Animal intelligence: Experimental studies

by Edward L. Thorndike

Introduction

The work to be described in this paper is a direct continuation of the work done by the author in 1897-1898 and described in Monograph Supplement No. 8 of the Psychological Review under the heading, ‘Animal Intelligence; an Experimental Study of the Associative Processes in Animals.’[23] This monograph affords by far the best introduction to the present discussion, and I shall therefore assume an acquaintance with it on the part of my readers.

It will be remembered that evidence was there given that ordinary mammals, barring the primates, did not infer or compare, did not imitate in the sense of ‘learning to do an act from seeing it done,’ did not learn various simple acts from being put through them, showed no signs of having in connection with the bulk of their performances any mental images. Their method of learning seemed to be the gradual selection of certain acts in certain situations by reason of the satisfaction they brought. Quantitative estimates of this gradualness were given for a number of dogs and cats. Nothing has appeared since the ‘Experimental Study’ to negate any of these conclusions in the author’s mind. The work of Kline and Small[24] on rodents shows the same general aspect of mammalian mentality.

Adult human beings who are not notably deficient in mental functions, at least all such as psychologists have observed, possess a large stock of images and memories. The sight of a chair, for example, may call up in their minds a picture of the person who usually sits in it, or the sound of his name. The sound of a bell may call up the idea of dinner. The outside world also is to them in large part a multitude of definite percepts. They feel the environment as trees, sticks, stones, chairs, tables, letters, words, etc. I have called such definite presentations ‘free ideas’ to distinguish them from the vague presentations such as atmospheric pressure, the feeling of malaise, of the position of one’s body when falling, etc. It is such ‘free ideas’ which compose the substance of thought and which lead us to perhaps the majority of the different acts we perform, though we do, of course, react to the vaguer sort as well. I saw definitely in writing the last sentence the words ‘majority of the different acts’ and thought ‘we perform’ and so wrote it. I see a bill and so take check book and pen and write. I think of the cold outside and so put on an overcoat. This mental function ‘having free ideas,’ gives the possibility of learning to meet situations properly by thinking about them, by being reminded of some property of the fact before us or some element therein.

We can divide all learning into (1) learning by trial and accidental success, by the strengthening of the connections between the sense-impressions representing the situation and the acts—or impulses and acts—representing our successful response to it and by the inhibition of similar connections with unsuccessful responses; (2) learning by imitation, where the mere performance by another of a certain act in a certain situation leads us to do the same; and (3) learning by ideas, where the situation calls up some idea (or ideas) which then arouses the act or in some way modifies it.

The last method of learning has obviously been the means of practically all the advances in civilization. The evidence quoted a paragraph or so back from the Experimental Study shows the typical mammalian mind to be one which rarely or never learns in this fashion. The present study of the primates has been a comparative study with two main questions in view: (1) How do the monkeys vary from the other mammals in the general mental functions revealed by their methods of learning? (2) How do they, on the other hand, vary from adult civilized human beings?

The experiments to be described seem, however, to be of value apart from the possibility of settling crucial questions by means of the evidence they give. To obtain exact accounts of what animals can learn by their own unaided efforts, by the example of their fellows or by the tuition of a trainer, and of how and how fast they learn in each case, seems highly desirable. I shall present the results in the manner which fits their consideration as arguments for or against some general hypotheses, but the naturalist or psychologist lacking the genetic interest may find an interest in them at their face value. I shall confine myself mainly to questions concerning the method of learning of the primates, and will discuss their sense-powers and unlearned reactions or instincts only in so far as is necessary to its comprehension.

It has been impossible for the author to make helpful use of the anecdotes and observations of naturalists and miscellaneous writers concerning monkey intelligence. The objections to such data pointed out in Chapter II, [pp. 22-26], hold here. Moreover it is not practicable to sift out the true from the false or to interpret these random instances of animal behavior even if assuredly true. In the study of animal life the part is only clear in the light of the whole, and it is wiser to limit conclusions to such as are drawn from the constant and systematic study of a number of animals during a fairly long time. After a large enough body of such evidence has been accumulated we may be able to interpret random observations.

The subjects of the experiments were three South American monkeys of the genus Cebus. At the time of beginning the experiment No. 1 was about half grown, No. 2 was about one fourth full size and No. 3 was about half grown. No. 1 was under observation from November, 1899, to February, 1900; No. 2 and No. 3 from October, 1900, to February, 1901. No. 1 was during the period of experimentation decidedly tame, showing no fear whatever of my presence and little fear at being handled. He would handle and climb over me with no hesitation. No. 2 was timid, did not allow handling, but showed no fear of my presence and no phenomena that would differentiate his behavior in the experiments discussed from that of No. 1, save much greater caution in all respects. No. 3 also showed no fear at my presence. Any special individual traits that are of importance in connection with any of the observations will be mentioned in their proper places. No. 1 was kept until June, 1900, in my study in a cage 3 by 6 by 6 feet, and was left in the country till October, 1900. From October, 1900, all three were kept in a room 8 by 9 feet, in cages 6 feet tall by 3 long by 2.6 wide for Nos. 1 and 2, 3 feet by 3 feet by 20 inches for No. 3. I studied their behavior in learning to get into boxes, the doors to which could be opened by operating some mechanical contrivance, in learning to obtain food by other simple acts, in learning to discriminate between two signals, that is, to respond to each by a different act, and in their general life.

Following the order of the ‘Animal Intelligence,’ I shall first recount the observations of the way the monkeys learned, solely by their own unaided efforts, to operate simple mechanical contrivances.

Besides a number of boxes such as were used with the dogs and cats (see [illustration on p. 30]), I tried a variety of arrangements which could be set up beside a cage, and which would, when some simple mechanism was set in action, throw a bit of food into the cage. [Figure 26] shows one of these. See description of QQ (ff) on [page 182].

Fig. 26. A, loop; BB, lever, pivoted at M. A bit of food put in front of C would be thrown down the chute DDD when A was released.