Monday, October 12, 2009

The Mind - Minding the Gap

One of our most basic tools, the No. 2 pencil, used by every test taker, illustrates the exceptional freedom of the human mind as compared with the limited scope of animal cognition. You hold the painted wood, write with the lead, and erase with the pink rubber held in place by a metal ring. Four different materials, each with a particular function, all wrapped up into a single tool. And although that tool was made for writing, it can also pin hair up into a bun, bookmark a page or stab an annoying insect. Animal tools, in contrast—such as the sticks chimps use to fish termites out from their mounds—are composed of a single material, designed for a single function and never used for other functions. None have the combinatorial properties of the pencil.

Another simple tool, the telescopic, collapsible cup found in many a camper’s gear, provides an example of recursion in action. To make this device, the manufacturer need only program a simple rule—add a segment of increasing size to the last segment—and repeat it until the desired size is reached. Humans use recursive operations such as this in virtually all aspects of mental life, from language, music and math to the generation of a limitless range of movements with our legs, hands and mouths. The only glimmerings of recursion in animals, however, have come from watching their motor systems in action.

All creatures are endowed with recursive motor machinery as part of their standard operating equipment. To walk, they put one foot in front of the other, over and over again. To eat, they may grasp an object and bring it to the mouth repeatedly until the stomach sends the signal to stop. In animal minds, this recursive system is locked away in the motor regions of the brain, closed off to other brain areas. Its existence suggests that a critical step in acquiring our own distinctive brand of thinking was not the evolution of recursion as a novel form of computation but the release of recursion from its motor prison to other domains of thought. How it was unlocked from this restrictive function links to one of our other ingredients—promiscuous interfaces—which I will turn to shortly.

The mental gap broadens when we compare human language with communication in other species. Like other animals, humans have a nonverbal communication system that conveys our emotions and motivations—the chortles and cries of little babies are part of this system. Humans are alone, however, in having a system of linguistic communication that is based on the manipulation of mental symbols, with each example of a symbol falling into a specific and abstract category such as noun, verb and adjective. Although some animals have sounds that appear to represent more than their emotions, conveying information about objects and events such as food, sex and predation, the range of such sounds pales in relation to our own, and none of them falls into the abstract categories that structure our linguistic expressions.

This claim requires clarification, because it often elicits extreme skepticism. You might think, for example, that animal vocabularies appear small because researchers studying their communications do not really understand what they are talking about. Although scientists have much to learn about animal vocalizations, and communication more generally, I think insufficient study is unlikely to explain the large gap.
Most vocal exchanges between animals consist of one grunt or coo or scream, with a single volley back. It is possible that animals pack a vast amount of information into a 500-millisecond grunt—perhaps equivalent to “Please groom my lower back now, and I will groom yours later.” But then why would we humans have developed such an arcane and highly verbose system if we could have solved it all with a grunt or two?

Furthermore, even if we grant that the honeybee’s waggle dance symbolically represents the delicious pollen located a mile north and that the putty-nosed monkey’s alarm calls symbolically represent different predators, these uses of symbols are unlike ours in five essential ways: they are triggered only by real objects or events, never imagined ones; they are restricted to the present; they are not part of a more abstract classification scheme, such as those that organize our words into nouns, verbs and adjectives; they are rarely combined with other symbols, and when they are, the combinations are limited to a string of two, with no rules; and they are fixed to particular contexts.

Human language is additionally remarkable— and entirely different from the communication systems of other animals—in that it operates equally well in the visual and auditory modes. If a songbird lost its voice and a honeybee its waggle, their communication would end. But when a human is deaf, sign language provides an equally expressive mode of communication that parallels its acoustic cousin in structural complexity.

Our linguistic knowledge, along with the computations it requires, also interacts with other domains of knowledge in fascinating ways that strikingly reflect our uniquely human ability to make promiscuous connections between systems of understanding. Consider the ability to quantify objects and events, a capacity that we share with other animals. A wide variety of species have at least two nonlinguistic abilities for counting. One is precise and limited to numbers less than four. The other is unlimited in scope, but it is approximate and limited to certain ratios for discrimination—an animal that can discriminate one from two, for instance, can also discriminate two from four, 16 from 32, and so on. The first system is anchored in a brain region involved in keeping track of individuals, whereas the second is anchored in brain regions that compute magnitudes.

Last year my colleagues and I described a third counting system in rhesus monkeys, one that may help us understand the origins of the human ability to mark the difference between singular and plural. This system operates when individuals see sets of objects presented at the same time—as opposed to individuals presented serially—and causes rhesus monkeys to discriminate one from many but not many from many food items. In our experiment, we showed a rhesus monkey one apple and placed it in a box. We then showed the same monkey five apples and placed all five at once into a second box. Given a choice the monkey consistently picked the second box with five apples. Then we put two apples in one box and five into the other. This time the monkey did not show a consistent preference. We humans do essentially the same thing when we say “one apple” and “two, five or 100 apples.”

But something peculiar happens when the human linguistic system connects up with this more ancient conceptual system. To see how, try this exercise: for the numbers 0, 0.2 and –5, add the most appropriate word: “apple” or “apples.” If you are like most native English speakers, including young children, you selected “apples.” In fact, you would select “apples” for “1.0.” If you are surprised, good, you should be. This is not a rule we learned in grammar school—in fact, strictly speaking, it is not grammatically correct. But it is part of the universal grammar that we alone are born with. The rule is simple but abstract: anything that is not “1” is pluralized.

The apple example demonstrates how different systems—syntax and concepts of sets—interact to produce new ways of thinking about or conceptualizing the world. But the creative process in humans does not stop here. We apply our language and number systems to cases of morality (saving five people is better than saving one), economics (if I am giving $10 and offer you $1, that seems unfair, and you will reject the dollar), and taboo trade-offs (in the U.S., selling our children, even for lots of money, is not kosher).

Source of Information : Scientific American September 2009

No comments: