The science of life : fully illustrated in tone and line and including many diagrams, стр. 803

THE CORTEX AT WORK

giving food immediately afterwards. Now Suppose that the buzzer is sounded several times in succession (at intervals of a couple of minutes), but without the usual meal. At first the dog responds to the sound, but gradually the reactions become feebler and less definite, until finally, after half-a-dozen consecutive disappointments, they cease altogether. This is the process of extinction! But it is not, he found, a simple fading-away or rubbing-out of a trace left in the brain ; it is something more complicated than that. It is not even a permanent extinction. If the original conditioned reflex was firmly established the extinction is found to be merely temporary ; the conditioned reflex has not been destroyed at all, it has only been covered over by something else ; two or three days later (or even two or three hours, in some cases) the buzzer will call forth a full expectant response, in spite of the previous disappointment.

Actually, during the extinction, a process of active restraint was going on in the dog’s brain. One might imagine the animal was thinking, “ Hallo—what’s wrong ? The food isn’t turning up as it shouldbetter ignore the buzzer for a bit until things become normal again.’ But to do that is, of course, to read more lucidity into the animal’s mind than we have any right to presume. We shall find it more satisfactory to say that the conditioned reflex becomes inhibited, much as the movements of the bowels are inhibited when occasion demands. By counting the drops of saliva produced by the animal each time the buzzer sounds, the progress of such an inhibition can be measured, and a curve can be plotted showing its steady intensification. But there may be complications. First, if, after a conditioned reflex has been thus extinguished and allowed to recover, the experiment is repeated, one finds that the second time the extinction occurs, it takes place more rapidly than the first. What took half a dozen buzzings to accomplish may now take three or four. And a third extinction will be swifter still. An inhibition, therefore, like any other kind of mental activity, can be improved by practice. Moreover, if extinction is accomplished, he buzzings being repeated until there is 10 response, and then immediately after here is a return to the normal procedurei.e., a buzz followed by food—it will be found that the conditioned reflex is at once restored, and in the future, as in the past, yuzzing will produce copious salivation,

Clearly here is a very delicate system of adjustment between the organism and its circumstances. Its practical value is evident enough. Consider a wild dog born into the world with its few necessary primary and unconditioned reflexes and with its wide capacity for making associations and conditioning these reflexes. Such a dog in forming conditioned reflexes learns to anticipate and prepare for the special accidents of its individual life. Its behaviour adapts itself to the fact that in some special part of the forest there is plenty of food, while in another there are painful thorns; that men and danger are about at certain seasons or at certain times of day and not at others. But in a changing world there must obviously be some means of revising this knowledge. There is no point in haunting a warren for rabbits if it becomes deserted, or in shunning a human dwelling-place if it has ceased to be inhabited. A signal that has lost its meaning must not be responded to ; it must be taken out of current use 5 and it is by means of “ extinction,” the inhibition of reflexes that no longer “ work,” that the guiding collection of conditioned reflexes is overhauled and revised.

And now for another aspect of this same tidying-up, adjusting process in the mind. Imagine that a fresh dog is taken into the laboratory and put into one of the usual quiet experimental rooms. From time to time a tuning-fork is sounded, and whenever this happens food appears. Soon, as we have learnt, it forms a conditioned reflex and regards the sound of the tuning-fork as a food-signal. But what we have not yet discussed is that, other things being equal, the stimulus is at first “ generalized.” The stimulus is not very precisely apprehended. Any tuning-fork—or, indeed, any similar noise—will produce signs of foodexpectation, and the reaction will be the greater the more closely the noise resembles the original tuning-fork. But the animal can be taught to discriminate more precisely between the noises. If a number of different notes are sounded pretty frequently, and only one of them is followed up with food, the others soon cease to evoke any response at all. The animal’s brain finds out which one of the noises matters. Obviously the brain-process here is analogous to the extinction of a conditioned reflex, since it consists essentially in suppressing the responses to those stimuli which, though similar, have no consequence of importance. This is another manifestation of inhibition in the cerebral cortex. Quite similarly in

777