THESIS I

._..___-_ 7-, ' 1.; ‘ ..._._.-...__

This is to certify that the
thesis entitled
.Ln experimental analysis of

abnormal fixation in the rat

presented by
Mr. Loo Ia Gladin

has been accepted towards fulfillment
of the requirements for

I. A. degree mm

Maj! professor 2

 

DateMZ/l/ffj
/ / /

 

 

on "M

AN EXPERIRENTAL ANALYSIS OF

ABNOHhAL FIXATIUN In THE HAT

By
Leo L. Gladin

A THESIS

Submitted to the School of Graduate Studies of Hichigan
State College of Agriculture and Applied Science in
partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Psychology \

1953

THESIS

The author wishes to express his sincere
thanks to Dr. M. Ray Denny, under whose direction
the present investigation was undertaken. His
assistance and guidance throughout the experiment

and thesis preparation has been deeply appreciated.

- .~\ ”'x *1 N'-
0 t

LB‘iclkitj‘E

List of Tables

TABLE OF CONTENTS

List of Figures

I.
II.
III.

IV.

VI.
VII.
VIII.

Introduction
Statement of the Problem

Methods and Procedure

A. Apparatus
B. Subjects
C. Procedure

Presentation of Results
Discussion of Results
Summary

Appendix

Bibliography

Page

ii

11

15
15
19
25
33
35

TABLE I.

TABLE II.

TABLE III.

TABLE IV.

LIST OF TABLES

MEDIAN LATENCY BEFORE FALL MINUS
MEDIAN LATENOY AFTER FALL ON DAY
ONE AND DAY EIGHT OF TPE INSOLU-
BLE PROBLEM.

COMPARISON OF LEAN LATENCIES OF
TI-E EIGHT! NON-LEARNING RATS ON
TWO OR THREE SUCCESSIVE PRESEN-
TATIONS OF THE NEGATIVE CARD ON
THE)HABITUAL SIDE. (SOLUBLE PROB-
LEM . 7

COMPARISON OF MEAN LATENCIES OF
THE EIGHT NON-LEARE‘IING RATS ON
PRESENTATION OF THE POSITIVE
CARD ON THE HABITUAL SIDE AFTER
ENTER AND AFTER FALL. (SOLUBLE
PROBLEM).

COMPARISON OF MEAN LATENC IES OF
THE EIGHT NON-LEARNING RATS ON'
TWO OR THREE SUCCESSIVE PRESENi-
TATIONS OF THE POSITIVE CAM) ON
THE)HABITUAL SIDE. (SOLUBLE PROB-
IEL‘I o

55

35

36

36

LIST OF FIGURES

Figure 1.:
Ideal expectancy pattern illustrating
variation in latency times after entry
and after falling to the net.

Figure 2.:
Illustrating certain typical sequential
probabilities of entry and reward dur-
ing the insoluble problem..

Figure 3.:
Illustrating.the sequential probabilities
of entry and reward after a fall to the
net during the soluble problem, for a rat
fixated on the right.

Figures 4,,5.:
Atypical response pattern of one of the
fixated rats..

Figure 6.:
Typical orientation response toward the
experimenter.

Figures 7 through 15.:
A generalization procedure used in break-
ing fixation.

ii

37

38

39

40

4o

41

INTRODUCTION

In a series of experiments using the Lashley jumping
apparatus, N. . F. Maier and his students (11) have succeeded
in producing in rats a response pattern of such persistence
as to warrant the designation fixation. While stereotypy of
response has been reported by many other investigators,

(h, S, 6) it is Meier's contention that the rigid behavior
pattern found in rats subjected to his experimental designs
is qualitatively different from that ordinarily subsumed
under the postulates of learning theory. Further, he has
held that the behavior patterns he has described should be

considered as abnormal. fixations which result from the

 

 

Operation of a dynamic mechanism of frustration and lead to

behavior without a goal.

 

The essential feature of the experimental design held
responsible for "abnormal fixation" is the insoluble problem;
the Lashley jumping apparatus was originally employed in

visual discrimination problems, in which it was observed

that some rats failed to make the required discrimination,

but fixated a position response instead. It appeared to Maier

that the jumping apparatus could possibly be used to create

neuroses in rats in a discrimination problem setting (I).

The general procedure employed in producing abnormal fixa-
tions is as follows:

The rat is trained to jump‘from a small stand to one of
two windows which give access to a platform upon which
food is diSplayed. When the rat learns to jump a distance
of about eight inches, it is trained to knock down a card
placed in a window. The cards which cover the window are
readily discriminable, one being black with a small white
circle in the center, the other being white with a small black
circle in the center. The rat is given experience in jumping
to both windows during the initial training period, care being
taken to prevent the development of position habits or selec-
tive preferences of one card over the other. After this
training, during which both windows are unlocked and give
access to food on each jump, the situation is altered so
that both windows are randomly locked fifty percent of the
time. No matter which card is selected or which position is
assumed, the rat encounters a locked window on five of the
ten daily trials. The rat usually refuses to jump after a
few trials, and is then forced (goaded) to jump by means
of electric shock, taps on the tail with a small stick, or
'by directing a jet of air under pressure at its posterior.
Any attempts to jump in a direction other than the windows

(abortive jumping) is discouraged.

After a brief period of variability, the rat begins to
express a consistent pattern of reSponse, usually in the form
of a position stereotype: the rat jumps to either the left
window or the right without further variation. This phase
of the experiment is carried on for from eighty to ZhO trials.
The problem is then "made soluble": one of the cards is con-
sistently left unlocked, the other is always locked. The
positive card is randomly shifted from one window to the
other throughout the ten daily trials. The arrangement is
so planned, however, that the positive card will be offered

on the side of the rat's position stereotype fifty percent

of the time, so that the number of rewarded jumps remains
constant, as does the number of encounters with a locked win-
dow and a subsequent fall to the net. If the rat persists

in the position stereotype for two hundred trials of the
soluble problem, it is considered to have abnormally fixated
its reSponse due to having exceeded its frustration tolerance
during the experiment. The criterion of two hundred trials
was established by Maier as a sufficient number of trials in
which to make the required discrimination, since it was

found that rats rarely learned if continued in the problem
beyond this number (11).

In a combined report of three studies by Maier and his

students (11, p. 32), 3h rats were subjected to the insoluble
problem situation; only ten (29.h percent) learned the sub-

sequent discrimination, while Zh (70.6 percent) fixated.

Such results have led to the proposal of a qualitative differ-
ence between fixations and habits (11, p. 89-91). Essentially,
it is assumed that frustration resulting from the problem
situation brings about a bimodal distribution of individuals;
"(a) those whose position responses are modified in an average
of eighty trials when an opportunity to learn was given, and
(b) those whose position responses were unmodified after two
hundred trials" (11, p. 89-90).

The contention is that frustration serves to Split a
population; the evidence for the operation of frustration
in any animal is the maintenance of a position reSponse to
a criterion of two hundred trials of the soluble problem.
It is assumed that the introduction of the insoluble problem
results in increments of frustration for the susceptible ani-
imals, which separates them qualitatively from such animals
as are not frustrated by the situation.

The specific nature of this qualitative difference is
further illustrated by a comparison between frustration-
induced and motivated behavior:

\

...a frustration—induced reSponse is stronger than a
response normally learned, despite the fact that the
learned reaponse is reinforced by reward and the frus-
trated response is not." (11, p. 91)

This conclusion is based on a comparison between rats

subjected to the insoluble problem and those given the following
treatment: Rats are trained to a position habit under one hun—

dred percent reward for from eighty to 260 trials, and are

then presented with a card discrimination to be made. Most
rats learn to alter their responses rather rapidly during

the discrimination problem, and only a few fixate the habitual
reSponse. These few are considered to have an initially low
frustration tolerance (11, p. 78).

Klee (10) compared the performance of two groups of rate,
one of which was required to establish a position reSponse
according to its natural preference to obtain one hundred
percent entry and reward, the other being subjected to the
insoluble problem, which randomly yielded fifty percent entry
and reward. After 160 trials under these separate conditions

f training, both groups were required to alter their position
responses in favor of a symbol discrimination. It was found
that 83.h percent of the rats which were allowed one hundred
percent entry learned to make the required symbol discrimina-
tion, while 75 percent of the rats exposed to the insoluble
problem failed to make the required discrimination. On the

basis of these results, Klee differentiates fixations from

 

habits. It is his contention that rats which receive one
hundred percent reinforcement should show more resistance to
alteration of reaponse, while rats which receive only fifty
percent reinforcement should have proportionately weaker habit
strength in the pattern of behavior expressed, and should
therefore more readily abandon their position responses when
presented with a symbol discrimination leading to one hundred

percent reward.

The point of view which equates habit strength with
consistency of primary reinforcement has been negated by
the results of many investigators. In a review of the
literature on partial reinforcement, Jenkins and Stanley
(9) have pointed out the efficacy of such reinforcement
under a variety of experimental conditions.

A second objection to the dichotomy of fixation and REBEL.
may be based on the amount of cue differentiation which is
present at the outset of the discrimination problem. A con—
trol group which receives one hundred percent reward for 160
trials is abruptly exposed to fifty percent non-entry followed
by a fall to the net and non-reward. Rats so trained are con-
fronted with a distinctly new learning situation, one in which
stimulus-generalization decrement should occur and less compe—
tition from previous learning should obtain. Entry and re-
ward is always gained by jumping to a discriminable card; a
fall to the net follows a jump to a different but equally
discriminable card. Alteration of response is therefore
greatly facilitated.

On the other hand, for these rats which have been ex-
posed to 160 trials of the insoluble problem, the subsequent
transition to a soluble discrimination problem is a subtle
one. The conditions of fifty percent reward, randomly re-
ceived, still obtain. It is assumed by Maier that the change

in the problem which makes one card consistently rewarding

should lead to alteration of response by a goal-oriented

rat. However, it has been pointed out by Eglash (2) that

the rat does utilize the negative card cues on the habitual
side by learning to make an abortive jump to this card. This

reSponse should reduce the punishing effects of striking the

locked window. Thus it appears that the rat uses the visual

card cues added by the soluble problem, but uses them in a

manner consistent with the pattern of response adopted during
the insoluble problem.

While Higard (8, p. 303-30h) has pointed out that this
subtle change in the problem confronting the rat might be a
crucial factor in the maintenance of the habitual reaponse,
Maier has minimized its importance. In defense of his position,
he introduces the results obtained with a group of rats for
whom the insoluble problem was followed by a locked window,
consistently blocked by the same card, on the habitual side.
In this group, there was less response-alteration than in any
other (11, p. hO). One hundred percent punishment produces
Vonly generalization decrement, however, unless the animal
shifts to the other side; then, as Maier found, in a few non-
fixated animals discrimination learning is very rapid.

In brief, it might be assumed that the number of fixations
which occur among a group of animals depends upon the nature
of the cue conditions to which the animals are exposed. One

may grant that frustration occurs in the insoluble problem,

but in a situation where at least two possible directions of
response exist, the fact that rats respond either by learning
a discrimination imposed by the experimenter or by maintaining
the habitual response does not necessarily imply that frustra-
tion has split the population.

The ultimate criterion of abnormality in this situation
must inevitably be a tautological one; failure of the animal
to learn the problem set by the experimenter is sufficient
evidence of susceptibility to frustration, while learning of
that problem is evidence of high stability. While Maier (11,
p. 28) suggests that the fixated animal may be "responding to
some goal that is not under the experimenter's control", he
prefers to consider the behavior as due to the Operation of
a frustration mechanism, which results in a "freezing" of a
sample of behavior. It is assumed by Maier that the typical
position response marks the inception of behavior without a
goal. He states that this response "has no adaptive value in
the sense that it is adequate to the situation or in the sense
that it is superior to any number of possible responses" (ll,
p. 27). However, it would appear more parsimonious to assume
that the same factors which lead to learning and retention of
reaponse also lead to fixation. In this framework, the effi-
ciency with which the animal learns to utilize the cue environ-
ment of the insoluble problem situation would determine whether

or not the habitual response is perpetuated.

The question of normal and abnormal, as designations
to be legitimately applied to rats, may be raised when the
phylogenetic level of the animal is considered. Hamilton
(6) has shown that the perseverative behavior of rats is
much more pronounced than that of humans exposed to roughly
comparable situations. It is noted by Hebb (7) that the
position habit "is a constant nuisance in animal experiments"
(p. 92). He suggests that the visual cues from the more dis--
tant, and therefore least variable, portion of the environment
tend to dominate behavior in rats, and illustrates this point
in the following manner:

"Teach a rat to jump from a small platform to another

one near by, the second platform is just large enough

for him to land on safely, and holds food. After he
has made ten jumps, move the second platform through

90 degrees. The rat hesitates, shows disturbance, but

finally Jumps -- into Space, in the former direction

of food. (p.93)

The naive observer would be inclined to label such a
performance as abnormal. It is by isolating the specific
factors upon response elicitation depend that such behavior
Inay be more parsimoniously explained without resorting to
a diagnosis of abnormality.

In Maier's orientation, the fall to the net functions
solely as "punishment". It has been noted by Mowrer, (15)
find Wolpe, (17) that the antecedent of such falls to the net
1J3 goading by air-blast, electric shock, or prodding with a
Silick. The rat thus is involved in a situation with both

Eijpproach (food reward), and avoidance (goading), components.

10

The single response develOped during the insoluble problem
subserves two primary drives. When the problem is made
soluble, a differential resistance to the two cards is noted.
Maier says:

"The differtial reactions to the positive and negative

cards, shown both by the degree of resistance and the

frequency of abortive jumps, reveal that the fixated

rat actually has learned which card punishes and which

does not. In other words, the rat has made the required

differentiation but is unable to practice the required
response. This property of the fixation makes it appear
as a form of compulsion. The rat executes an unadaptive
reaponse even when it knows better." (p. MB)

The implication that the rat "knows better" has not won
wide acceptance. Hilgard (8) suggests that the alternative
window is no longer under consideration by the rat, so that
the problem is now reduced to one window, without the alter-
native possibility so obvious to the experimenter.

Extrapolating from the conclusions drawn from fixation
experiments, Maier and Ellen suggest "that learning theory
must exclude frustrated behavior, that social psychology must
distinguish between groups organized around frustration, and
groups organized around common goals" (13, p. M36). Such a
statement demands a radical reorientation in psychological
theorization. Further investigation is necessary to determine

whether "when frustration occurs, the behavior expressed is

an and rather than a means to an end" (13, p, 3A6),

ll

STATEMENT OF THE PROBLEM

In the literature on fixation, a number of factors
which tend to enhance the stereotypy of reaponse sequences
have been considered. It has been found that designs which
necessitate a jump response by the rat (5), or which involve
punishment (h), or a heightened hunger drive (3) result in a
greater stereotypy of response. The fixation problem as de-
signed by Maier includes all these factors.

Klee (10) has considered these factors under the heading,
"degree of involvement of the organism in the total situation",
stating, however, that:

”In and of itself, the degree 2f involvement is not to
”be thought of as a cause of abnormal behavior. Rather,
it must be considered as a condition essential in most
situations for the production of such reactions only
insofar as it would enable frustration to occur. From
this, it would be expected that, the greater the degree
of involvement, the greater the effect the frustrating
situation is likely to have on the animal as measured
by the abnormal reaction to the situation. Also, the
frustrating situation need not be so severe where the
involvement is greater." (p. S)

 

In other words, the adherents of frustration theory
-minimize those characteristics of the design which have been
found in learning theory to lead to increased habit strength.
The point of departure which makes fixation abnormal is
posited as frustration. The question may be raised, then,
whether the situation is consistently frustrating to the rat,

as far as may be ascertained from the behavior of the animal,

12

apart from its success or failure in learning the problem
set by the experimenter. Is it possible that the rat may
be goaléoriented without learning that problem?

In an experiment by Klee (10), the sole incentive to
performance in the insoluble problem, and the subsequent
soluble problem was a food reward. That the animals which
fixated under this condition were 223 goal oriented appears
to be a tenuous assumption. That consummatory responses
are merely incidental consequences of successful entry to
the food platform may also be questioned from the standpoint
of the behavior of rats observed in the preliminary phase
of the present study. Of twenty rats, only one failed to
eat consistently when it gained access to the food platform.
Such consummatory responses were often of such vigor that
it was necessary to forcibly separate the rat from its food.
Furthermore, it was observed that rats, after falling to the
net, appeared to approach the experimenter, contrary to ex-
pectation, since it might be assumed that an avoidance reSponse
would be more apprOpriate subsequent to the "punishment" of
falling to the net.

Analysis of the data sheets used in previous fixation
experiments employing the Lashley jumping apparatus revealed
the possibility that the rat might learn to use a temporal
or sequential cue during the insoluble problem. Furthermore,

this cue, being uncontrolled, would continue to be effective

13

in the ensuing soluble problem, thus serving to perpetuate
the pattern of response earlier developed.

In the soluble problem, four standard patterns (see
Figure 2, 3, Appendix) of ten trials each, are repeated
over a period of two hundred trials. In the first pattern,
after a fall to the net, the probability of an entry and
food reward on the succeeding jump is three times the prob-
ability of a fall to the net. In two patterns, the probabil-
ity of entry after a fall is three to two, and on the fourth
pattern, the probabilities are even, two jumps following
falls to the net result in entry, and two jumps following
a fall result in a second fall. If an expectancy of reward
were to deveIOp on this basis, it would be randomly rein-
forced in the following fashion: the animal would fall once
or twice, but no more than three times consecutively, with
the next jump being followed by entry and food reward. Con-
versely, after entry and reward, the probability of a fall
to the net on the following trial is increased in a similar
ratio.

Since the experimenter assumes an insoluble problem
situation when he varies the visual cues, 1.9., the cards
which block the two windows, his primary concern in the sub-
sequent soluble problem is focused on the ability or inability
of the rat to utilize the visual cues of camdsymbol and bright-
ness to gain access to one hundred percent reward. The assump-

tion here made is that the critical cues from the standpoint

13

in the ensuing soluble problem, thus serving to perpetuate
the pattern of response earlier developed.

In the soluble problem, four standard patterns (see
Figure 2, 3, Appendix) of ten trials each, are repeated
over a period of two hundred trials. In the first pattern,
after a fall_to the net, the probability of an entry and
food reward on the succeeding jump is three times the prob-
ability of a fall to the net. In two patterns, the probabil-
ity of entry after a fall is three to two, and on the fourth
pattern, the probabilities are even, two jumps following
falls to the net result in entry, and two jumps following
a fall result in a second fall. If an expectancy of reward
were to deveIOp on this basis, it would be randomly rein-
forced in the following fashion: the animal would fall once
or twice, but no more than three times consecutively, with
the next jump being followed by entry and food reward. Con-
versely, after entry and reward, the probability of a fall
to the net on the following trial is increased in a similar
ratio.

Since the experimenter assumes an insoluble problem
situation when he varies the visual cues, i.e., the cards
which block the two windows, his primary concern in the sub-
sequent soluble problem is focused on the ability or inability
of the rat to utilize the visual cues of cardsymbol and bright-
ness to gain access to one hundred percent reward. The assump-

tion here made is that the critical cues from the standpoint

of the experimenter must also be the critical cues involved
in response elicitation from the standpoint of the rat. If,
however, the rat learns to respond on the basis of uncon-
trolled critical cues of the insoluble problem and these
cues continue to lead to reinforcement during the soluble
problem, then the discrimination problem posed by the ex-
perimenter need not exist for the rat. The expectancy de-
veloped during the insoluble problem continues into the
soluble problem, maintaining the response in progress. An
experiment was designed to test the possibility that such an

expectancy is learned.

METHOD AND PROCEDURE

AIPARATUS:

The standard Lashley jumping appartus was used in the
present study (see photographs, Appendix). The data sheets
contained a representation of the two windows to facilitate
recording of the approximate nature of the jump made on each
trial. Sufficient space was allowed on each sheet to record
the behavior of the rat during every trial. A stOpwatch was
used to record the latency of each response, from the time tne

rat was placed on the stand until a jump was made.

15

SUBJECTS:
Ten adult albino rats approximately 150 days old,
five male and five female, were used. None of the animals

had been previously used for experimental purposes.

PROCEDURE:

The rats were placed on a reduced daily diet of nine
grams of Purina checkers and trained under a mean of nine-
teen hours of food deprivation.

The animals were given preliminary experience in being
handled by the experimenter. After handling, they were
placed on the jumping stand, which was located directly
against the two windows, so that the animals could step
easily to the food platform, where they were fed their en-
tire daily food ration before being removed. After four days
of preliminary training, the rats were required to learn to
jump to the food platform from a progressively greater distance,
care being taken to distribute the jumps to both windows, so

that a position habit could not develop during this phase.
On each trial, the two cards covered one third of each window,
the cards being alternated to prevent a selection of cards
during jump-training. .

When all rats had learned to jump an eight inch distance
to the food platform, the windows were gradually obscured by
the two cards, the rats learning to knock down a card to gain

access to the food. After the rats had learned to jump to

16

windows completely obscured by the cards, they were condi-
tioned to make the required jump to the tap of a pencil on
the jumping stand. This was accomplished by tapping alter-
nately on the stand and on the rat's tail during forty trials
of one hundred percent reward. This procedure was employed
to reduce the amount of painful goading required to elicit
the jump response during and after the insoluble problem.

The animals were then given eighty trials in the insol-
uble problem, in which both windows were randomly locked
fifty percent of the time. No matter which window was chosen,
no more than five of the ten daily trials could lead to en-
try. The mean ratio of entries following falls, as compared

to falls followed by second falls was twelve to five, dis-
tributed in the following fashion: On the first day, the
number of entries following a fall was four, the number of
falls following a fall were zero. On the following two days,
the number of entries following a fall was three, the number
of falls followed by a second fall to the net was one. On
the next two days, the number of entries following a fall was
three, the number of falls followed by a second fall was two.
on the sixth and seventh days, the number of entries following
a fall was two, the number of falls followed by a second fall
to the net was also two. On the eighth day, the simple al-
ternation of falls and entries prevailed as on the first day

Of the insoluble problem, to allow a comparison of performance

17

at the beginning and end of the insoluble problem, as judged
by the differences in latency times on the two days.

After falling, the rats were allowed to remain in the
net for about thirty seconds, while the experimenter changed
cards or simulated a change of cards. Upon gaining access
to food, the rats were allowed to eat for about thirty seconds
before being replaced on the jumping stand. The apparatus
was so arranged that the experimenter could only pass the
stand on the right side when going to the rear of the stand
to manipulate the cards.

A record was kept of the latency of reSponse, measured
from the time the animal was placed on the stand until a
jump was made. Activity of the rat on the jumping stand,
consummatory responses after entry to food, the type of re-
Sponse made, and location of the rat in the net at the time
of being picked up, were also recorded.

The soluble problem was based on the four standard pat-
terns reported on page thirteen, with probabilities of three
to one on the first pattern, three to two on the following two
patterns, and two to two on the fourth pattern. The mean
ratio of entries following falls as compared to falls fol-
lowed by second falls was eleven to seven. The four patterns
(see Figure 3, Appendix, for illustration) were rotated
throughout the two hundred trials of the soluble problem.

Each animal was allowed sixty seConds free time on the

jumping stand. Failure to jump within that time was followed

18

by thirty seconds of tapping on the stand. After a total of
ninety seconds on the stand, the animal was subjected to
actual goading by being tapped on the tail with a pencil un-
til a jump response was made.

Animals who learned the discrimination problem to a
criterion of fifty successful entries were presented with a
ckuaice between the discriminated positive card, which was
rwuidomly locked fifty percent of the time, and.an Open window,
'with.no card displayed. This procedure was employed to test
the possibility that failure to jump to an open window was

:not exclusively a characteristic of only such rats as fixated.

19

PHESEQTATION OF RESULTS

All animals continued to eat throughout the experiment.

All ten rats developed position stereotypes, and at the
<3onclusion of the insoluble problem were presented with a
<3ard.discrimination, permitting one hundred percent reward.
wao rats learned the card discrimination, one after forty
‘trials, the other after seventy trials. One rat remained
‘variable throughout the experiment. Seven rats reached the
fixation criterion of two hundred trials.

Two fixated rats developed unique patterns of response,
'which.took the form of exerting extreme force during jumping.
Owing to the design of the problem, this reSponse resulted in
.fifty percent successful entries, and thus might constitute
a ferm of solution to the rat. One animal abandoned this
iform of resyonse after 120 trials in favor of an abortive
SPesponse to the negative card; the other maintained extreme
iforce in jumping throughout the problem.

The difference between the median latency on the jumps
following a fall to net and the median latency of the jumps
followed by entry was computed for the first ten trials and
the last ten trials of the insoluble problem. On the first
ten trials, this difference was in the direction of longer

latencies for the jumps following falls. On the last ten trials,

the difference was reversed, and the difference between these

2O

differences is significant at the one percent level of
confidence (see Table I, Appendix). Thus it may be inferred
that learning to the suggested cue took place during the in—
soluble problem.

That such learning affected the behavior of the animals
during the soluble problem is indicated by the latencies of
the fixated rats when confronted with the nepative card on
two or more successive trials. It was found throughout the
soluble problem that the animals showed decreased delay in
jumping to the negative card after a fall to the net. In other
words, a fall to the net appears to result in a faster jump
on the following trial, even when the negative card has appar-
ently been discriminated from the positive, in terms of longer
latency times and abortive style of jumping.

For the eight non-learning animals, the mean time of the

first jump to the negative card was h9.6 seconds; the mean

time of the following jump to the negative card was 3h.2 seconds
(see Table II, Appendix). This difference is significant

at the 0,1 percent level of confidence. It was also found

that animals jumped more rapidly to the positive card after
falling to the net than after eating. Mean latency of the
second of two successive jumps to the positive card was 28.1
seconds; mean latency of jumps to the positive card following
falls to the net was 19.7 seconds. This difference was also
significant at the 0.1 percent level of confidence (Table III,

Appendix). A comparison of the latency of reSponse in those

21

daily trials in which the positive card was presented two or
more successive times shows a greater latency on the second
presentation. Mean latency on the first jump to the positive
card was 22.6 seconds; as noted above, mean latency on the
second successive jump to the positive card was 28.1 seconds.
This difference was significant at the one percent level of
confidence (Table IV, Appendix). It should be specifically
noted that the first jump to the positive card is in most
instances preceded by a trial which resulted in a fall to the
net. Figure I (Appendix) illustrates these latency comparisons
in a typical pattern of ten trials taken from the record of
one rat.
With succeeding trials, the rats, after falling to the

net, also appeared to learn a particular approach reSponse

to the experimenter. The experimenter always appeared from
behind the apparatus on the right side after changing the
cards (or simulating a change of cards) on each trial. Prior
to the forty-first trial of the soluble problem, the mean
number of times the rat was located at the right edge of

the net at the time of being picked up was lh.h times at the
right edge and hl times elsewhere in the net. After the
forty-first trial, the mean location at the time of being
picked up was 53.5 times at the right edge and 16.5 times else-
Where in the net. Thus, it would appear that the rat is goal-

oriented toward the experimenter and toward the jumping stand

subsequent to a fall to the net.

22

While it is known that a fixated rat will jump to a
locked window on the habitual side in preference to an open
one (no card present), this test has apparently not been made
with rats who have learned the problem. After fifty success-
ful entries, the two rats which learned the discrimination
were presented with a choice between the positive card, which
was now locked fifty percent of the time and an open window.
One rat jumped to the card 28 times out of thirty, falling 1h
times. When presented with a choice between the negative
card and an open window, the rat jumped to the open window
five out of six times. Presented with the positive card
again, the rat chose the card six times out of seven, falling
three times. The other rat chose the open window in preference
to the positive card, jumping with little delay. However, when
presented with a choice between the negative card and an open
window, the rat required goading on two successive trials to
elicit response, although it jumped to the open window on
each trial. Thereafter the rat jumped easily to the open
window in the presence of either card.

One rat (Figure h, 5, Appendix) began to abort more
strongly to the negative card during the closing trials of
the experiment, until it began to fall to the floor beyond
the left side of the net. After two falls to the floor, the
rat began to jump to a cage 2h inches away when confronted

with the negative card on the habitual side, but it did jump

23

to the window on that side when the positive card was displayed.
It was found that even when a negative card was removed from
the window, allowing obvious access, the rat still jumped to
the cage. Evidently, the positive card on the non-preferred
side is sufficient to signify a locked window, since it is
always present on the non-preferred side when the locked
negative card is presented on the habitual side.

Evidence of frustration during the insoluble problem,
as inferred from the behavior of the rats: Two of the rats
in the non-learning group made attempts to escape from the
jumping stand by jumping to the floor early in the problem.
This behavior occurred during one trial in each instance. One
of the animals who subsequently learned the soluble problem
aborted to the left of the left window seven times in succession
on the second day, and three times in succession on the fourth
day. Thus, as far as may be determined from the behavior of.
the rats, the problem may be frustrating to those which learn
as well as to those which do not. It may also be significant
to note that this latter rat continued to show increased
latency of reSponse after falling until it learned the problem.

At the conclusion of the experiment, a method of breaking
fixation was devised and applied to five of the fixated animals.
This was a generalization technique, in which the habitual re-
sponse was gradually moved over to the non-preferred window.

It is best understood by study of the photographs, Figures 7-15,

in the Appendix. The procedure was found effective for

breaking the fixation in all animals in terms of entering
the open window on the non-preferred side. No attempt was
made to utilize the technique in training animals to make

the card discrimination.

DISCUSSION OF RESULTS

While all of the animals continued to eat throughout
the experiment, consummatory behavior is presumably not
sufficient evidence to diaprove the theory of behavior without
a goal. However, when the experimental findings indicate a
sequential relationship exists between those trials which are

rewarded by entry and those which result in a fall to the
net, the frustration-fixation theory is seriously jeopardized.
Since the responses learned to the cues of fall and entry have
received a large number of reinforcements during the insoluble
problem, and are reinforced more often than the positive card
one during the soluble problem, that is, more than fifty
percent of the time while the rat maintains a position pref-
erence, there seems to be no reason why a rat which has learned
approach and avoidance responses to these cues should not
remain fixated.

It is plausible to assume then, that the rat has maxi-
mally employed certain cues during the insoluble problem,
which, owing to the physical design of the subsequent soluble
problem, continue to be effective in maintaining the habitual
response. That the majority of rats exposed to these conditions
do not learn the problem set by the experimenter is thus not

necessarily a sign of abnormality. It may be argued, however,

26

that the rats have not solved gay problem: fixated rats
continue to perform under fifty percent reward, even though

a differential cue exists which can lead to one hundred

percent reward. Such a position demands a certain prescience
on the part of the rat, a certain expectancy about the be-
havior of experimenters. The expectancies developed by the

rat appear to be based on its own responses in a situation,

not considered as a problem per se, but as a maximal adjustment
to a situation to which the animal is exposed. That such an
adjustment occurs during the insoluble problem is affirmed by
Maier, who states however, that "Any adjustment accomplished
must be regarded as purely incidental, and not as a factor

that contributes to the development of fixations" (ll, p. 53).
If such an adjustment is based on uncontrolled cues utilized
by the rat, there would seem to be no valid reason for main-
taining that it "must be regarded as purely incidental".
Further, Maier and Feldman (1h) found that a fixated response
is stronger when the rat is exposed to 160 trials of the in-
soluble problem than when it is subjected to eighty, and that
more fixations resulted from the longer Series of trials. It
was supposed that frustration produced "increments of rigidity"
(11, p. 65) which strengthened the position response. From the
standpoint of the present analysis, the cues which establish

an expectancy of entry after a fall to the net would be more

frequently reinforced during a longer series of trials, thus
both strengthening'the habitual reaponse and increasing the

number of fixations.

27

Since the sequential cue of entry after fall continues
to be reinforcing during the soluble problem, even though a
card discrimination is offered which leads to one hundred
percent reward, the continuation of the habitual reaponse need
not be considered abnormal. Two possibilities of response
exist, one which is arbitrarily defined as learning 3 problem,

the other being labeled fixation. Statistically speaking,

 

the normal response to the problem is fixation, since the
majority of animals maintain the habitual response throughout
the problem. There is no reason to suppose that in a complex
situation all animals should learn the reSponse required by
the experimenter in order to be considered as normal.

The two rats which employed extreme force in jumping
illustrate the point that even fixated rats may not all be—
have alike. Although exposed to the same objective situation
as the others, these rats jumped so vigorously that each
entry invariably carried them to the rear of the food plat-
form. Impact against the locked window was much harder than
that of the other fixated rats. But since extreme force in
jumping resulted in fifty percent successful entries, in a
random order of reinforcement, it could well constitute a
form of solution to the rat. In each case, this reaponse
was also combined with decreased latency of response after
falling to the net.

Failure of fixated rats to jump to the non-habitual

window even when no card obstructs the view of the food platform,

28

has been taken as an indication of a compulsive abnormality

(11, p. 53). It might be expected, then, that rats which

learn the required discrimination, being "normal", could

easily jump to the Open window when circumstances made such

a response "desirable". Yet one of the rats which learned

the discrimination continued to jump at the positive card

in preference to an open window even when such jumps led to

falls to the net on half the trials, as described on page

twenty-two. This rat did not alter its response in favor of

the open window until after a total of 1? falls to the net.

Such behavior indicates that fixation is not necessarily

qualitatively different from a fairly well learned reSponse.
That fixation does not represent a compulsion to jump

to a specific card is demonstrated by the behavior of rats

when the jumping stand is shifted to a point midway between

one of the windows and the blank face of the apparatus (Figure

8, Appendix). Given a choice between the positive card in

the window toward which the habitual reSponse is made, and

a blank wall, the rat jumps against the wall. Even without

a card in place in the habitual window (Figure 9, 10, Appendix),

the rat still jumps against the wall. While such behavior

appears strange, it must be remembered that it is the behavior

of initially naive rats which have practiced a highly specific

response over a period of more than two hundred trials, in a

limited environment. A similar form of behavior may be observed

in the human infant of nine or ten months, according to Piaget:

29

"When the baby is placed between two.pillows and he has
succeeded in finding an object hidden under the right one,
the object can be taken from his hands and placed under
the left pillow before his very eyes, but he will look
for it under the right pillow where he has already found
it once before, as if the permanence of the object were
connected with the success of the former action, and not
witg a system of external displacements in space.” (16,

p. 3 -

In the same sense, it is the rat's response, practiced
over many trials, which has become highly specific to the
situation. Whatevever the cues are which lead to response
elicitation, they are not necessarily the symbol cards which
block the windows. As in Hebb's experiment, previously cited
(p. 9). the visible cues of the goal need not be present
to elicit the habitual response, as long as the general
environment is sufficiently stable to allow the establish-
ment of the orientation from which the response is to be
made..When the response is allowed to generalize to the
othervwindow, response to that window is easily made, in the
presence of either card in the habitual window (as illustrated
in figure 15, Appendix)..As has been suggested by Hilgard,,
it appears that fixation is actually a failure to generalize,
and not the expression of a compulsion.

That the responses of fixated animals are compulsive in
nature may also be questioned on the basis of the atypical
response alteration of the rat whose abortion resulted in a
fall to the floor..After two such_experiences, it was quite
able to alter its response, expressing no "compulsion" to

Jump to the negative card. It is significant also that the

3O

rat elected to jump 2h inches to the cage instead of to the
positive card on the non-preferred side. Since the rat was
not compelled to jump to the negative card, the question may
be raised as to why it did not jump to the other, which had
apparently been discriminated as the positive card. It would
appear that the positive card on the non-preferred side may
come to serve only as a cue of non-entry, since it is always
in the other window when the rat jumps to the negative card.
This hypothesis is further supported by the behavior of the
rat when the negative card is removed (Figure 5, Appendix).
The rat still jumps to the cage although "obvious" access

is afforded. The positive card on the right side appears

to be sufficient as a cue of expectancy of a locked window,
eliciting the response of jumping to the cage.

The behavior of rats following a fall to the net adds
supporting evidence to the hypothesis that an expectancy of
entry and reward after a fall is developed and maintained by
the underlying cue conditions inherent in the design of the
problem. The implication that the rats invariably made an
approach response toward the experimenter after falling is
difficult to reject. Such approach responses often occurred
directly after the appearance of the experimenter from behind
the apparatus. That such movement toward the experimenter

was related to an expectancy of food reward is suggested by

the following:

In working with rats which have been placed on a reduced
diet investigators commonly experience being bitten by the
animals. Such biting may range from tentative "nips" to
skin-puncturing bites. Since the experimenter is the person
who feeds the animals, such "biting of the hand that feeds"
may be assumed to be an anticipatory food response. The hand
is used in placing the daily ration of food in the cage, and
is therefore both a visual and olfactory cue followed by food

reward. One animal (number A, Table II, Appendix) showed
this biting tendency to a marked degree when fed in the cage,

seizing the hand of the experimenter each day as food was
placed in the cage and relinquishing it when the food pellets
were dropped on the cage floor. Over the last forty trials.
of the soluble problem, this animal was offered the hand of
the experimenter at two points: (1) when picked up from the
food platform after successful entry and food reward, (2) at
the right edge of the net following a fall. After eating,
when the probability of entry on the next trial is lower,
this rat bit the offered hand three times; after falling and
approaching the right edge of the net, the animal bit the of-
fered hand thirty times. Such reaponses could of course be
taken for "hostility" on the part of the rat, but since the
rat desisted after two or three bites, and did not indulge in
such behavior to an appreciable extent after eating it is more
likely that the response was an expectancy for entry and food

reward after a fall. Thus, although only one rat was tested

lIl'thlS manner, the approach response of all animals to the
_rigjit edge of the net could also be considered as expectancy
cufiiented. In other words, in spite of having failed to learn
this problem imposed by the experimenter, the rats 3“305Ted
{final-oriented on the basis of cues other taan those assumed
to Ice the most critical cues of the situation. The present
pussition is perhaps best summed up in an observation made by
Maier:

"Before one can conclude that an animal has failed to

learn, one must be sure that it hasn't learned something
different from what was intended." (12, p. 22)

Analysis of the data sheets used in previous fixation
experiments suggested the possibility that rats might learn
to use a sequential cue during the insoluble problem. It
is possible that a fall to the net may serve as an important
cue because of the increased probability of entry and reward
following such a fall. Owing to the pattern of card presenta—
tion, the cue would continue to be effective in the ensuing
soluble problem, serving to interfere with the learning of a
card discrimination.

Results of the present experiment indicated that such a
cue operated in the manner suggested. Rats jumped more rapidly
on trials following a fall to the net on the last ten trials.
During the soluble problem, rats jumped more rapidly on the
next trial following a fall to the net regardless of whether
the next trial offered entry and reward or another fall to
the net.

Although eight of ten rats failed to learn the required
discrimination, reaching the fixation criterion of two hundred
trials of the soluble problem, these animals appeared to be
goal-oriented throughout the problem. They continued to
eat on each trial when entry was made, and subsequent to a

fall to the net, they learned to make approach response to

the experimenter.

Of the two rats who learned to a criterion of fifty
successful entries, one continued to jump to the positive
card in preference to an open window, even when the positive
card was locked fifty percent of the time, resulting in a
fall to the net on half the trials. While this animal ulti-
mately learned to jump to the open window, it may be inferred
that similar behavior in fixated rats is not qualitatively
different from such fairly—well learned reSponse.

The results of the present study offer evidence against
the view held by Maier and his associates that fixated be-

havior is "behavior without a goal."

APPEND IX

35

TABLE I. MEDIAN LATENCY BEFORE FALL MINUS
_ MEDIAN LATENCY AFTER FALL ON DAY
ONE AND DAY EIGHT OF THE INSOL-
UBLE PROBLEM. (TEN RATS)

 

Rat Day 1: Day 8: Difference:
1. -27.5 seconds 12.5 seconds 40
2. - 0.5 " 3.5 " 4
3o: '36 n 1 n 37
2.. -33 3 g 3 3g
.. - 9 -

6. 13 fl _ 2 ll _15
7.. ~43 " - l " 42
8. - 5.5 " 7 " 12.5
9 e — -5405 n -16 II 38.5
10. “2205 N . 12 .5 N 42
Mean: -22.5 ' 2.15 24.7

The number on day one should be algebraically larger
than on day eight,.as it invariably is..

t ratio: 3.53. Differences are significant at the 1%
level of confidence.

 

*- fl' *

TABLE II. COMPARISON OF MEAN LATENCIES OF THE
EIGHT NONALEARNING RATS ON TWO OR
THREE SUCCESSIVE PRESENTATIONS OF
THE NEGATIVE CARD ON THE HABITUAL
SIDE DURING THE SOLUBLE PROBLEM.

 

Rat Mean time of first mean time of 2nd
Jump to neg. card: Jump to neg. card:

1. 39.3 seconds 26.9 seconds

2.. 30.9 " 23.9 "

3.. _ 36.1 " 26.2 "

4.. 42.5 " 22.5 "

5. 49 .9 fl 3704 II

6.. 71.5 " 51.1 "

7.. 48.1 " 35.5 "

8. 10.2 " 50.2 "

Mean: 48.6 34.2

t.ratio: 7. 55. Difference between means is significant
at the 0.1 level of confidence..

36

TABLE III. COLPARISON OF MEAN LATENCIES OF THE
EIGHT NON-LEARNILG RATS OL PRLSEH—
TATION OF THE POSITIVE CATZD OL THE
HABITUAL SIDE AFTER ENTRY AND AFTER
FALL DURING THE SOLUBLE PROBLEM

 

Rat Mean time of jump . Mean time of jump
to pos. card after to pos. card after
a previous entry: a previous fall:
1 2S. 6 seconds 16.1 seconds
2 18.1L " 1L1 "
i 23 7 n 13 6 n
13.7 II 2%:5 H
5 '7 fl '9 n
6 2.0 n 33.2 n
7 39.7 .. . 33.9 n
,8 22.1 18.2
Mean: 28.1 19-7

t ratio: 7. 21. Difference between means is significant
at theCLl%level of confidence.

 

 

TABLE IV. COMPARISON OF MEAL LAlW JQCIES Or THE
EIGHT NON-LEARNING RATS OL TH'O CR
THREE SUCCESSIVE PRLSENTATIQLS OF
TIE I‘OSITIVE CARD Om TIIE HABITUAL
SIDE DURILG THE SOLUBLE ILLSLLL

 

Rat Mean time of Mean time of
first jump to second jump to
pos. card: pos. card:

1 18. 2 seconds 25. 6 seconds
2 17:5 n 18 u n

g lh.8 n 23 7 n

5 33.8 n ' 15.; n

6 3827 I: -220 g

7 34.2 .. 39.7 ..

8 21.g 22.1
Mean: 22.6 28.1

t ratio: H.29. Difference between means is significant
at the 1% level of confidence.

 

 

Windows

Wb ;- @714? 28
Bw - Kb 12
Bw -'EQ 22
Wb - §g_ 15
Bw - E 5
Bw - Eb I7
Wb - E! 17
Bw e E2. 4
Wb - 2!. 14
Wb - §!_ 7
Fig.4

latency time:
(in~seconds)

37

Rat strikes locked window,
falls to net.

Rat enters right window, eats.

Rat enters right window, eats.
Note increased latency time on
second successive entry.

Falls to net..

Entry after fall..Note decrease
in latency time after a fall.-

Entry after entry, increased
latency time..

Rat strikes locked window on*
right, falls to net.

Rat enters right window, eats.
decreased latency after a falla

Rat strikes locked window on
right,,falls to net..

Second successive Jump to a
locked window. Note thatzlateny
cy time is decreased following
a fall to the net,,even:when
the non-entry card is display-
ed in the habitual window..

l..IdeaI?expectanny pattern~illustrating variation
in latency times after entry and after falling
to the net..This is the performance record of a
rat fixated on the right, for ten daily trials.

*Bw means "Black card with white circle";.Wb
stands for "White card with black circle".-

For this rat,,Bw was always locked,,an either
side, while Wb was always unlocked,.allowing
entry to food on either side. Symbol on the
left refers to left window; symbol on the right
refers to right window.

Bw

5

Fall
Entry
Fall
Fall
Entry
Entry
Fall
Entry
Entry
Fall

38

Both windows locked..
Both windows open..Entry after fa11.e

Fall,

trial 3, followed by 2nd fall..

Entry after fall..
Entry after entry..

Entry after fall.

falls to the net are followed by entry; one fall
to the net is followed by a second fall..

 

L
2..
3.
5:
a
7-
a-
9.
1m

5§$$¥§§$§¥

$3§§§$9§$§

Entry
Fall
Fall
Entry
Fall
Fall
Entry
Entry
Entry
Fall

Fall,

.trial 2,.followed by second fall.

Entry after fall.

Fall, trial 5,.followed by seoOnd fall.
Entry after fall.-

Two falls to the net are followed by entry; two falls
to the net are followed by a second fall..

1. Bw
2.~Wb
3. Wb
4. Wb
5. Bw
6..Bw
7..Wb
8..Wb
9. Bw
Three

Wb

Bw
. Bw

Bw
Wb~
Wb

Bw
Bw

Wb
Wb

Fall
Fall
Fall
Entry
Entry
Fall
Entry
Fall
Entry
Entry

Fall,
Fall,
Entry
Entry

Entry

trial 1, followed by second fall.
trial 2, followed by third fall..
after fall..
after fall.

after fall.

falls to the net are followed by entry; two falls
to the net are followed by another fall..

 

Fig..2..

Illustrating certain typical sequential prob-

abilities of entry and reward during the insol-
uble problem, when both windows are randomly
locked 50% of the time, so that neither card
can lead to 100% entry and reward..

39

1. Wb - §w_ Fall
2..Bw - Wb Entry Entry after fall..

3. Bw - Wb Entry

4..Wb - 21, Fall

5. Bw - Wb. Entry Entry after fall..

6. Bw - Wb Entry

7e Wb - §E{ Fall

8. Bw - Wb Entry Entry after fall.

9.;Wb - BE. Eall
10. Wb - Bu Fall Fall,,tria1,9,,followed by 2nd fall..

Three falls to the net are followed by entry; one fall
to the net is followed by a second fall..

A—__

1e\BW - Wb Entry

2..wu.- g! Fall A
-3..Bw - wb Entry Entry after fall.

#O.Wb - a! Fall

5..Wb - 31 Fall Fall, trial 4,,followed by 2nd fall.
6..Bw - Wb Entry Entry after fall.

7e-Bw "' Wb mtry

8e Bw - Wb Entry

9..Wb - g1 Fall

-Wb - 2!

Fall Fall,,trial 9, followed by 2nd fall..

Two falls to the net are followed by entry; two falls to
the net are followed by second falls..

 

1e Bw ‘ Wb Entry

2. Wb - 2! Fall

3..Wb - E! Fall Fall, trial 2,.followed by 2nd falls
4..Wb - Bw Fall Fall,trial Bufollowed by third fall..
5. Bw - WE Entry Entry after fall.

6. Bw - Wb Entry

7e Wb - E! F311

8. Bw - Wb Entry Entry after fall.

9..wn - 2!, Fall .
10. Bw — Wb Entry Entry after fall.

Three falls to the net areFfollofiEd'by‘entry;"two'falls
to the net are followed by another fall.

 

Fig..3.. Illustrating the sequential probabilities of
entry and reward after a fall to the net dur-
ing the soluble problem, for a rat fiXated on
the right. Although the Nb card is always une
locked” in either window, the sequnltisl one
of entry after a fall to the net is still in
operation..

Fig. 4. This rat,.shown making
an abortive entry to its pos-
itive card,,began to abort more
strongly at the close of the
experiment when confronted with
the negative card on the left
side. It began to miss the net
entirely, falling to the floor.
After two such falls, it began
to Jump to a cage 24 inches a-
way (see below) when confronted
with the negative card. It con»
tinned to Jump to the positive
card on the left side, however,
as shown.

Fig. 5. It was found that even
when the negative card was re-
moved from the habitual window,.
the rat still Jumped to the cage.
It appears that the positive

card on the nonrpreferred side may
come to serve only as a further
cue to the situation of a locked
window on.the habitual side, e-
liciting an:avoidance res onse
even when that window is obvious-
ly" open.

Fig. 6. A rat at the right
edge of the net,.exhibiting
the typical orientation to-
ward the experimenter, who
always appears from behind
the apparatus on the right
side.

 

 

 

 

41

GENERALIZHTION PROCEDURE USED IN BREAKING FIXATION

Fig. 7. A rat making an abortive
Jump to the negative card. For

this rat, the negative card, which
is always locked, is the black
card with the small white circle..
The Jumping stand is eight inches
from the windows.. ‘

 

Fig. 8..When the Jumping stand is
shifted to the right,.the habitual
window assumes the relative posi-
tion of the nonspreferred window..1
The rat cannot be claimed to be
expressing a "compulsion" to Jump
to the negative card,.since it is
not present..The elicited response
is that which has been practiced
in the past..

 

 

Fig. 9..Ewen.with no card pres-
ent in the window, the habitual
response is made, and occurs
without goading. The rat is shown
here Just at the point of leaving
the stand, as it Jumps to the
wall.

 

 

GENERALIZKTION PROCEDURE (CONTINUED) 42

Fig. 10. With the Jumping stand
directly before the habitual wine
dow, the fixated response is

- still elicited. The determinants
of response here seem to be 1hr
ternalized rather than objective.

 

Fig. 11. With the stand returned
nearly to the normal position,.
the rat Jumps readily to the ha-
bitual window.This trial is fol-
lowed by one in which the stand
is in the normal position (not
shown). The Jump is also made

to the habitual window..

Fig. l2..With the stand again~
directly before the habitual wine
dow,.the response is now directed
to the window. The cues of the ex-
ternal environment,.on the basis
of the previous responses, are be-
ginning to determine the response.

 

 

 

43

GENERALIZATION PROCEDURE (CONCLUDED)

Fig. 13. With the stand placed
as in figure 9, the rat still
Jumps to the window, although
it previously could not do so.
In terms of the relative posi-
tion of the habitual window,
the response may now be cone
sidered as being made to the
non-preferred window.

 

Fig..l4. Even when the stand is
placed far to the right, the rat
still enters the window.

 

Fig. 15. With the stand restored
to the normal position, the lock-
ed negative card in place in the
habitual window, the rat is now
able to enter the non-preferred
window. The rat can also enter
the open nonspreferred window

in the presence of the positive
card in the habitual window (not
shown). The response has been
generalized to the other window.

 

10.

ll.

12.

13.

BIBLIOGRAPHY

Eglash, A.£ Abnormal fixations.

 

Unpublished Ph. D. thesis,

University of Michigan, 1951.

Eglash, A.: Perception, association and reasoning in animal
fixations. Psychol. Rev., 1951, vol. 58, EZh-hBh.

 

Elliott, M. H.: The relationship of drive to learning.
Psychol, Bull., 1932, vol. 29, 6AA.

 

Farber, I. E.: Response fixations under anxiety and non-
anxiety conditions. g, Exp. Psychol., l9h8, vol. 38,

 

111-131.

‘

Gilhousen, H. C.: Fixation of excess distance patterns

in the white rat. Q. Comp.

Psychol., 1933, vol. 16,

 

1-23.

Hamilton, G. V.: A study of perseverance reactions in

primates and rodents. Behavior Monographs, 1916, vol. 3.

 

Hebb. D. 0.: The organization of behavior. John Wiley and

 

Sons, New York, 19h9.

Hilgard, E. H.: Theories of learning, Appleton-Century-

Crofts, New York, l9h8:

Jenkins, W. 0., and Stanley, J.

0.: Partial reinforcement,

a survey and critique. Psychol. Bull., 1950, vol. h7,

 

193-231 e

Klee, J. B.: The relation of frustration and motivation
to the production of abnormal fixation in the rat.

Psychol. Monog., l9hh, vol.

L<6

/ O

Maier, N. R. P.: Frustration: the study of behavior without

 

 

a goal. McGraw-Hill Book Company, New York, l9h9.

Maier, N. R. F.: Qualitative differences in learning in
rats. '3. Comp. Psychol., 1939, vol. 27, 289—332.

 

and Ellen, P.: Can the anxiety-reduction theory

explain abnormal fiXations?

V01 e 58 3 LL35'J4LLS-

Psychol. Rev., 1951,

 

45

1h. and Feldman, R. 8.: Studies of abnormal behavior
in the rat: XXII. Relationship between strength of
fixation and duration of frustration. J. Comp. and
Physiol. Psychol., 19MB, vol. hl, 3i8-3E3.

 

 

15. Mowrer, 0. H.: Learning theory and personality dynamics.
Ronald Press, NEW York, 1950.

 

16. Piaget, J.: Principal factors determining intellectual
evolution from childhood to adult life. Outside
readinns in psychology, (ed. by E. L. Hartley, H. G.
Birch, R. E. HartleyTwThomas Y. Crowell Company, New
York, 1950.

 

l7. holpe, J.: Learning theory and abnormal fixations. Psvchol.
Rev., 1953, Vol. 00, 111-115.

\ .
‘‘‘‘‘

  

J r .c
V . ‘g‘l (
- 4 ,.) ' I
.K .
,
.
‘ 4
V
I
I — ' ‘
b I‘.
.
.
r.
_ -.
.
.
' I
. . ‘:
‘ ;
.I
' I
, . a p
.‘ ' .
I V ..A -
I' .v ‘ ‘ I
. ‘ffia
. ‘ l
. "
(J. .3 '.
.- ~A>- _ V .
'. , v.' ",w'
- . .v 1 I
.._ .1 . Of,"
. .~ I

2'. ’4' .' ' ~

.‘ ‘ _ ’
I_‘ . f'
.r' '

I - ‘ ‘

_ P -,I$ , "r

. ;

f A‘ --_ ‘

_ \ . A, A “

‘. I . . f)
e ' P ‘ -

~~. ~ : ~ U
. n rL -.'.
. - ". L7
,' -/ ”‘C.
, . _. ‘_.. ‘.
‘._-
' ,"‘ u.
_ \
v ' \.
.I ~." A . \
.-. ,a ' \“ --2_.
. I '
. A .
. ,< ,. .
f ,-='..
._ . ‘ _
. 1' - , , \
,_ Na. f-kv‘in
‘ - .1 ' I‘
v ~" "l" '
/.“~
.‘ r'
' Q .
‘ :‘ 41 \
~".-‘\ . ' "
V,-_, ',~ "1'. - V.
i ,_ ..i
n , »_ .» 01,-.- . .
J . e. . .v‘,‘\‘z ‘ (Ia

I! “ .' ~_ 'V' ‘*

. ‘f' , ,_v r-

. .. ‘- -‘ -‘ '~ “7"
_' ‘ .- ' ”I
4. . I "1'.

_ ‘ - .

, at v ‘\’$
0" ’ i.
ll, _. 1' I .
,_ ‘ g V] , ' '
I a . ‘.
-\ . — '0‘“
. ”(In W ‘
, .
.‘e"
. a ‘ \I
’ .
_.
_ .
. n. , p
. l.‘; f
’ _ -
f. .u‘
.
'- ’v '1-
"
'l ‘ .
v
. D'-
_ . .
J. .41 .L’
_ .. ...Nvfi‘
. ' .. I":
“x“! . , 'I ‘
, J'.’
1 _ " v
V
'v , M:-
‘,’ I I ‘
. 1' J

  

e 1“} “‘1’

L;

'1 .‘( —
I ‘ '
e I |
1 ' 1 b v‘
Yr - ‘ a I "a" e I ‘.
‘ 4' . , e
I . ‘ 4" ‘ " i '
i . ~ " .‘ '-' q. ‘ “‘

ROOM. US“. ONLY

  
  
 

an“:

   
   
  

in

'32?ng K”
vii-3‘”, .. au-

  

   
   
  

7:

‘03:

I

I
‘ J
o
1
e
u
I
I
‘ .
,,
. L
.,
.
‘.
e
. A
1 .
l
J .

~.
~.
.2‘
\ l
. ‘ I
.‘ >~
'1 , .. .
"t . ‘ ‘
_ r _
2-, N.
v‘; a
I, .
.V '
’fl

.‘w
1.,1'3
.4 '> '7
1"~‘;."V
: ‘ h
’I‘ 'I '
"ik'qv‘
";| ‘v
“it.“ ,‘1
_ ' ..1' ..
3" I‘.‘ \‘
'>.'-r, - v\'.
\ he I: 5’
.. 01$- .
d :40" _'
‘l'n ‘ ~33

3.2

~

~-
x
'~ 51‘

V
e

§

4' -, As
.‘,'.. .. ',
no -

.e l
.' \ r“

T ‘ "

.3.) ' '
-~’ I. _l

'.'-4Iv '
‘ \' Q

. , e ‘.'

.- ‘1 .
hi ,
x .
I‘d» Wm?»
man " ~
I ". .3.‘
_,.~ I-I. -.
' ;--'l v '
e . ' .' .
a}: '. \' -‘
r 7) "i
' '1“ f0
‘ ‘ (v. 1"‘5f-Q
"1' ' (‘4'.
‘3. 'i her
,, M " 'r.
_r.1'4,:'_‘_‘..-‘V ' '
, l. . O y I
‘(-:.;‘f‘.‘ '-
-1 l‘.-‘.‘.“ ‘
."'~ :‘5
." .4 .
#4:;- ~‘
‘ i‘x-‘nt‘ ‘
b 1' “I". .
‘.,\‘fl‘ .
‘1‘"r ‘ J"
flIfif,
"in; .. k
If 'A‘ 0.: h
5 +4.:

 

o..-

1"»be

MMWM

“1'

\
V

w whOMWEMWtMW

u

 

J

HUIHJIH

'I
I,
l
l
l
l
I
l.

I
I
I
ll