What You Can Change _. And What You Can't - Martin E. Seligman [37]
The same irrationality is true of Garcia’s phenomenon. Consciousness, the whole elaborate apparatus that dominates much of our life, is irrelevant to learning a taste aversion. In one experiment, rats tasted saccharin and were then made unconscious with anesthesia. While asleep, brief stomach illness was induced. After the rats woke up, they hated saccharin.4
The last problem was that Pavlovian conditioning extinguishes easily, a side effect of its being a conscious expectation. All Pavlov had to do was show up but not feed his dogs a few times, and they stopped salivating when they saw him. But my dislike of sauce béarnaise lasted a decade, and remained alive through a dozen dinner parties in which skeptical psychologists gave me sauce béarnaise to sample. I can eat it now, twenty-five years later, but still, I had to change from Craig Claiborne’s recipe to Julia Child’s.
The message from Garcia. All these problems can be reconciled with Pavlovian conditioning, because evolution is at work. For millions of years our ancestors repeatedly encountered, among their other woes, stomach illness. Plants and animals they ate sometimes contained poison, their food sometimes spoiled, the streams they drank out of sometimes carried germs. Each of these challenges could have killed them; but if they got sick and survived, they had to learn to avoid the toxin in the future.
The problem is: What to avoid? Toxins in nature are avoidable only because they have distinctive tastes, as do the foods that carry them. Those of our ancestors who after stomach illness learned rapidly and well to hate the most distinctive taste that accompanied sickness passed on their genes. Those who didn’t died out.
We are prepared by our evolutionary history to learn some things well. We learn, as I did, that sauce béarnaise goes with illness—in one trial and across long delays. This learning occurs at levels deeper and less fallible than rationality, is very strong, and is illogical. It took years for me to start liking sauce béarnaise again.
We do not “know” any of these things when we are born—they are not “instinctive.” Garcia’s rats loved saccharin, and I loved sauce béarnaise the first time I tasted it. If we have but a single aversive experience with prepared associations, however, we absorb it immediately and will not easily let it go. Evolution has shaped our sensory apparatus, and it has shaped our response system. A keen sense of taste and the response of retching are both the product of natural selection. We have known this ever since Darwin. What Garcia told us that was new is that how quickly we learn something or how slowly we learn it—or whether we learn it at all—is subject to natural selection. Learnability itself is shaped by evolution.
In response to Garcia’s findings, a long and bitter quarrel ensued in the journals, perhaps the bitterest in the history of learning theory. It is still not over. Most psychologists now accept the idea that our genes constrain learning, but for the staunch hard core, this is still heresy, for two reasons: The first is a special-interest-group concern, and the second is a fundamental concern.
The guild of behaviorists had by 1965 painted itself into a corner. We did experiments on pigeons pecking disks for grain, on rats pressing bars for Purina pellets, on rats becoming afraid of tones, and very little else. But we dubbed our findings, grandly, “the laws of learning”—not “the laws of hungry pigeon learning” or “the laws of rat learning.” We declared our laws to be as universal as Newton’s gravity. Garcia, very much not a member of the guild, told us that the kind of stimulus (taste) and its evolutionary relation to the response (stomach illness) mattered a lot. This, he contended, produced a violation of the laws: one trial learning over long delays. This meant that all the years that B. F. Skinner, Clark Hull, Ivan Pavlov,