Fig. 1.
What, now, would we expect to observe if a reasoning animal, who is surely eager to get out, is put, for example, into a box with a door arranged so as to fall open when a wooden button holding it at the top (on the inside) is turned from its vertical to a horizontal position? We should expect that he would first try to claw the whole box apart or to crawl out between the bars. He would soon realize the futility of this and stop to consider. He might then think of the button as being the vital point, or of having seen doors open when buttons were turned. He might then poke or claw it around. If after he had eaten the bit of fish outside he was immediately put in the box again he ought to remember what he had done before, and at once attack the button, and so ever after. It might very well be that he would not, when in the box for the first time, be able to reason out the way to escape. But suppose that, in clawing, biting, trying to crawl through holes, etc., he happened to turn the button and so escape. He ought, then, if at once put in again, this time to perform deliberately the act which he had in the first trial hit upon accidentally. This one would expect to see if the animal did reason. What do we really see?
To save time we may confine ourselves to a description of the twelve cats experimented with, adding now that the dogs presented no difference in behavior which would modify our conclusions. The behavior of all but No. 11 and No. 13 was practically the same. When put into the box the cat would show evident signs of discomfort and of an impulse to escape from confinement. It tries to squeeze through any opening; it claws and bites at the bars; it thrusts its paws out through any opening, and claws at everything it reaches; it continues its efforts when it strikes anything loose and shaky; it may claw at things in the box. The vigor with which it struggles is extraordinary. For eight or ten minutes it will claw and bite and squeeze incessantly. With No. 13, an old cat, and No. 11, an uncommonly sluggish cat, the behavior was different. They did not struggle vigorously or continually. (In the experiments it was found that these two would stay quietly in the box for hours, and I therefore let them out myself a few times, so that they might associate the fact of being outside with the fact of eating, and so desire to escape. When this was done, they tried to get out like the rest.) In all cases the instinctive struggle is likely to succeed in leading the cat accidentally to turn the button and so escape, for the cat claws and bites all over the box. These general clawings, bitings, and squeezings are of course instinctive, not premeditated. The cats will do the same if in a box with absolutely no chance for escape, or in a basket without even an opening—will do them, that is, when they are the foolishest things to do. The cats do these acts for just the same reason that they suck when young, propagate when older, or eat meat when they smell it.
Each of the twelve cats was tried in a number of different boxes, and in no case did I see anything that even looked like thoughtful contemplation of the situation or deliberation over possible ways of winning freedom. Furthermore, in every case any cat who had thus accidentally hit upon the proper act was, after he had eaten the bit of fish outside, immediately put back into the box. Did he then think of how he had got out before, and at once or after a time of thinking repeat the act? By no means. He bursts out into the same instinctive activities as before, and may even fail this time to get out at all, or until a much longer period of miscellaneous scrabbling at last happens to include the particular clawing or poking which works the mechanism. If one repeats the process, keeps putting the cat back into the box after each success, the amount of the useless action gradually decreases, the right movement is made sooner and sooner, until finally it is done as soon as the cat is put in.
This sort of a history is not the history of a reasoning animal. It is the history of an animal who meets a certain situation with a lot of instinctive acts. Included without design among these acts is one which brings freedom and food. The pleasurable result of this one gradually stamps it in in connection with the situation "confinement in that box," while their failure to result in any pleasure gradually stamps out all the useless bitings, clawings, and squeezings. Thus, little by little, the one act becomes more and more likely to be done in that situation, while the others slowly vanish. This history represents the wearing smooth of a path in the brain, not the decisions of a rational consciousness.
Fig. 2.
Fig. 3.
We can express graphically the difference between the conduct of a reasoning animal and that of these dogs and cats by means of a time-curve. If, for instance, we let perpendiculars to a horizontal line represent each one trial in the box, and let their heights represent in each trial the time it took the animal to escape (each three millimetres equaling ten seconds), the accompanying figure (Fig. 2) will tell the story of a cat which, when first put in, took sixty seconds to get out; in the second trial, eighty; in the third, fifty; in the fourth, sixty; in the fifth, fifty; in the sixth, forty, etc. This figure represents what did actually happen with one cat in learning a very easy act. Suppose the cat had, after the third accidental success, been able to reason. She would then have the next time and in all succeeding times performed the act as soon as put in, and the figure would have been such as we see in Fig. 3. The thing is still clearer if, instead of drawing in the perpendiculars, we draw only a line joining their tops. Fig. 4 shows, then, the curve for the real history, and Fig. 5 shows the abrupt descent, due to a rational comprehension of the situation. I kept an accurate record of the time, in seconds, taken in every trial by every cat in every box, and in them all there appears no evidence for the presence of even the little reasoning that "what let me out of this box three seconds ago will let me out now." Surely, if an animal could reason he would, after ten or eleven accidental successes, think what he had been doing, and at the eleventh or twelfth trial would at once perform the act. But no! The slope of the curves, as one may see in the specimens shown in Fig. 6, is always gradual. So, in saying that the behavior of the animals throughout the experiments gave no sign of the presence of reasoning I am not giving a personal opinion, but the impartial evidence of an unprejudiced watch. The curves given in Fig. 6 are for cats learning to escape from the box already described, whose door was held by a wooden button on the inside.