An Introduction to Comparative Psychology
Chapter 16: Do Animals Reason?
C. Lloyd Morgan
WE are now in a position to consider the question which I have prefixed as a heading to this chapter. It is obvious that, as we attempt to answer the question, we must steadily bear in mind the sense in which the term "reasoning" is employed. If we apply this term to the process by which an animal, profiting by experience, adapts his actions to somewhat varying circumstances, there can be no hesitation whatever in giving an affirmative answer to the question. There is no doubt that animals not only profit by past experience, but that they can apply this experience to concrete situations as they severally arise. So much is implied by the attribution to them of intelligence. Where a situation is assimilated with those in which experience has already been gained, expectations arise which are a sufficient guide to practical behaviour. As we have seen, such expectations are sometimes called practical inferences: if the word is to be used with a wide meaning, we may regard them as intelligent inferences. Mr L. T. Hobhouse calls them, in a recent work on Mind in Evolution, practical judgments as contrasted with the universal judgments of conceptual thought. It is true that his "practical judgment may seem to imply more than I am disposed to allow to animal psychology. It is true that he credits the animal with a power of perceiving relations which appears to be more than I am myself prepared to grant. But then he says: " In such a perception, the relations contained contribute to the character of the whole as much as the
(289) elements that are related, and in that sense the relations may be said to be perceived, It does not follow that the character of any of the relations concerned is analyzed out and distinguished from the terms which compose it." This distinguishing of the relation from its terms, which is just what I mean by the perception of relations as considered in the thirteenth chapter, he does not claim for the animal. As the outcome of a very careful consideration of the whole question, supplemented by a number of interesting and valuable experimental observations, he concludes that " the highest animals have as much capacity for dealing with the practical exigencies of their surroundings as can be attained by an intelligence limited in its scope to the concrete and the practical. Intelligence as we conceive it in this stage is capable of forming what we have called practical judgments."
If then behaviour which is the outcome of concrete sense-experience, is placed in the game category as rational conduct based on the conceptual thought which results from the analysis of experience and the synthesis of ideal construction, we must freely admit that animals can and do reason. But I have used the term reason in a more restricted sense. Mr Hobhouse regards such restriction as arbitrary. The whole question, in his view, is a matter of degree. " It is not that new faculties are introduced, but that old faculties receive a fresh development." " A chicken avoids a caterpillar because he dislikes the taste. We perhaps refuse to allow that the chicken reasons because he does not know what it is that makes the caterpillar taste bad. After the chicken follows ;he chemist, who finds that the caterpillar secretes a certain acid. But will the chemist explain why a given acid has an acrid taste, or show how the experience of unpleasantness should modify subsequent action? A horse learns to lift a latch. We do not think he reasons. He. merely. has found out how it is done, and does it. A man explains to a child the action of the latch, and
(290) shows how by pressing it at one point you lift it out of a catch at another. He, we say, reasons because he analyzes the process and how it is done. But a physicist 'might point out that the man knows nothing whatever about it unless he sees that the principle of the lever is involved in a simple form; and a metaphysician might add that the physicist cannot be said to understand the principle -of the lever unless he is prepared to decide whether it is a principle which holds true of reality, and if so, on what epistemological grounds. If we allow reason to the human-species in general, and yet restrict it to that species, it must be by identifying the term reason arbitrarily with a certain grade in the development of analysis."
But there are no grounds for supposing that in the chicken or the horse there is-any development of analysis. It is not a question of a certain grade but of any grade. I may call Mr Hobhouse himself as a witness. Under the heading Absence of Analysis he says: " At the same time it must be understood that, if we attribute ideas to an animal, they are not ideas arrived at by any breaking-up, analysis", or other elaboration of what is given in perception [i.e., concrete experience]. None of my animals (with the possible exception now and again of the monkeys) showed the least understanding of the how or why of their actions, as- distinct from the crude fact that to do such and such a thing produced the result they required. It is the want of what wemay call analysis that made, for example, the push-back bolt [in certain experiments] such a difficulty. What Jack [a dog] and the elephant knew was, crudely, that they had to push this bolt. That the reason why they had to push it was to get it clear of the staple they obviously' never grasped." Such is Mr Hobhouse's testimony. Of course all definition and restriction of terms is arbitrary, The object is to attain such clearness of thought as will enable us to understand exactly what we are discussing. To this
(291) end I have restricted the term "reason," not indeed so as to identify it with a certain grade of analysis, but rather to that process which, as the result of analysis and resynthesis, affords a scheme in the light of which action is taken. The process which I have described in an earlier chapter under the heading The Perception of Relations is the avenue to analysis; and this process, as we have seen, is closely bound up with the beginnings of conceptual thought. Mr Hobhouse, indeed, says that when, by an act of analysis, we make the relation a distinct object of thought, independent of the terms which it connects in a particular case, we pass from perception to conception, and this passage takes place in such close connection with the focussing of the relation in the particular case, that the question whether animals reason, in the restricted sense of the term, and the question whether they are capable of analysis are very closely related. I have already recorded my opinion that animals do not focally perceive relations: from this it follows that they do not utilize conceptions so as to reason. But it will he well to discuss the question further from the somewhat different standpoint which we have now reached.
Tony, the fox-terrier pup already introduced to my readers, when he wanted to go out into the road, used to put his head under the latch of the gate, lift it, and wait for the gate to swing open.' -Now an observer of the dog's intelligent action might well suppose that he clearly perceived how the end in view was to be gained, and the most appropriate means for effecting his purpose. But here much depends on the sense in which this statement is understood. It may be understood in the sense that the situation had acquired what Dr Stout calls -meaning," so that certain concrete surroundings suggested directly, and without analysis, a given modeof practical behaviour. Or it may be understood in the sense that the dog formed a
(292) general conception of means such as could be profitably applied to this particular end. If the former interpretation be correct, I should say that Tony acted intelligently as the result of sense-experience; if the latter, I should regard his conduct as rational. And it may be said that it is quite impossible to decide between the two views, since we cannot ascertain what passed through the dog's mind. Once more, therefore, I must draw attention to the canon of interpretation adopted at the outset of our inquiries concerning other minds than ours, namely, that in no case is an animal activity to be interpreted in terms of higher psychological processes, if it can be fairly interpreted", in terms of processes which stand lower in the scale of psychological evolution and development. The question is therefore whether Tony's behaviour can be fairly explained, without his forming any conception of the relation of the means employed to the end attained. It appears to me that it can. I watched the development of the habit. The gate is of iron and has iron bars running vertically with interspaces of five or six inches between. On either side is a wall or low parapet, on which are similar vertical rails. The latch of the gate is at a level of about a foot above that of the top of the low wall. When it is lifted, the gate swings open by its own weight. When the dog was put out of the front door he naturally wanted to get out into the road, where there was often much to interest him; cats to be worried, other dogs with whom to establish sniffing acquaintance, and so forth. I watched the dog at very early stage of the development of the habit. He then ran up and down the low wall, and put his head out between the iron bars, now here, now there, now elsewhere, keenly gazing into the road. This he did for quite three or four minutes. Although he had gone out of that gate many times, although he had opportunities for seeing me lift the latch (a matter that probably had no interest what-
(293) -ever for him, following me out being a matter of course in his experience), he did not specially look out at or near the gate. He certainly did not seem to have any notion of means to attain an end; nor indeed did he seem to be trying to get out. He appeared only to be looking restlessly and wistfully at the familiar road. At length it so happened that he put out his bead beneath the latch, which, as I have said, is at a convenient height for his doing so, being about a foot above the level of the wall. The latch was thus lifted. He withdrew his head and began to look out elsewhere, when he noticed that the gate was swinging open, and out he bolted. After that, whenever I took him out, instead of opening the gate for him, I waited until he lifted the latch. Gradually he went, after less frequent poking of his head in the wrong place, to the one opening from which the latch could be lifted. But it was nearly three weeks, during which I took him out about a dozen times, before he went at once and without hesitation to the right place and put his head without any ineffectual fumbling beneath the latch. Why did he take so long? I think partly because there was so little connection between gazing out into the road and getting out into the road. He did not, at first at any rate, seem to do the former in order to effect the latter. The relation between means and end did not appear to take form in his mind, even subconsciously as means to the end. And I take it that he never had the faintest notion of how or why looking out just there came to mean walking forth into the road.
With regard to this particular trick, then, I venture to affirm that, when we know the whole history of it, Tony's action is quite similar in kind to' that of my little, chick, Blackie, which, profiting by a chance experience, pulled down the corner of the newspaper and escaped from my experimental poultry-yard. As it stands, it is quite within the range of sense-experience-, nay, more, it affords a very
(294) pretty example of the application of sense-experience to new circumstances. It is typically intelligent.
One of my own students, Mr Edward J. Shellard, informs me of a case which fell within his own observation, of an action which " appeared at first to be the result of thought," but which on closer investigation was 'clearly seen to be the outcome of sense-experience. 'A Scotch staghound living in a yard was often shut out, and in order to enter raised the latch of the door opening into the yard. Here again the chance observer -would be likely to fix upon the fact of the latch being purposely raised, but the investigator would seek to know the exact manner in which the action was performed, and how the habit was acquired. In this instance the manner was as follows : The staghound " at first raised his paws to the door and scratched violently, manifesting various signs of impatience. His scratches, which extended from the top of the door downwards and over the whole area, would thus inevitably at some time or other reach the handle of the latch, which was thus struck forcibly downwards, the latch itself rising upwards. The door would then open from the weight of the dog pushing against it. The dog always opened the door in this manner from the time when the incident was first noticed until he left, a period of about three years. The door was opened with no greater ease at the expiration of that period than at the commencement. His paws would strike various parts of the door, and he never appeared to exercise any degree of judgment in the localization of his strokes, the fact of his paws striking the handle of the latch being a necessary result, provided the dog had sufficient patience and strength to continue"
I have elsewhere described subsequent observations on my fox-terrier. I watched his behaviour when a solid
(295) india-rubber ball was thrown towards a wall standing at right angles to its course- -At first he followed it right up to the wall, and then back as it rebounded. So long as it travelled with such velocity as to be only just ahead of him he pursued the same course. But when it was thrown more violently, so as to meet him on the rebound as he ran towards the wall, lie learnt that he was able to seize it as it came towards him. And, profiting by the experience thus gained, he acquired the habit, though for long with some uncertainty of reaction, of slowing off when the object of his pursuit approached the wall, thus awaiting its rebound. Again, when the ball was thrown so as to glance at a wide angle from a surface, at first, when the velocity was such as to keep it just ahead of him, he followed its course. But when the velocity was increased he learnt to take a short cut along the third side of a triangle, so as to catch the object at some distance from the wall. Another series of experiments was made at a spot where a right angel was formed by the meeting of two surfaces. One side of the angle, the left, was dealt with for a day or two. At first the ball was closely and directly followed. Then a short cut was taken to meet its deflected course. On the fourth day this behaviour was well established. On the fifth the ball was thrown so as to strike the other or right side of the angle, and thus be deflected in the opposite direction. The dog followed the old course (the short cut to the left), and was completely nonplussed, searching that side, then more widely, and not finding the ball for eleven minutes. On repeating the experiment thrice, similar results were obtained. On the following day the ball wag thrown just ahead of him, so as to strike to the right of the angle, and was followed and caught. This course was pursued for three days, and he then learnt to take a short cut to the right. On the next day the ball was sent, as at first, to the left, and the
(296) dog was again nonplussed, having raced off to the right, in accordance with the more recent results of experience. Up to the time of his death I did not succeed in getting Tony to associate a given difference of initial direction with a resultant difference of deflection.
A well-known writer, Dr. Andrew Wilson, describes the case of a dog which used to hunt a rabbit nearly every morning down a curved shrubbery, and each time ran it into a drain at the end. "The dog then appears to have come to the conclusion "-I quote Dr Wilson's words "that the chord of a circle is shorter than its arc, for he raised the rabbit again, and instead of following' him through the shrubbery, as usual, he took the short cut to the drain, and was ready and in waiting for the rabbit when he arrived, and caught him." It is here assumed that the dog perceived the relation between a chord and its arc. I do not myself believe that he did; but that is not the question; that- is a matter of individual opinion. The question is: Can we or can we not explain the dog's action as the outcome of sense-experience, as indicative of intelligence profiting by association? I do not see how this can be denied. The terrier used to start the rabbit nearly every morning, and each time saw it escape into the old drain. That the sight of the rabbit should suggest, the drain into which it daily escaped, and that when the, idea was suggested the dog should run there directly, is a sequence not impossible, one would think, to se rise-(experience. And if so, the canon of interpretation, so often referred to, makes it imperative for us who adopt it to accept the. interpretation of the action as due to the simpler exercise of intelligence based on sense-experience rather than that according to which the dog perceived the relation between the chord and its arc.
Commenting on this suggested explanation, Mr Hob-
(297) -house says: "I do not, any more than Mr Lloyd Morgan, suppose a dog to know that a chord is shorter than its arc, but neither -do I think an association of the idea of the drain with that of the rabbit adequate nor even relevant to the case. Why should an 'idea of the drain' cause the dog to run to it ? The dog does not want to catch a drain, but a rabbit. What is needed is the idea of the I rabbit at the drain,' and that, moreover, as an event that will shortly take place. In fact, as we analyze the idea, it turns into the judgment, 'the rabbit will run to the drain' as it did yesterday, and then it becomes an intelligible basis for the dog's action in taking the nearest course. In any case, whether we call it association or judgment, we' have to admit that the dog acted, not as he acted before, but with a difference; and this difference is explained if we conceive him as applying the results won from previous experience to present circumstances in subservience to his desire." I confess I should not have expected such a criticism from Mr Hobhouse. It shows, however, how difficult it is to make one's meaning plain in such matters as those which are under discussion. I had said:  "Sense-experience and association afford the basis of a great number of expectations of the greatest practical value in the conduct of life. Such expectations may be described as intelligent inferences or inferences in the field of sense-experience." And I thought that rabbit at the drain," as the expectation resulting from previous I experience, and the application of this experience, without analysis, to present circumstances, was just what might be inferred from my discussion of the nature of in telligence. So, too, in criticising my interpretation of another piece of animal behaviour, Mr Hobhouse says 
(298) that "out of past experience the animal picks a way of satisfying -its desires. The practical judgment is not independent of associations, for association supplies the whole of the material. But out of the material it, selects what it wants, and shapes it as required." But ten years before Mr Hobhouse's work appeared I had said: "The ability to perform acts in special adaptation to special circumstances, the power of exercising individual choice between contradictory promptings, and the individuality or originality manifested in dealing with the complex conditions of an ever-changing environment-these seem to be distinctive features of intelligence" as contrasted with instinct. I wish, to avoid controversy, and therefore only add that Mr Hobhouse has stated in other words, and perhaps better, an interpretation in all essential features that which I had striven, ineffectually it would seem, to express.
Dr Thorndike has laid great stress on the importance of what he terms the impulse, which he defines  as " the direct feeling of the doing as distinguished from the idea of the act. done, gained through eye, etc." He regards the current' associationist interpretation as implying " that an animal, whenever he thinks of an act, can supply the impulse to do the act," and claims that the, groundwork of animal associations is not the association of ideas, but the association of, idea or sense-impressjon with impulse." The point is an. important one, and worthy of emphasis. But presumably few who had been led to deny to animals the formation of: free ideas, the products of analysis, who had been forced- to. lay stress upon practical, concrete experience as the basis of behaviour, would have dreamt of suggesting that " the impulse which actually does it " could be "supplied at will.",
They might even be tempted to question whether, as the result of conceptual thought, it could be supplied at will in the absence of previous association. In any case, in animal psychology the " feeling of doing " must be an. inalienable part of the practical situation where behaviour is concerned. And Dr Thorndike has done well to bring it into clear prominence and relief.
The experiments described by Dr Thorndike in his monograph are of considerable interest. Animals, chiefly cats, but also one or two dogs, were placed, hungry, in cages rudely constructed of wooden laths, and forming somewhat cramped prisons about twenty inches long by fifteen broad and twelve high. From these the animal could escape by some simple act, such as pulling a loop of cord, pressing a lever or standing on a platform, or by two or more such acts where the door was held to by more than one bolt. The animal was put into the enclosure, food was left outside in sight, and his actions observed. In some cases the door was opened for the cat when it licked itself or scratched itself. The net result of the experiments is that the animal claws about aimlessly under the impulse to escape from confinement, until it chances to claw the string or loop or button so as to allow the door to swing open. " Gradually all the other non-successful impulses will be stamped out, and the particular impulse leading to the successful act will be stamped in by the resulting pleasure, until, after many trials, the cat will, when put in the box, immediately claw the button or loop in a definite way. . . . Starting, then, with its store of instinctive impulses, the cat hits upon the successful movement, and gradually associates it with the sense-impression of the interior of the box until the connection is perfect, so that it performs the act as soon as confronted with the sense- impression." " Associations between licking or scratching itself and escaping are similarly established, and there was a noticeable tendency
(300) to diminish the act Until it becomes the mere vestige of a lick or a scratch."
More recently Dr Thorndike has extended his observations to monkeys. The apparatus used cannot here be described. "In their method of learning," says this observer, "the monkeys do not advance far beyond the generalized mammalian type, but in their proficiency in that method they do. They seem at least to form associations very much faster, and they form very many more. They also seem superior in the delicacy and in the complexity of the associations formed, and the connections seem to be more permanent." "In discussing these facts," he says in an earlier part of his monograph, "we may first clear our way of one popular explanation, that this learning was due to ' reasoning.' If we use the word reasoning in its technical psychological meaning as the function of reaching conclusions by the perception of relations, comparison and inference, if we think of the mental content involved as feelings of relation, perceptions of similarity, general and abstract notions and judgments, we find no evidence of reasoning in the behaviour of the monkeys towards the mechanisms used. . . . The argument that successful dealings with mechanical contrivances imply that the animals reasoned out the properties of the mechanisms, is destroyed when we find mere selection from their general instinctive activities sufficient to cause success with bars, hooks, loops, etc. There is also in the case of the monkeys, as in that of the other mammals, positive evidence of the absence of any general function of reasoning. We find that at least many simple acts were not learned by the monkeys in spite of their having seen one perform them again and again; that the same holds true of many simple acts which they saw other monkeys do, or were put through by me. We find that
(301) after having abundant opportunity to realize that one signal meant food at the bottom of the cage and another none, a monkey would not act from the obvious inference and consistently stay up or go down, as the case might be, but would make errors such as would be natural if he acted under the growing influence of an association between sense-impression and impulse or sense-impression and idea, but quite incomprehensible if he had compared the two signals and made a definite inference. We find that, after experience with several pairs of signals, the monkeys yet failed when a new pair was used to do the obvious thing to a rational mind-viz., to compare the two, think which meant food, and act on the knowledge directly. . . . The monkeys learn quickly, it is true, but not quickly enough for us to suppose the presence of [free, or analytic] ideas, or the formation of associations among them. For if there' were such ideas they should in the complex acts do even better than they did. The explanation then is a high degree of facility in the formation of associations of just the same kind as we found in the chicks, dogs, and cats."
Mr A. J. Kinnaman has made valuable and interesting observations on the "Mental Life of two Macacus rhesus Monkeys in Captivity."  Some of these had for their object to test how far the monkeys could discriminate between the forms of vessels and establish associations between form and food contained or not contained therein. In his summary of the results of these observations he says that "the monkeys are able to discriminate these forms and to associate food with them consecutively. The associations Are not formed by a single trial, but conic about more or less gradually through much repetition. It is easier to form an association de novo than to break an established one and
(302) form a new one. The necessity of forming a new association induces a revival of former associations of the same general kind. The learning process, upon the whole, is still that of trial, happy accident, recollection of the fortunate movements, and an elimination of the useless ones." Experimenting with his monkeys in a maze, similar to that used by Dr Willard Small in his observations on rats, Mr Kinnaman says  that the results offer no new problem above that of working a combination lock or associating food with one of a series of glasses by number, form, or colour. . . . In the learning process we have here again a more or less definitely directed effort spurred by the food stimulus, fortunate accidents, memory of them, and the elimination of useless efforts."
Mr Kinnaman says  that his monkeys did not reason inthe higher sense of the term. But he thinks that they may proceed on a method of reasoning by analogy. "The ruling out of reasoning by analogy with all lower animals, he remarks, "is often due to a failure to differentiate sufficiently the psychological process of analogical reasoning, resulting in practical activity, from a subsequent logical analysis, accounting for the intelligent act. Of course the animals cannot do the latter. In part their reasoning is like that of the human being. Yesterday a man saw a vine and handled it without evil results. To-day he sees another quite like it, handles it, and is poisoned. He does not say, ' Lo, now, here is this and this likeness, therefore it is safe to handle this vine.' He was just dimly conscious of a resemblance. He may not possibly be able to name a single likeness if put to the test. So far in his process he and the monkey have gone along together." And he prefaces these remarks by saying that however true it may be of chicks, cats, and dogs that they cannot reason by analogy,
I very much doubt whether it is true of rhesus monkeys." But when a chick, having had experience of bees, avoids the mimicking Eristalis or drone-fly, might we not on similar grounds say that it too reasoned by analogy? That animals are able to sense resemblance as a matter of practical experience is unquestionable. Their behaviour is largely based on their doing so. But there is much difference between sensed resemblance and perceived similarity of relations. And it is to procedure based on. this higher process to which I should restrict the phrase reasoning by analogy. Of this Mr Kinnaman does not furnish conclusive evidence. But his treatment of the subject is eminently fair and careful. And he says: "Whether these animals have 'free ideas' and general notions beyond the mere -'recept' [or generic image], and are capable of real analogical reasoning, cannot be positively determined If they do, the processes certainly do not rise to the level of full reflective consciousness."
Mr L. T. Hobhouse, in his able work on Mind in Evolution, has recorded well-devised experiments on several of the higher mammals, including a rhesus monkey and a chimpanzee. In the animals below the Primates he finds evidence of what he terms practical ideas "of a very crude and unanalyzed character." "Evidence," he says,  "of more articulate ideas is much more restricted, and so far as decisive experimental tests are concerned, confined, I believe, to apes. By a more articulate idea is meant one in which comparatively distinct elements are held in a comparatively distinct relation." But the data on which behaviour is founded are essentially concrete. " That is to say we deal in this stage not with the relation as such, but with two or more related objects of experience. . . . In a given case a consequence is anticipated on the basis of a parallel experience, but there is no
(304) consciousness of the implied generalisation, nor even an analysis revealing the point of identity as against the individual differences between the two cases."  As before stated, he concludes that "the higher animals have as much capacity for dealing with the practical exigencies of their surroundings as can be, attained by an intelligence limited in its scope to the concrete and the practical." " Caution, cunning, and sagacity of the kind of which animal stories are so full do not as a rule imply anything more or less than the concrete experience that we have described." I find some difficulty in following his treatment of relations. He says that the particular relations are explicit, the universal that connects them operates unconsciously, and, again, that the related term which in the previous stage merely influences action, is now, in the higher phases of animal behaviour, brought explicitly into consciousness. But he also says that "if we attribute ideas to an animal, it must be understood that they are not ideas arrived at by any process of analysis." In my own treatment of the perception of relations the beginnings of analysis are involved. The relation, implicit in the body of concrete experience, is through comparison rendered explicit as a focal object in consciousness. I do not think, therefore, that Mr Hobhouse's interpretation differs in essentials from that which I suggested. The distinction between concrete experience and conceptual thought holds good for him as it does for me. But he has brought out the nature of the advance in intelligent behaviour and the distinction between its lower and its higher phases in a way which I did not' attempt. I am not wholly satisfied with his treatment ; but I cannot deal with it more fully here.
The characteristic feature of the conclusions above quoted is that they are based on the results won by careful experiments and on observations directed ad hoe. The body of
(305) experimental evidence is much larger and more reliable than it was when this work first appeared nine years ago. And yet further work is urgently needed. I may illustrate. the kind of observation which is open to those who have intelligent animals by the following account which Dr. Alexander Hill, F.R.S., contributed to Nature:- 
"That an animal,"he says in his prefatory remarks, "can form a perceptual judgment, which leads to action suitably adapted to its circumstances, no one doubts ; but this is hardly reasoning in the usually accepted meaning of the term. We may, for the sake of simplicity, term the forming of a perceptual judgment putting one and one together. But can an animal compare an inference with an inference ? Can it 'put two and two tegether' within the common meaning of this phrase? My meaning will, I think, make itself sufficiently clear in the description of the following experiment :-
"An exceptionally intelligent fox-terrier was taught to open a box by lifting a wooden latch with its nose. Some care was spent upon the design of this box (Fig. 20). The latch was in the first instance long, and therefore easily lifted. Behind the door was placed a spiral spring, which could be twisted until it exerted any degree of pressure
(306) which seemed desirable. As the dog learnt to lift the latch, the length of the latch was curtailed. At the same time the spring was tightened until it pressed against the door with a degree of force which made the latch so stiff that the dog could not lift it without deliberate effort. There was no risk of its being opened by a chance movement. The dog was rewarded with food for performing the trick, which soon became so familiar as to be a game. As often as the door was closed the dog opened it. If be found the box on the floor he invariably opened it without waiting for any sign. Frequently he examined the interior of the box when he had opened it, but food was never placed inside it. One evening, after the trick had been shown to a number of friends in order that the dog's almost ridiculous familiarity with it might be noted, Peter was sent to bed without his supper. He is fed but once a day. Next morning a hot grilled bone was placed in the box. The box was placed in a small yard surrounded by the house. The I boot-room ' opens into the yard on one side, and into a passage on the other. After the dog had had a run in the garden the passage door into the boot-room was opened. We were watching the yard from an upper window. Two minutes after entering, the boot-room Peter smelled the bone, ran through into the yard, and approached the box. When he saw the latch he ducked his head as if intending to lift it, but desisted. He then sniffed excitedly at the box and pushed it with his nose. He returned to the boot-room. After a few minutes he came out again into the yard and sniffed in the same way at the box. Twice he pushed the latch from behind, but did not put his head beneath it. After a while he returned to the boot-room and showed no signs of revisiting the box. Ile was then taken for a twelve-mile run in the country. - As he seemed to be tired when he reached home, he was left for half an hour in the boot-room to Test. After a run in the garden he was re-admitted to the boot-room, with the yard door open. Unluckily the wind blew the door to before Peter bad gone into the yard. After we had watched for some time my son went down to see what had happened, opened the door, and pushed, the dog through it, backwards. Ile went straight to the box, lifted the latch in the most businesslike way, and took out the bone.
" The experiment was repeated a fort night later with identical results, The dog ran into the yard, sniffed at the box, pushed it with his nose, was very eager to get the meat, but, this time, be showed no sign of remembering the way to open the box. He returned a second time, and then desisted altogether. During the morning the dog remained about the house. He constantly asked to be admitted into the boot-room, and showed in the clearest manner that he remembered
(307) that the grilled bone was to be found that way. At twelve o'clock the door was opened for him. He went straight through into the yard,' opened the box, and took out the bone, which be attacked without any sign of doubting his legal right to its possession. It may be noticed that he is frequently fed in this yard.
In this experiment the dog knew two things. He knew how to open the box. Indeed, the sight of the latch was so strongly associated in the dog's mind with the action of lifting it that it is surprising that the usual, almost mechanical, response to sensation did riot occur. Had he lifted the latch it would not necessarily have implied that be did it with the object of securing the food. He knew that the box contained meat. Eager as he was to secure the meat, he did not reason , 'The way to secure the meat is to lift the latch.' I have described the experiment in detail, because all details are, as it appears to me, of great importance. It is to be noted that the opening of the box was associated in the dog's mind with the approbation of a human being. Great care was taken that no person should be present when the dog found the box. The sight of the box was strongly suggestive to the dog's mind of the action of opening it. With a view to diminishing the urgency of this sensori-motor association a piece of hot meat with a strong 'brown smell' was placed in the box. Its rich scent distracted his attention from the Latch. When the dog was readmitted to the yard later in the morning be was aware that the box was in the yard, and he went straight from a person to the box. By this time the bone was cold, and its scent less striking. It is impossible to repeat the experiment upon Peter, because now, when he opens the box, he invariably searches for food inside it."
As originally published in 1894, the concluding paragraph of this chapter stood as follows:--
On the whole, I am inclined to conclude that when we separate observed facts from observers' inference there is a remarkably small percentage of cases in which the interpretation, on the assumption of sense -experience only, will not hold good. Such are the cases which should, wherever possible, be made the basis of an experimental investigation. As matters stand at present, I think it far more probable that the small percentage of outstanding cases would, on complete investigation, be shown to be the result of the exercise of intelligence, than that they involve reason in the
(308) sense in which I have used this word. I am very far from wishing to occupy the false position of dogmatic denial of rational powers to animals. I think it is a subject 'for, further and fuller investigation. Put I do express the opinion that the fuller and more careful the investigation, the less is the satisfactory evidence of processes of reasoning; and that, though the question is still an open one, the probabilities are that animals do not reason..
I regard the results of the further and fuller investigation briefly summarised above as confirmatory of my main thesis, though they have unquestionably led to a more complete and adequate interpretation of many of the details of the working of animal intelligence. There are other investigations to which 1 have not had space to allude. Dr Thorndike has shown that the associative processes leading to intelligent adaptation through sense-experience are present" in the fish; Mr Yerkes has observed them in the turtle and the green frog; and Dr Willard Small has fully and ably discussed the role they play in the behaviour of rats. We have definitely entered on a new phase of Comparative Psychology, that in which careful and conscientious observation is made the basis of a discussion founded on an adequate knowledge of psychology.