferior or less evolved than us.

Some efforts to teach animals language skills seem to stem from a sense of misguided solidarity. It is as if the more features we share with animals, the closer we feel we are to them and the better we are likely to treat them. It is undoubtedly true that the closer we identify with some other living thing, the more empathy we have for it. But the solution to that is to have empathy for all living creatures, and not try to convince ourselves that we are alike in some specific ways.

As Pinker says:

What an irony it is that the supposed attempt to bring Homo sapiens down a few notches in the natural order has taken the form of us humans hectoring another species into emulating our instinctive form of communication, or some artificial form we have invented, as if that were a measure of biological worth. The chimpanzees’ resistance is no shame to them; a human would surely do no better if trained to hoot and shriek like a chimp, a symmetrical project that makes about as much scientific sense. In fact, the idea that some species needs our intervention before its members can display a useful skill, like some bird that could not fly until given a human education, is far from humble! (p. 351)

While any animal lover would dearly love to think that they can talk with animals, we may have to reconcile ourselves to the fact that it just cannot happen, because they lack the physical and perhaps cognitive apparatus to do so.

Next: The differences between animal and human communication.

POST SCRIPT: Superstitions

One of the negative consequences of religious beliefs is that it leads to more general magical thinking, one form of which is superstitions. Steve Benen lists all the superstitions that John McCain believes in.

It bothers me when political leaders are superstitious. Decision-makers should not be influenced by factors that have no bearing whatsoever on events.

Trying to discover the origins of language is a fascinating scientific problem but the evidence is necessarily indirect. Clearly our bodies’ physical capacity to articulate sounds is a biological development. Language had to be preceded by the evolution of the physical organs responsible for vocalization. Those organs must have co-evolved with those parts of the brain that can process language. But this evolutionary history is hard to reconstruct since the voice organs and brains are made of soft tissue and are thus unlikely to fossilize. Even if we could get an accurate fix on when the actual physical ability to speak came into being, this the could only be used to set a limit on the earliest time at which language could have occurred, but tells us nothing of when it actually did.

Since humans have these language organs and our closest existing cousins the chimpanzees do not, and since our branch of mammals split off from chimpanzees about 5-7 million years (or about 350,000 generations) ago, it is theoretically possible for language to be that old and still be consistent with only humans being able to speak.

At the other end, the discovery of cave art in Europe consisting of depictions of animals and humans in carved and painted and sculpted forms by Cro-Magnon humans in the Upper Paleolithic era about 35,000 years ago indicate complex social thinking indicative of the presence of language, suggesting that this sets a limit on the latest time for the origin of language.

But 35,000 to 5-7 million years is a huge time interval and attempts have been made to get a more precise fix on the origin of language. Various approaches have been attempted. One avenue of exploration comes from linguistics: the study of languages themselves and how they evolved. Another is to look at the physiological development of the human body. A third method is to look at the development of lifestyles to discern levels of complexity that suggest the kinds of social organization that would require language. A fourth is to look at the use of tools, to see if there is sophistication and uniformity over a wide area suggesting that knowledge was being shared and transmitted to distant locales.
While these are all promising avenues of research, unfortunately the lines of evidence from these different approaches currently do not converge on a single time, suggesting that we still have a long way to go in determining when language might have arisen.

Starting with linguistics, it is known that the structure of languages is very analogous to the biological tree of living organisms. Just as the fossil and DNA evidence all point to all living things being descended from a common ancestor, the approximately five thousand languages that currently exist exhibit grammar and vocabulary relationships strongly suggestive of the fact that they are all derived from a single common proto-language that existed long ago that evolved and split into branches the way that living organisms did. By tracing that linguistic tree back in time, we may be able to fix narrower bounds on the date of origin of that proto-language.

Steven Pinker argues that since modern humans Homo sapiens first appeared about 200,000 years ago and spread out of Africa about 100,000 years ago, and since all modern humans have identical language abilities along with a universal grammar, it seems likely that language appeared concurrently with the first appearance of modern humans. (Steven Pinker, The Language Instinct, 1994, p. 363, 364) Furthermore, there was a more than a tripling of brain size (from 400cc to 1350cc) during the period between the first appearance the genus Homo (in the form of Homo habilis) about two million years ago until Homo sapiens appeared, suggesting that the brain developed in that period partly in order to accommodate the new language centers. Pinker suggests that since Homo sapiens are us, it seems reasonable that language came into being as long ago as 200,000 years ago.

As for biological development. Richard Leakey explains what it is about the human body that enables speech. (The Origin of Humankind, 1994)

Humans are able to make a wide range of sounds because the larynx is situated low in the throat, thus creating a large sound-chamber, the pharynx, above the vocal chords . . . the expanded pharynx is the key to producing fully articulate speech . . . In all mammals except humans the larynx is high in the throat, which allows the animal to breathe and drink at the same time. As a corollary, the small pharyngeal cavity limits the range of sounds that can be produced. . . Although the low position of the larynx allows human to produce a greater range of sounds, it also means that we cannot drink and breathe simultaneously. We exhibit the dubious liability for choking.

Human babies are born with the larynx high in the throat, like typical mammals, and can simultaneously breathe and drink, as they must during nursing. After about eighteen month, the larynx begins to migrate down the throat, reaching the adult position when the child is about fourteen months old. (p. 130)

The unique position of the larynx in human speech suggests that if were able to identify when it got lowered to its present position, we might be able to determine when we first had the ability to speak. But the problem is that those parts of the body are made of soft tissues and do not fossilize easily. However, the shape of the bottom of the skull called the basicranium is arched for humans and essentially flat for other mammals and this part of the skull is an indicator of how well it can articulate sounds. “The earliest time in the fossil record that you find a fully flexed basicranium is about 300,000 to 400,000 years ago, in what people call archaic Homo sapiens.” (Leakey, p. 132)

But of course that does not mean that language developed simultaneously with the basicranium. Leakey says that it is unlikely that language was fully developed among archaic Homo sapiens.

The brain is another indicator of possible language origins. The part of the brain known as Broca’s area is a raised lump near the left temple associated with language and the use of tools. Furthermore, the left hemisphere of the brain (which is associated with language) is larger than the right. So if we can find fossilized skulls that indicate the presence of either of these features, that would also indicate the onset of possible linguistic ability. A fossil found nearly two million years ago seems to have just such features. Combined with the discovery of tool-making around this time Leakey thinks it is possible that it was with the advent of Homo habilis (the handyman) about two million years ago that language first started to appear, at least in a very crude form. (Leakey, p.129)

Another strategy is to look at the various tools and other artifacts that humans created and see if there is an increase in sophistication and increased spread of similar designs, which would suggest the sharing of knowledge and ideas and thus speech. The more complex the social structures in which people lived, the greater the need for language. As for tools, although they started being made about two million years ago, the earliest kinds were opportunistic in nature. More conscious tool making began about 250,000 years ago but then stayed static for about 200,000 years. The kinds of ordering of tools that are really suggestive of language does not seem to occur until suddenly about 35,000 years ago, coinciding with the sudden spurt in cave art in the Upper Paleolithic period. (Leakey, p. 134)

So basically the situation is confused. While it is possible that language began to appear in some primitive form as early as two million years ago, it seems more likely that real language skills began about 200,000 years ago. Also it is not clear whether language evolved gradually since that time or whether it remained in a low and more-or-less static state before suddenly exploding about 35,000 years ago into the complex language structures that we now have.

Next: Can animals talk?

POST SCRIPT: Fred and Wilma? Who knew?

The most unforgettable act of the 1969 Woodstock festival was Joe Cocker’s rendering of the Beatles’ A little help from my friends, a gentle song sung by Ringo Starr, which Cocker turned into an over-the top, weird, air-guitar-playing, frenzied, incoherent performance that looked like he was having some kind of seizure. Throughout it, you kept wondering what the hell he was singing since the lyrics seemed to have only a passing resemblance to the original.

Some helpful soul has now provided captions for Cocker’s words. It all makes sense now. Or maybe not.

(Thanks to Jesus’s General.)