Source: Gould, James L. and Peter Marler. Learning by instinct. Scientific American, January 1987. Reprinted in William S.-Y. Wang, ed.. 1991. The emergence of language. Development and evolution. New York: Freeman. 88–103.
"Learning is often thought of as the alternative to instinct, which is the information passed genetically from one generation to the next. Most of us think the ability to learn is the hallmark of intelligence. The difference between learning and instinct is said to distinguish human beings from "lower" animals such as insects. Introspection, that deceptively convincing authority, leads one to conclude that learning, unlike instinct, usually involves conscious decisions concerning when and what to learn.

"Work done in the past few decades has shown that such a sharp distinction between instinct and learning—and between the guiding forces underlying human and animal behavior—cannot be made. For example, it has been found that many insects are prodigious learners. Conversely, we now know that the process of learning in higher animals, as well as in insects, is often innately guided, that is, guided by information inherent in the genetic makeup of the animal. In other words, the process of learning itself is often controlled by instinct. …

"It now seems that many, if not most, animals are "preprogrammed" to learn particular things and to learn them in particular ways. In evolutionary terms innately guided learning makes sense: very often it is easy to specify in advance the general characteristics of the things an animal should be able to learn, even when the details cannot be specified. For example, bees should be inherently suited to learning the shapes of various flowers, but it would be impossible to equip each bee at birth with a field guide to all the flowers it might visit.

"Innately guided learning—learning by instinct—is found at all levels of mental complexity in the animal kingdom. … There is strong evidence, for example, that the process of learning human speech is largely guided by innate abilities and tendencies.

Two theoretical frameworks

"The distinction often made between learning and instinct is exemplified by two theoretical approaches to the study of behavior; ethology and behaviorist psychology. Ethology is usually thought of as the study of instinct. …

"Classical behaviorist psychologists see the world quite differently from ethologists. Behaviorists are primarily interested in the study of learning under strictly controlled conditions and have traditionally treated instinct as irrelevant to learning. Behaviorists believe nearly all the responses of higher animals can be divided into two kinds of learning called classical conditioning and operant conditioning.

"Classical conditioning was discovered in dogs by the Russian physiologist Ivan P. Pavlov. In his classic experiment he showed that if a bell is rung consistently just before food is offered to a dog, eventually the dog will learn to salivate at the sound of the bell. The important factors in classical conditioning are the unconditioned stimulus (the innately recognized cue, … which in this case is food), the unconditioned response (the innately triggered behavioral act, … which in this case is salivation) and the conditioned stimulus (the stimulus the animal is conditioned to respond to, which in this case is the bell. Early behaviorists believed any stimulus an animal was physically capable of sensing could be linked, as a conditioned stimulus, to any unconditioned response.

"In operant conditioning, the other major category of learning recognized by most behaviorists, animals learn a behavior pattern as the result of trial-and-error experimentation they undertake in order to obtain a reward or avoid a punishment. In the classic example a rat is trained to press a lever to obtain food. The experimenter shapes the behavior by rewarding the rat at first for even partial performance of the desired response. For example, at the outset the rat might be rewarded simply for facing the end of the cage in which the lever sits. Later the experimenter requires increasingly precise bahavior, until the response is perfected. Early behaviorists thought any behavior an animal was physically capable of performing could be taught, by means of operant conditioning, as a response to any cue or situation.

Challenges to behaviorism

"By 1970 several disturbing challenges to the behavioristic world view had appeared. The idea that any perceptible cue could be taught, by classical conditioning, as a conditioned stimulus was dealt a severe blow by John Garcia, now at the University of California at Los Angeles. He showed that rats could not associate visual and auditory cues with food that made them ill, even though they could associate olfactory cues with such food. On the other hand, he found that quail could associate not auditory or olfactory cues but visual ones—colors—with dangerous foods. Later work by other investigators extended these results, showing, for example, that pigeons readily learn to associate sounds but not colors with danger and colors but not sounds with food. The obvious conclusion was that these animals are predisposed to make certain associations more easily in some situations than in others.

"The same kind of pattern was discovered in experiments in operant conditioning. Rats readily learn to press a bar for food, but they cannot learn to press a bar in order to avoid an electric shock. Conversely, they can learn to jump in order to avoid a shock but not in order to obtain food. Similarly, pigeons easily learn to peck at a spot for a food reward but have great difficulty learning to hop on a treadle for food; they learn to avoid shock by hopping on a treadle but not by pecking. Once again it seems that in certain behavioral situations animals are innately prepared to learn some things more readily than others.

"The associations that are most easily learned have an adaptive logic. In the natural world odor is a more reliable indicator than color for rats (which are notoriously nocturnal) trying to identify dangerous food; the color of a seed is a more useful thing for a pigeon to remember than any sounds the seed makes. Similarly, a pigeon is more likely to learn how to eat novel seeds if it experiments on food with its beak rather than with its feet. Animals that have innate biases concerning which cues they rely on and which procedures they attempt are more likely to ignore spurious cues, and they will learn faster than animals without inherent biases. The idea that animals are innately programed to attend to specific cues in specific behavioral contexts and to experiment in particular ways in other contexts suggest a mutually reinforcing relation between learning and instinct. This relation helps … to reconcile the approaches of behaviorists and ethologists."