========+==========+=========+=========+===========+=============+===========+===========+===========+=============+===========+===========+===+===+========
| No. | S.1 | S.2 | S.3 | S.4 | S.5 | S.6 | S.7 | S.8 | S.9 | S.10 | | | Ratio
Date | of | | | | | | | | | | | R | W | of
| trials | 1.2.3 | 8.9 | 3.4.5.6.7 | 7.8.9 | 2.3.4.5.6 | 6.7.8 | 5.6.7 | 4.5.6.7.8 | 7.8.9 | 1.2.3 | | | R to W
--------+----------+---------+---------+-----------+-------------+-----------+-----------+-----------+-------------+-----------+-----------+---+---+--------
April | | | | | | | | | | | | | |
19 | 1- 10 | 1 | 9.8 | 6.7.3 | 9.7 | 6.2 | 7.8.6 | {6.7.7.7 | 4 | 7 | 2.3.3.1 | 3 | 7 | 1:2.33
| | | | | | | | {6.5 | | | | | |
| | | | | | | | | | | | | |
20 | 11- 20 | 3.2.1 | 9.8 | 5.3 | 7 | 4.2 | 8.8.6 | 5 | 8.4 | 7 | 3.1 | 3 | 7 | 1:2.33
21 | 21- 30 | 3.1 | 8 | 3 | 8.7 | 6.2 | 6 | 5 | 6.4 | 9.7 | 1 | 5 | 5 | 1:1.00
22 | 31- 40 | 1 | 9.8 | 3 | 7 | 6.2 | 6 | 6.7.5 | 5.8.4 | 9.8.9.8.7 | 2.1 | 4 | 6 | 1:1.50
23 | 41- 50 | 2.3.1 | 8 | 5.7.3 | 7 | 4.2 | 6 | 5 | 7.8.4 | 7 | 3.1 | 5 | 5 | 1:1.00
24 | 51- 60 | 1 | 8 | 4.5.7.3 | 9.7 | 5.6.2 | 6 | 6.7.5 | 6.4 | 8.9.7 | 1 | 4 | 6 | 1:1.50
26 | 61- 70 | 1 | 8 | 6.7.4.7.3 | 7 | 4.5.6.2 | 6 | 5 | 8.4 | 7 | 3.2.3.1 | 6 | 4 | 1: .67
27 | 71- 80 | 3.1 | 8 | 3 | 9.7 | 4.6.2 | 7.6 | 6.5 | 5.8.4 | 7 | 1 | 4 | 6 | 1:1.50
28 | 81- 90 | 2.3.1 | 8 | 3 | 7 | 4.5.6.2 | 6 | 5 | 5.8.4 | 7 | 1 | 7 | 3 | 1: .43
29 | 91- 100 | 1 | 8 | 3 | 9.7 | 6.2 | 6 | 5 | 4 | 7 | 1 | 8 | 2 | 1: .25
30 | 101- 110 | 1 | 8 | 4.3 | 7 | 5.6.2 | 6 | 5 | 4 | 7 | 2.3.1 | 7 | 3 | 1: .43
May | | | | | | | | | | | | | |
1 | 111- 120 | 2.3.2.1 | 8 | 3 | 7 | 2 | 6 | 5 | 4 | 7 | 1 | 9 | 1 | 1: .11
3 | 121- 130 | 1 | 8 | 5.6.3 | 7 | 4.5.2 | 6 | 5 | 4 | 7 | 1 | 8 | 2 | 1: .25
4 and 5 | 131- 140 | 3.2.1 | 8[1] | 3 | 7 | 2 | 6 | 5 | 4 | 7 | 1 | 9 | 1 | 1: .11
5 | 141- 150 | 1 | 8 | 4.3 | 7 | 2 | 6 | 5 | 4 | 7 | 1 | 9 | 1 | 1: .11
--------+----------+---------+---------+-----------+-------------+-----------+-----------+-----------+-------------+-----------+-----------+---+---+--------
| | | | | | | | 2.3.4.5 | | | 1.2.3.4.5 | | |
| | 2.3.4 | 6.7.8.9 | 3.4.5 | 4.5.6.7.8.9 | 6.7.8.9 | 1.2.3.4.5 | 6.7.8 | 3.4.5.6.7.8 | 5.6.7 | 6.7.8.9 | | |
| +---------+---------+-----------+-------------+-----------+-----------+-----------+-------------+-----------+-----------+---+---+--------
6 | 1- 10 | 2 | 6 | 3 | 4 | 6 | 3.2.1 | 6.2 | 5.6.7.8.3 | 5 | 6.1 | 6 | 4 | 1: .67
========+==========+=========+=========+===========+=============+===========+===========+===========+=============+===========+===========+===+===+========
[Footnote 1: End of series on May 4.]
By reading downward in any particular column of results, one obtains a
description of the changes in the animal's reaction to a particular
setting of the doors. Thus, for instance, in the case of setting 1,
which was presented to the animal in trials numbered 1, 11, 21, and so
on to 141, it is clear from the records that no definite improvement
occurred. But oddly enough, in the case of setting 10, which presented
the same group of open doors, almost all of the reactions are right in
the lower half of the column. For setting 2, it is evident that mistakes
soon disappeared.
Comparison of the data of table 1 indicates that the number of correct
first choices is inversely proportional to the number of doors in use,
while the number of choices made in a given trial is directly
proportional to the number of doors in use.
During the first week of work on this problem, Skirrl improved markedly.
His performance was somewhat irregular and unpredictable, but on the
whole the experiment seemed fairly satisfactory. Cold, cloudy, or rainy
days tended to diminish steadiness and to increase the number of
mistakes. Similarly, absence of hunger was unfavorable to continuous
effort to find the right box.
The period of confinement, as punishment for wrong choices, was
increased from thirty seconds to sixty seconds on April 26. But there is
no satisfactory evidence that this favored the solution of the problem.
Work on May 4 was interrupted by a severe storm, the noise of which so
distracted the monkey that he ceased to work. Consequently, observations
were interrupted on the completion of trial 132, and on May 5, the
series was begun with setting 3. On this date, eighteen trials were
given in succession, and in only one of them did a mistake occur. Since
the ten trials numbered 133 to 142 were correct, Skirrl was considered
to have solved problem 1, and systematic training was discontinued.
On the following day, as a measure of the extent to which the animal had
learned to select the first door at the left no matter what its position
or the number of doors in the group presented, a control series was
given in which the settings differed from the regular series of
settings. These supplementary settings are presented at the bottom of
table 1 together with the records of reaction in ten trials.
Since in only six of these ten control settings was the first choice
correct, it is scarcely fair to insist that the animal was reacting on
the basis of an ideational solution of the problem. Rather, it would
seem that he had learned to react to particular settings. A careful
study of all of the data of response, together with notes on the varied
behavior of the animal during the experiments, justifies the statement
that Skirrl's solution of problem 1 was incomplete and unreliable. It
was highly dependent upon the particular situation, or even the
particular door at the left end of the group, and slightly if at all
dependent upon anything comparable to the human idea of first at the
left of the group.
This particular series of observations has been described and discussed
in some detail in order to make the chief points of method clear. It
will be needless, hereafter, to refer explicitly to many of the
characteristics of reaction or to the important points in the
construction of tables which have been mentioned.
A graphic representation of Skirrl's learning process in problem 1 is
presented in figure 18. The irregularities are most striking, and fairly
indicate the erraticness of the animal. The curve is based upon the data
in next to the last column of table 1, that is, the column presenting
the errors or wrong first choices in each series of trials.
Unquestionably, the form of such a curve of learning should be
considered in connection with the method or methods of selecting the
right box employed by the animal during the course of experimentation.
It appears from an analysis of the behavior of Skirrl in problem 1 that
there developed a single definite and persistent method, namely, that of
going to one box in the group, and in case it happened to be a wrong
one, of choosing, on emergence from it, the next toward the right end of
the group, and so on down the line. Having reached the extreme right
end, the tendency was to follow the side of the reaction-chamber around
to the opposite end and to enter the first box at the left end of the
group, which was, of course, the right one. This method appears, with
certain slight variations, in approximately ninety per cent of the
trials which involved incorrect choices. Thus, in the case of trials 121
to 130, of which eight exhibit right first choices, the remaining two
exhibit the method described above except that the final member at the
right end of the group was in each case omitted.
[Illustration: FIGURE 18.--Error curves of learning for the solution of
problem 1 (first box at left end).]
On the whole, Skirrl's behavior in connection with this problem appears
to indicate a low order of intelligence. He persisted in such stupid
acts as that of turning, after emergence from the right box, toward the
right and passing into the blind alley I, instead of toward the left,
through G and H, to D. In contrast with the other animals, he spent much
time before the closed doors of the boxes, instead of going directly to
the open doors, some one of which marked the box in which the reward of
food could be obtained. It is, moreover, obvious that his responses, as
they appear in table 1, are extremely different from those of a human
being who is capable of bringing the idea of first at the left end to
bear upon the problem in question.
_Problem 2. Second from the Right End_
Following the series of control trials of problem 1 given to Skirrl on
May 6, a period of four days was allowed during which the animal was
merely fed in the boxes each day. This was done in order that he should
partially lose the effects of his previous training to choose the first
box at the left before being presented with the second problem, the
second box from the right.
On May 11 regular experimentation was begun with problem 2. Naturally
the situation presented unusual difficulties to the monkey because of
his previously acquired habit, and on the first day it was possible to
give only five trials, in all except the first of which Skirrl had to be
aided by the experimenter to find the right box. He persistently, as
appears in the first line of records of table 2, entered the first box
at the left. The series was continued on May 13, but with very
unsatisfactory results, since he apparently had been greatly discouraged
by the unusual difficulties previously met. Only four trials could be
given, and in these the showing made was very poor. It is noteworthy,
however, that in trials 6, 7, and 8, May 13, there was no marked
tendency to choose the first box at the left. Thus quickly had the force
of the previous habit been broken.
For problem 2, the total number of open doors in the ten settings is
fifty, as appears from the data on page 18, and as ten of these fifty
open doors may be defined as right ones, the expected ratio of right to
wrong first choices in the absence of previous training is 1 to 4. The
actual ratio for the first series given in problem 2 is 1 to 8, while in
the second series it is 0 to 10.
On the morning of May 13, work was interrupted in the ninth trial by
what seemed at the moment a peculiarly unfortunate accident, but in the
light of later developments, an incident most fruitful of valuable
results.
Skirrl, in trial 9, directly entered box 1. Since this was not the right
box, he was punished by being confined in it for ten seconds. While in
the box he howled and when the entrance door was raised for him to
retrace his steps, he came out with a rush, showing extreme excitement
and either rage or fear, I could not be sure which. At intervals he
uttered loud cries, which I am now able to identify as cries of alarm.
Repeatedly he went to the open door of box 1 and peered in, or peered
down through the hole in the floor which received the staple on the
door. He refused to enter any one of the open boxes and continued, at
intervals of every half minute or so, his cries. For thirty minutes I
waited, hoping to be able to induce him to complete the series of
trials, but in vain. Although it was obvious that he was eager to escape
from the apparatus, he would not enter any of the boxes even when the
exit doors were raised. Instead, he gnawed at the door (12 in fig. 17)
to the alleyway D and attempted to force his way through, instead of
taking the easy and clear route to the alleys, through one of the boxes.
His behavior was most surprising and puzzling. Finally, I gave up the
attempt to complete the series and returned him to his cage by way of
the entrance door to the response-compartment E.
I then entered the apparatus to seek some explanation of the animal's
behavior, and my search was rewarded by the finding of two sharp pointed
nails which protruded for an inch or more in the middle of the floor of
box 1. My assistant, who had been charged with the task of installing
the locks for the several doors, had used nails instead of screws for
attaching staples underneath the floor and had neglected to clinch the
nails. Skirrl, in the dim light of the box, doubtless stepped upon one
of the nails and inflicted a painful, although not serious, injury upon
himself. It was impossible for him to see clearly the source of his
injury. He was greatly frightened and expressed the emotion most
vigorously. His behavior strongly suggested a superstitious dread of
some unseen danger. It may be that the instinctive fear of snakes, so
strong in monkeys, was partly responsible for his response.
The first result of this accident was that more than two weeks were
lost, for it was impossible, during the next few days, to induce the
animal to enter any of the multiple-choice boxes voluntarily. From May
14 to May 24, I labored daily to overcome his newly acquired fear. The
usual procedure was to coax him through one box after another by
standing at the exit door with some tempting morsel of food. After
several days of this treatment, he again trusted himself to the boxes,
although very circumspectly and only when both entrance and exit doors
were raised. Not until May 24 was it possible to resume regular
experimentation, and on that day it was found necessary to indicate the
right box by raising the exit door slightly and then immediately
lowering it. Trials in which this form of aid was given are indicated in
table 2 by a star following the last choice.
Gradually, Skirrl regained his confidence in the apparatus and began to
work more naturally. For a long time he would not stand punishment, and
it was necessary for the experimenter to be very careful in locking the
doors, since the sound of the bar sliding beneath the floor often
frightened and caused him to quit work. Day after day the tendency to
peer through the holes in the floor at the entrance to the boxes
rendered it clear that the animal feared some danger from beneath the
floor. This behavior was so persistent that much time was wasted in the
experiments.
On the last day of May, punishment by confinement for ten seconds in
wrong boxes was introduced, but since this tended to discourage the
monkey, there was substituted for it on June 1 the punishment of forcing
him to work his way out of each wrong box by raising the entrance door
which had been closed behind him. This he could fairly readily do, and
his stay in a box rarely measured more than ten seconds.
As a variation in the mode of procedure, confinement for thirty seconds
was tried on June 5, but it worked unsatisfactorily and had to be
abandoned. During this series, the animal was startled by the sound from
one of the sliding bars under the floor, and in the sixth trial he
refused to work.
As improvement was very slow, varied modes of rewarding and punishing
the animal were tried in the hope of discovering a means of facilitating
the work. Among the former are the use of banana, grapes, peanuts, and
other eagerly sought foods in varying quantities, and in the latter are
included periods of confinement ranging from ten seconds to sixty
seconds. In the end, confinement of about thirty seconds, combined with
a small quantity of food which was much to the monkey's taste, gave most
favorable results.
All this time Skirrl's attention to the task in hand was seldom good. He
was easily diverted and even when extremely hungry, often stopped work
in the middle of an early trial, yawned repeatedly and finally sat down
to wait for release from the apparatus.
The results obtained during the long continued trials with this animal
in problem 2 are presented in table 2, which differs from the previously
described table, first, in that several of the trials are followed by an
asterisk to indicate that aid was given by the experimenter, and second,
in that two additional columns, headed, respectively, R and W, are
presented. These give the right and wrong first choices for each day,
whereas the two columns preceding them give the same data for each
series of ten trials. Similarly, the ratio of right to wrong choices is
presented for each day in table 2, instead of for each series of ten
trials as in table 1.
From the results of table 2, several peculiarly interesting facts
appear. In the first place the influence of the habit of choosing the
first box at the left disappears with surprising suddenness, and in the
second place, there are remarkable contrasts in the results for
different settings as they appear in their respective vertical columns.
Thus, in the case of setting 1, after the first trial mistakes became
relatively infrequent, whereas in setting 6, which involved the same
number of doors, mistakes continued to be the rule until nearly a
thousand trials had been given. The most likely explanation of this
difference is that for some reason the animal avoided box 9.
The _reactive tendencies_, or better, the _methods of reaction_ which
manifested themselves during this long series of observations may be
described as follows: (a) choice of the first box at the left; (b)
random choice with tendency to choose first, a box near the middle of
the group; (c) choice of first box at the right followed by the one next
to it on the left; (d) direct choice of the right box.