Experiments on the Abilities of the Monkeys to Learn Without Tuition

I will describe a few of the experiments with No. 1 as samples and then present the rest in the form of a table. No. 1 was tried first in BB (O at back) on January 17, 1900, being put inside. He opened the box by pulling up the string just above the bolt. His times were .05, 1.38, 6.00, 1.00, .10, .05, .05. He was not easily handled at this time, so I changed the experiment to the form adopted in future experiments. I put the food inside and left the animal to open the door from the outside. He pulled the string up within 10 seconds each time out of 10 trials.

I then tried him in MM (bolt). He failed in 15. I then (January 18th) tried him in CC (single bar outside). He got in in 36.00 minutes; he did not succeed a second time that night, but in the morning the box was open. His times thenceforth were 20, 10, 16, 25 and on January 19th, 40, 5, 12, 8, 5, 5, 5 seconds.

I then tried him (January 21, 1900) in CCC (double bar). He did it at first by pushing the old bar and then pulling at the door until he worked the second bar gradually around. Later he at times pushed the second bar. The times taken are shown in the time-curve. I then (January 25th) tried him in NN (hook). See time-curves on [page 185]. I then (January 27th) tried him in NNN (hook on other side). He opened it in 6, 12 and 4 seconds in the first three trials. I then (20 minutes later) tried him with NNNN (double hook). He opened the door in 12, 10, 6 and 6 seconds. I then (January 27th) tried him with PP (string across). He failed in 10. I then (February 21st) tried him with apparatus OO (string box). For his progress as shown by the times taken see the time-curve. His progress is also shown in the decrease of the useless pullings at the wrong string. There were none in the 9th trial, 14th, 15th, 16th, 18th, 24th, and following trials.

No. 1 was then (February 24th) tried with OOO (string box with box on upper string). No. 1 succeeded in 2.20, then failed in 10.00. The rest of the experiment will be described under imitation.

He was next tried (March 24th) with apparatus QQ (chute). He failed in 10.00, though he played with the apparatus much of the time. Other experiments were with box RR (wood-plug) (April 5th). He failed in 10.00. After he had, in a manner to be described later, come to succeed with RR, he was tried in box SS (triple; wood-plug, hook and bar) (April 18th); see time-curve. No more experiments of this nature were tried until October, 1900.

The rest of the experiments with No. 1 and all those with No. 2 and No. 3 may best be enumerated in the form of a table. (See Table 9 on [page 187].) It will show briefly the range of performances which the unaided efforts of the animals can cope with. It will also give the order in which each animal experienced them. F means that the animal failed to succeed. The figures are minutes and seconds, and represent the time taken in the first trial or the total time taken without success where there is an F. In cases where the animal failed in say 10 minutes, but in a later trial succeeded, say in 2.40, the record will be 2.40 after 10 F. There are separate columns for all three animals, headed No. 1, No. 2 and No. 3. Im. stands for a practically immediate success.

The curves on pages 185 and 186 ([Figs. 27 and 28]) show the progress of the formation of the associations in those cases where the animal was given repeated trials, with, however, nothing to guide him but his own unaided efforts. Each millimeter on the abscissa represents one trial and each millimeter on the ordinate represents 10 seconds, the ordinates representing the time taken by the animal to open the box. A break in the curve, or an absence of the curve at the beginning of the base-line represents cases where the animal failed in 10 minutes or took a very long time to get out.

Fig. 27.

Fig. 28.

In discussing these facts we may first of all 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 behavior of the monkeys toward the mechanisms used. And this fact nullifies the arguments for reasoning in their case as it did in the case of the dogs and cats. 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 shall find that at least very many simple acts were not learned by the monkeys in spite of their having seen me 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 shall find that 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 shall 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.

Table 9

No. 1.No. 2.No. 3.
Min. Sec.Min. Sec.Min. Sec.
Box TT (nail plug)Oct. 19, 19000.40Oct. 21, 190014.10Oct. 21, 190036.00
Box UU (old plug at side)Oct. 19, 1900F 60.00
Box VV (wire loop)Oct. 20, 1900{F 10.00Oct. 24, 1900F 10.00Oct. 22, 1900{F 10.00
{F 10.00Oct. 25, 1900F 10.00{F 10.00
{F 10.00{F 10.00
Box WW (bar inside)Oct. 20, 1900F 10.00Oct. 21, 19005.00after
F 30.00		Oct. 22, 1900{F 10.00
Oct. 24, 1900{F 5.00
{F 10.00
{F 15.00
Box XX (bar outside)Oct. 23, 1900im.after [25]
F 10.00		Oct. 24, 19003.40Oct. 23, 1900.30
Box YY (push bar)Oct. 30, 19002.00[26]
Box Beta (single hook)Oct. 30, 19009.00after F 10.00 and 10.00Oct. 24, 1900im.
Box LL (triple; nail plug, hook and bar outside)Nov. 4, 190016.00[27]Oct. 3, 19002.00Nov. 3, 19001.45
Box Alpha (catch at back)Nov. 5, 1900.35Oct. 5, 19006.00Nov. 5, 1900
Box KK (triple; bolt, side-plug and knob)Nov. 7, 1900F 10.00
F 10.00		Oct. 7, 1900F 60.00Nov. 7, 1900F 10.00
Box Theta (bolt at top)Nov. 19, 1900F 10.00Jan. 8, 1901F 10.00
Box Eta (ring at back)Dec. 17, 1900im.Dec. 17, 19004.20
App. QQ (push chute)Dec. 17, 1900F 60.00
Box Gamma (wind)Jan. 3, 1901.20Jan. 4, 1901F 10.00
F 10.00
Box Delta (push back)Jan. 4, 1901F 5.00
F 5.00		Jan. 4, 19012.10after[28]
F 10.00
App. QQ (a) (bar chute)Jan. 6, 19018.00Jan. 7, 1901F 10.00
Box Zeta (new side plug)Jan. 7, 19011.10after F 5.00Jan. 8, 1901.50
App. QQ (b) (2½ revolution chute)Jan. 9, 19013.00Jan. 8, 1901F 10.00
App. QQ (c) (nail-plug chute)Jan. 11, 1901F 5.00
F 5.00		Jan. 11, 1901F 5.00
F 5.00
Box Epsilon (push down)Jan. 12, 1901F 5.00
F 10.00		Jan. 12, 1901F 10.00
App. QQ (d) (ring chute)Jan. 16, 1901F 5.00
F 5.00		Jan. 16, 1901im.
App. QQ (e) (hook chute)Jan. 16, 1901F 5.00
App. QQ (f) (string chute)Jan. 17, 1901F 5.00
App. QQ (ff) (string-wire chute)Jan. 17, 1901.20Jan. 19, 1901F 5.00
F 5.00

The methods one has to take to get them to do anything, their general conduct in becoming tame and in the experiments throughout, confirm these conclusions. The following particular phenomena are samples of the many which are inconsistent with the presence of reasoning as a general function. No. 1 had learned to open a door by pushing a bar around from a horizontal to a vertical position. The same box was then fitted with two bars. He turned the first bar round thirteen times before attempting to push the other bar around. In box LL all three monkeys would in the early trials do one or two of the acts over and over after they had once done them. No. 1, who had learned to pull a loop of wire off from a nail, failed thereafter to pull off a similar loop made of string. No. 1 and No. 3 had learned to poke their left hands through the cage for me to take and operate a chute with. It was extremely difficult to get either of them to put his right hand through or even to let me take it and pull it through.

A negative answer to the question “Do the monkeys reason?” thus seems inevitable, but I do not attach to the question an importance commensurate with the part it has played historically in animal psychology. For I think it can be shown, and I hope in a later monograph to show, that reasoning is probably but one secondary result of the general function of having free ideas in great numbers, one product of a type of brain which works in great detail, not in gross associations. The denial of reasoning need not mean, and does not to my mind, any denial of continuity between animal and human mentality or any denial that the monkeys are mentally nearer relatives to man than are the other mammals.

So much for supererogatory explanation. Let us now turn to a more definite and fruitful treatment of these records.

The difference between these records and those of the chicks, cats and dogs given on [pages 39-65] passim is undeniable. Whereas the latter were practically unanimous, save in the cases of the very easiest performances, in showing a process of gradual learning by a gradual elimination of unsuccessful movements, and a gradual reënforcement of the successful one, these are unanimous, save in the very hardest, in showing a process of sudden acquisition by a rapid, often apparently instantaneous, abandonment of the unsuccessful movements and a selection of the appropriate one which rivals in suddenness the selections made by human beings in similar performances. It is natural to infer that the monkeys who suddenly replace much general pulling and clawing by a single definite pull at a hook or bar have an idea of the hook or bar and of the movement they make. The rate of their progress is so different from that of the cats and dogs that we cannot help imagining as the cause of it a totally different mental function, namely, free ideas instead of vague sense-impressions and impulses. But our interpretation of these results should not be too hasty. We must first consider several other possible explanations of the rapidity of learning by the monkeys before jumping to the conclusion that the forces which bring about the sudden formation of associations in human beings are present.

First of all it might be that the difference was due to the superiority of the monkeys in clear detailed vision. It might be that in given situations where associations were to be formed on the basis of smells, the cats and dogs would show similar rapid learning. There might be, that is, no general difference in type of mental functioning, but only a special difference in the field in which the function worked. This question can be answered by an investigation of the process of forming associations in connection with smells by dogs and cats. Such an investigation will, I hope, soon be carried on in the Columbia Laboratory by Mr. Davis.[29]

Secondly, it might be that the superior mobility and more detailed and definite movements of the monkeys’ hands might have caused the difference. The slowness in the case of the dogs and cats might be at least in part the result of difficulty in executing movements, not in intending them. This difficulty in execution is a matter that cannot be readily estimated, but the movements made by the cats and dogs would not on their face value seem to be hard. They were mostly common to the animals’ ordinary life. At the same time there were certain movements (e.g. depressing the lever) which were much more quickly associated with their respective situations by the cats than others were, and if we could suppose that all the movements learned by the monkeys were comparable to these few, it would detract from the necessity of seeking some general mental difference as the explanation of the difference in the results.

In the third place it may be said by some that no comparison of the monkeys with dogs and cats is valid, since the former animals got out of boxes while the latter got in. It may be supposed that the instinctive response to confinement includes an agitation which precludes anything save vague unregulated behavior. Professor Wesley Mills has made such a suggestion in referring to the ‘Animal Intelligence’ in the Psychological Review, May, 1899. In the July number of the same journal I tried to show that there was no solid evidence of such a harmful agitation. Nor can we be at all sure that agitation when present does not rather quicken the wits of animals. It often seems to. However I should, of course, allow that for purposes of comparison it would be better to have the circumstances identical. And I should welcome any antagonist who should, by making experiments with kittens after the fashion of these with the monkeys, show that they did learn as suddenly as the latter.

Again we know that, whereas the times taken by a cat in a box to get out are inversely proportional to the strength of the association, inasmuch as they represent fairly the amount of its efforts, on the other hand, the times taken by a monkey to get in represent the amounts of his efforts plus the amount of time in which he is not trying to get in. It may be said therefore that the time records of the monkeys prove nothing,—that a record of four minutes may mean thirty seconds of effort and three minutes thirty seconds of sleep,—that one minute may really represent twice as much effort. As a matter of fact this objection would occasionally hold against some single record. The earliest times and the occasional long times amongst very short ones are likely to be too long. The first fact makes the curves have too great a drop at the start, making them seem cases of too sudden learning, but the second fact makes the learning seem indefinite when it really is not. And in the long run the times taken do represent fairly well the amount of effort. I carefully recorded the amount of actual effort in a number of cases and the story it tells concerning the mental processes involved is the same as that told by the time-curves.

Still another explanation is this: The monkeys learn quickly, it is true, but not quickly enough for us to suppose the presence of 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.

Such an explanation we could hardly disapprove in any case. No one can from objective evidence set up a standard of speed of learning below which all shall be learning without ideas and above which all shall be learning by ideas. We should not expect any hard and fast demarcation.

This whole matter of the rate of learning should be studied in the light of other facts of behavior. My own judgment, if I had nothing but these time-curves to rely on, would be that there was in them an appearance of learning by ideas which, while possibly explicable by the finer vision and freer movements of the monkey in connection with ordinary mammalian mentality, made it worth while to look farther into their behavior. This we may now do.

What leads the lay mind to attribute superior mental gifts to an animal is not so much the rate of learning as the amount learned. The monkeys obviously form more associations and associations in a greater variety than do the other mammals. The improved rate assists, but another cause of this greater number of associations is the general physical activity of the monkeys, their constant movements of the hands, their instinctive curiosity or tendency to fool with all sorts of objects, to enjoy having sense-impressions, to form associations because of the resulting sound or sight. These mental characteristics are of a high degree of importance from the comparative point of view, but they cannot be used to prove that the monkeys have free ideas, for a large number of associations may be acquired after the purely animal fashion.

What is of more importance is the actual behavior of the animals in connection with the boxes. First of all, as has been stated, all the monkey’s movements are more definite, he seems not merely to pull, but to pull at, not merely to poke, but to push at. He seems, even in his general random play, to go here and there, pick up this, examine the other, etc., more from having the idea strike him than from feeling like doing it. He seems more like a man at the breakfast table than like a man in a fight. Still this appearance may be quite specious, and I think it is likely to lead us to read ideational life into his behavior if we are not cautious. It may be simply general activity of the same sort as the narrower activities of the cat or dog.

In the second place the monkeys often make special movements with a directness which reminds one unavoidably of human actions guided by ideas. For instance, No. 1 escaped from his cage one day and went directly across the room to a table where lay a half of a banana which was in a very inconspicuous place. It seemed as if he had observed the banana and acted with the idea of its position fully in mind. Again, on failing to pull a hook out, No. 1 immediately applied his teeth, though he had before always pulled it out with his hand. So again with a plug. It may be that there is a special inborn tendency to bite at objects pulled unsuccessfully. If not, the act would seem to show the presence of the idea ‘get thing out’ or ‘thing come out’ and associated with it the impulse to use the teeth. We shall see later, however, that in certain other circumstances where we should expect ideas to be present and result in acts they do not.

The fact is that those features in the behavior of the monkeys in forming associations between the sight of a box and the act needed to open it which remind us of learning by ideas may also be possibly explained by general activity and curiosity, the free use of the hand, and superior quickness in forming associations of the animal sort. We must have recourse to more crucial tests or at least seek evidence from a number of different kinds of mental performances. The first of these will naturally be their behavior toward these same mechanisms after a long time-interval.

The Permanence of Associations in the Case of Mechanisms

My records are too few and in all but one case after too short an interval to be decisive on the point of abrupt transition from failure to success such as would characterize an animal in whose mind arose the idea of a certain part of the mechanism as the thing to be attacked or of a certain movement as the fit one. The animals are all under observation in the Columbia Laboratory, however, and I trust that later satisfactory tests may be made. No. 2 was not included in the tests because he was either unwell or had become very shy of the boxes, entering them even when the door was left open only after great delay. The time-curves for the experiments performed will be found on [page 186] among the others. The figures beside each pair represent the number of days without practice.

The records show a decided superiority to those of the cats and dogs. Although the number of trials in the original tests were in general fewer in the case of the monkeys, the retention of the association is complete in 6 cases out of 8 and is practically so in one case where the interval was 8 months.

Experiments on the Discrimination of Signals

My experiments on discrimination were of the following general type: I got the animal into the habit of reacting to a certain signal (a sound, movement, posture, visual presentation or what not) by some well-defined act. In the cases to be described this act was to come down from his customary positions about the top of the cage, to a place at the bottom. I then would give him a bit of food. When this habit was wholly or partly formed, I would begin to mix with that signal another signal enough like it so that the animal would respond in the same manner. In the cases where I gave this signal I would not feed him. I could then determine whether the animal did discriminate or not, and his progress toward perfect discrimination in case he did. If an animal responds indiscriminately to both signals (that is, does not learn to disregard the ‘no food’ signal) it is well to test him by using two somewhat similar signals, after one of which you feed him at one place and after the other of which you feed him at a different place.

If the animal profits by his training by acquiring ideas of the two signals and associates with them ideas of ‘food’ and ‘no food,’ ‘go down’ and ‘stay still,’ and uses these ideas to control his conduct, he will, we have a right to expect, change suddenly from total failure to differentiate the signals to total success. He will or won’t have the ideas, and will behave accordingly. The same result could, of course, be brought about by very rapid association of the new signal with the act of keeping still, a very rapid inhibition of the act of going down in response to it by virtue of the lack of any pleasure from doing so.

For convenience I shall call the signals after which food was given yes signals and those after which food was not given no signals. Signals not described in the text are shown in [Fig. 29], below. The progress of the monkeys in discriminating is shown by [Figs. 30 and 31, on pages 199 and 201]. In [Figs. 30 and 31] every millimeter along the horizontal or base line represents 10 trials with the signal. The heights of the black surface represent the percentages of wrong responses, 10 mm. meaning 100 per cent of incorrect responses. Thus the first figure of the set, Left hand, a, presents the following record: First 10 trials, all wrong; of next 10, 7 wrong; of next 10, 6 wrong; of next 10, 7; of the next, 9; of the next, 9; of the next, 4; of the next, none; of the next, 3; of the next, 2, and then 70 trials without an error.

Fig. 29.

I will describe some of the experiments in detail and then discuss the graphic presentation of them all.

Experiments with No. 1

Having developed in No. 1 the habit of coming down to the bottom of his cage to get a bit of food when he saw me reach out and take such a bit from my desk, I tested his ability to discriminate by beginning to use now one hand, now the other, feeding him only when I used the left. I also used different sets of words, namely, ‘I will give some food’ and ‘They shall not have any.’ It will be seen later that he probably reacted only to the difference of the hands. The experiment is similar to that described on [pages 129 and 130] of Chapter II. At the beginning, it should be remembered, No. 1 would come down whichever hand was used, no matter what was said, except in the occasional cases where he was so occupied with some other pursuit as to be evidently inattentive. He did come to associate the act of going down with the one signal and the act of staying still or continuing his ordinary movements with the other signal. His progress in learning to do so is best seen in the curves of his errors. To the ‘yes’ signal he responded correctly, except for the occasional lapses which I just mentioned, from the start and throughout. With the ‘no’ signal his errors were as shown in [Fig. 30], a. The break in the curve at 110 and 120 is probably not significant of an actual retrograde as the trials concerned followed an eight days’ cessation of the experiments.

I next tried No. 1 with an apparatus exposing sometimes a card with a diamond-shaped piece of buff-colored paper on it and sometimes a card with a similar black piece. The black piece was three fourths of an inch farther behind the opening than the other. The light color was the ‘yes’ signal. The error curves for both signals are given, as No. 1 at the beginning of the experiment did not go down always ([Fig. 30], b and b₁).

I next tried No. 1 with the same apparatus but exposing cards with YES and N in place of the buff and black diamonds. The record of the errors is given in [Fig. 30], c and c₁. At the start he came down halfway very often. This I arbitrarily scored as an error no matter which signal it was in response to. It should not be supposed that these curves represent two totally new associations. It seems likely that the monkey reacted to the position of the N card in the apparatus (the same as that of the black diamond card) rather than to the shape of the letters. On putting the black diamond in front he was much confused.

I next gave No. 1 the chance to form the habits of coming down when I rapped my pencil against the table twice and of staying where he was when I rapped with it once. He had 90 trials of each signal but failed to give evidence of any different associations in the two cases.

Experiments of this sort were discontinued in the summer. In October I tried No. 1 with the right and left hand experiment, he being in a new room and cage, and I being seated in a different situation. He came down at both signals and failed to make any ascertainable progress with the no signal in 80 trials. (October 20-24.)

Fig. 30.

I then tried him with the black and buff diamonds, the black being in front (October 25-29). The reaction to the ‘yes’ signal was perfect from the start. The progress with the ‘no’ signal is shown in [Fig. 30], d.

I then tried him with an apparatus externally of different size, shape and color from that so far used, showing as the ‘yes’ signal a brown card and as the ‘no’ signal a white and gold card one half inch farther back in the apparatus. The ‘yes’ signal was practically perfect from the start. His progress with the ‘no’ signal is shown in [Fig. 30], e.

I then tried a still different arrangement for exposure, to which, however, he did not give uniform attention.

I then tried cards 1 and 101, 101 being in front and 1 in back. 1 was the ‘yes’ signal. ‘Yes’ responses were perfect from the start. For ‘no’ responses see [Fig. 30], f. I then put the ‘yes’ signal in front and the ‘no’ signal behind. ‘Yes’ responses perfect; for ‘no’ responses see [Fig. 30], f, a.

From now on I arranged the exposures in such a way that there was no difference between the ‘yes’ and ‘no’ signals in distance or surroundings.

The following list shows the dates, signals used, and the figures on [page 199] presenting the results. Where there is only one figure drawn, it refers to progress with the ‘no’ signal, the ‘yes’ signal being practically perfect from the start.

Table 10
‘Yes’ Signal‘No’ SignalFigure
Nov. 13-15, 1900.2102g g₁
Nov. 14-16, 1900.3103i i₁
Nov. 16-19, 1900.4104h
Nov. 19, 1900.5105j
Nov. 20, 1900.6106k
Nov. 21, 1900.7107l
Nov. 23(?), 1900.8108m
Nov. 27-29, 1900.9109n
Nov. 30, 1900.10110o

[Fig. 29] gives facsimiles of the different signals reduced to one sixth their actual size. The drawing of 101 is not accurate, the outer ring being too thick.

Experiments with No. 2

Fig. 31.

I first secured the partial formation of the habit of coming down when I took a bit of food in my hand. I then used the apparatus for exposing cards, YES in front being the ‘yes’ signal and a circle at the back being the ‘no’ signal. I gave No. 2 25 trials with the ‘yes’ signal and then began a regular experiment similar to those described. After about 90 trials (November 9-12, 1900) there was no progress toward differentiation of response, and it was evident from No. 2’s behavior that he was reacting solely to the movements of my hand. So I abandoned the exposing apparatus and used (November 11-13, 1900) as the ‘yes’ signal the act of taking the food with my left hand from a pile on the front of the box and for the ‘no’ signal the act of taking food with my right hand from a pile 4 inches behind that just mentioned. No. 2 did come to differentiate these two signals. The record of his progress is given in [Fig. 31] by A and A₁.

I then made a second attempt with the exposing apparatus, using cards 2 and 102 (November 6, 14-21). No. 2 did react to my movements in pulling the string but in over 100 trials made no progress in the direction of a differential reaction to the ‘no’ signal. I then tried feeding him at each signal, feeding him at the bottom of the cage as usual when I gave the ‘yes’ signal and at the top when I gave the ‘no’ signal. After a hundred trials with the ‘no’ signal there was no progress.

I then abandoned again the exposing apparatus and used as signals the ordinary act of taking food with my left hand (yes) and the act of moving my left arm from my right side round diagonally (swinging it on my elbow as a center) and holding the hand, after taking the food, palm up (no) (November 26, 27, 1900). No. 2 did come to differentiate these signals. His progress is given in the diagram in [Fig. 31] entitled ‘Palm up’ (B).

I next used (November 27, 1900) as the ‘yes’ signal the same act as before and for the ‘no’ signal the act of holding the food just in front of the box about four inches below the edge. No. 2’s progress is shown in [Fig. 31] in the diagram entitled ‘low front’ (C and C₁).

I next used (November 27-30) the same movement for both ‘yes’ and ‘no’ signals save that as the ‘yes’ signal I took the food from a brown pasteboard box 3 by 3 by 0.5, and as the ‘no’ signal I took it from a white crockery cover two inches in diameter and three eighths of an inch high which was beside the box but three inches nearer me. No. 2’s progress is shown in [Fig. 31] in the diagram entitled ‘Box near’ (D).

I next used for the ‘yes’ signal the familiar act and for the ‘no’ signal the act of holding the food six inches above the box instead of a quarter or a half an inch. The progress is shown in [Fig. 31], E and E₁. I then tried taking the food from a saucer off the front of the box for the ‘yes’ signal and from a small box at the back for the ‘no’ signal. ‘Yes’ was perfect from the start (10 trials given). ‘No’ was right once, then wrong once, then right for the remaining eight.

Experiments with No. 3

No. 3 was kept in a cage not half so big as those of 1 and 2. Perhaps because of the hindrance this fact offered to forming the habit of reacting in some definite way to ‘yes’ signals, perhaps because of the fact that I did not try hand movements as signals, there was no successful discrimination by No. 3 of the yellow from the black diamond or of a card with YES from a card with a circle on it. I tried climbing up to a particular spot as the response to the ‘yes’ signal and staying still as the response to the ‘no’ signal. I also tried instead of the latter a different act, in which case the animal was fed after both signals but in different places. In the latter case No. 3 made some progress, but for practical reasons I postponed experiments with him. Circumstances have made it necessary to postpone such experiments indefinitely.

Permanence of the Ability to Discriminate

No. 1 and No. 2 were tried again after intervals of 33 to 48 days. The results of these trials are shown in [Fig. 32]. Here every millimeter along the base line represents one trial with the ‘no’ signal (the ‘yes’ signals were practically perfect), and failure is represented by a column 10 mm. high while success is represented by the absence of any column. Thus the first record reads, “No. 1 with signal 104 after 40 days made 5 failures, then 2 successes, then 1 failure, then 1 success, then 3 failures, then 1 success, then 1 failure, then 3 successes, then 1 failure, then 10 successes.” The third record (106; 40 days) reads, “perfect success in ten trials.”

Fig. 32.

Discussion of Results

The results of all these discrimination experiments emphasize the rapidity of formation of associations amongst the monkeys, which appeared in their behavior toward the mechanisms. The suddenness of the change in many cases is immediately suggestive of human performances. If all the records were like c, f, h, i, j, k, l, m, B, E, and memory trials 103, A, B, and C, one would have to credit the animals with either marvelous rapidity in forming associations of the purely animal sort or concede that from all the objective evidence at hand they were shown to learn as human beings would. One would have to suppose that they had clear ideas of the signals and clean-cut associations with those ideas. The other records check such a conclusion.

In studying the figures we should remember that occasional mistakes, say 1 in 10 trials, are probably not significant of incomplete learning but of inattention or of precipitate action before the shutter had fairly exposed the card. We must not expect that a monkey who totally fails to discriminate will always respond wrongly to the ‘no’ signal, or that a monkey who has come to discriminate perfectly will always respond rightly. A sudden drop from an average high level of error to an average low level will signify sudden learning. Where the failure was on the first trial of a series a few hours or a day removed from the last series, I have generally represented the fact not by a column 1 mm. high and 1 mm. broad, but by a single 10 mm. perpendicular. See i and A. Such cases represent probably the failure of the animal to keep his learning permanent rather than any general inability to discriminate.

K was to some extent a memory trial of d (after over half a year).

The experiment with 10 and 110 is noteworthy. Although, as can be seen from the figures, the difference is obvious to one looking at the white part of the figure, it is not so to one looking at the black part. No. 1 failed to improve appreciably in fifty trials, probably because his previous experience had gotten him into the habit of attending to the black lines.

Before arguing from the suddenness of the change from failure to success we have to consider one possibility that I have not mentioned, and in fact for the sake of clearness in presentation have rather concealed. It is that the sudden change in the records, which report only whether the animal did or did not go down, may represent a more gradual change in the animal’s mind, a gradual weakening of the impulse to go down which makes him feel less and less inclined to go down, though still doing so, until this weakening reaches a sort of saturation point and stops the action. There were in their behavior some phenomena which might witness to such a process, but their interpretation is so dependent on the subjective attitude and prepossessions of the observer that I prefer not to draw any conclusions from them. On the other hand, records c, g, n, A and D seem to show that gradual changes can be paralleled by changes in the percentage of failures.

In the statement of conclusions I shall represent what would be the effect on our theory of the matter in both cases, (1) taking the records to be fairly perfect parallels of the process, and (2) taking them to be the records of the summation points of a process not shown with surety in any measurable objective facts. But I shall leave to future workers the task of determining which case is the true one.

If we judge by the objective records themselves, we may still choose between two views. (1) We may say that the monkeys did come to have ideas of the acts of going down to the bottom of the cage and of staying still, and that their learning represented the association of the sense-impressions of the two signals, one with each of these ideas, or possibly their association with two other ideas (of being fed and of not being fed), and through them with the acts. Or (2) we may say that the monkeys had no such ideas, but merely by the common animal sort of association came to react in the profitable way to each signal.

If we take the first view, we must explain the failure of the animals to change suddenly in some of the experiments, must explain why, for instance, No. 1 in g should, after he had responded correctly to the ‘no’ signal for 27 trials out of 30, fail in one trial out of four for a hundred or more trials. If the 27 successes were due to ideas, why was there regression? If the animal came to respond by staying still on seeing the K (card 104), because that sight was associated with the idea of no food or the idea of staying still, why did he, in his memory trial, act sometimes rightly, sometimes wrongly, for eleven trials after his acting rightly twice. If he stayed still because the idea was aroused, why did he not stay still as soon as he had a few trials to remind him of the idea? It is easy, one may say, to see why, with a capacity to select movements and associate them with sense-presentations very quickly, in cases where habit provides only two movements for selection and where the sense-presentation is very clear and simple, an animal should practically at once be confirmed in the one act on an occasion when he does it with the sense-impression in the focus of attention. It is easy, therefore, to explain the sudden change in i, l, m, B, C and E. But our critic may add, “It is very hard to suppose that an animal that learned by connecting the sight of a card with the idea ‘stay still’ or the idea ‘no food,’ should be so long in making the connection as was the case in some of these experiments, should take 10, 20 or 40 trials to change from a high percentage of wrong to a high percentage of right reactions.”

If we take the second view, we have to face the fact that many of the records are nothing like the single one we have for comparison, that of the kitten shown in [Fig. 30], and that the appeal to a capacity to form animal associations very quickly seems like a far-fetched refuge from the other view rather than a natural interpretation. If we take the records to be summation points in a more gradual process, this difficulty is relieved.

If further investigation upheld the first view, we should still not have a demonstration that the monkeys habitually did learn by getting percepts and images associated with sense-impressions, by having free ideas of the acts they performed; we should only have proved that they could under certain circumstances.

The circumstances in these experiments on discrimination were such as to form a most favorable case. The act of going down had been performed in all sorts of different connections and was likely to gain representation in ideational life; the experience ‘bit of banana’ had again been attended to as a part of very many different associations and so would be likely to develop into a definite idea.

These results then do not settle the choice between three theories: (1 a) that they were due to a general capacity for having ideas, (1 b) that they were due to ideas acquired by specially favoring circumstances, (2) that they were due to the common form of association, the association of an impulse to an act with a sense-impression rather roughly felt.

It would be of the utmost interest to duplicate these experiments with dogs, cats and other mammals and compare the records. Moreover, since we shall find (1 a) barred out by other experiments, it will be of great interest to test the monkeys with some other type of act than discrimination to see if, by giving the animal experience of the act and result involved in many different connections, we can get a rate of speed in the formation of a new association comparable to the rates in some of these cases.

Of course here, as in our previous section, the differences in the sense-powers of the monkeys from those of the kitten which I have tested with a similar experiment may have caused the difference in behavior. Focalized vision lends itself to delicate associations. Perhaps if one used the sense of smell, or if the dogs and cats could, preserving their same mental faculties in general, add the capacity for focalized vision, they would do as well as the monkeys.