Beswick Curiosity 1964

The Department of Social Relations
in partial fulfilment of the requirements
for the degree of

iii
Acknowledgements
The initial stages of this work were supported by the School
Mathematics Study Group, Yale University, under the direction of
Dr. E . Begle. My work at Harvard, in this connexion, was supervised
by Richard Alpert. I am indebted to him and other members of the
S. M . S. G . psychological assessment group for many suggestions and
critical comments which guided construction of the curiosity test.
David C. McClelland also gave valuable guidance at this stage.

I am indebted to Ralph Metzner for his co-operation in the
curiosity and delay of gratification project. This study was
supported, in part, by funds from USPH Project M-3344 (Rl), directed
by Walter Mischel, and a major portion of my work was made possible
by the George W . Naumberg Fellowship. Thanks are due to Kenneth
Keniston for permission to use unpublished scales, and to Lauren
B. Resnick for administering the curiosity test and providing
additional data.

The work would not have been possible without the willing cooperation
of school principals, teachers and students in Watertown
and Brookline, Massachusetts and in Mornington, Victoria, Australia.
The oversight given by Robert W . White during the final stages
of preparing the thesis has been most valuable.

The task in theory and measurement.
Experimental literature: history, reviews, alternation
behaviour, tension reduction, general theory of
drives, expectancies, cognitive theory, diversity or
unity of curiosity, genesis of curiosity, significance
of experimental findings,
Freud's theory: review, two themes, ego or eros, maschocism,
anality and curiosity, cognitive theory
implications.
The question of method: need for test of individual
differences, appropriateness of the TAT, cognitive
not projective assumptions.

Pictures and administration instructions: dependency
on conceptual definition and scoring criteria.
Selection of scoring criteria: conceptual definition,
subjective preparation, pilot study selection of
pictures, first scoring manual, correlation with
picture cues, data continency modifications, implications
of literature in history of science and
religion, effect of experimentally induced curiosity
on imagery, definition of major and minor categories,
correlation with asking 'Why?', techniques of
modification, scoring reliability.
The question of validity: correlations with external
criteria in development, aim of construct validity.

General instructions to scorers.
Major imagery criteria: wonder-interest, perceptual
instrumental acts, cognitive instrumental acts,
exploratory role behaviour, cue-response sequence.
Minor imagery criteria: intrapsychic response of
excitement, interesting stimulus, overt instrumental
acts.
Contrary indications.
Total score: instructions, table.

Introduction: motive or cognitive style.
Definition of cognitive style terms: category system,
signal, coding operation, efficiency of category
system, redundancy of signal, difficulty of coding,
accommodation, assimilation, rejection, individual
differences in the coding operation.
Conceptual conflict and tension reduction: uncertainty
and conflict, accommodation and learning, conflict
tension reduction and learning, pleasure and
excitement.
Maintenance of conceptual conflict: combination of
openness and discrimination, interaction between
orderliness and novelty, instrumental acts delay
resolution of conflict.
Creation and resolution of conceptual conflict:
delay and compulsive characteristics, exploratory
behaviour and other forms, generation of conflict,
growth of organization, effects of negative reinforcement,
learned strategies of learning, anti-
thesis between curiosity and avoidance, summary.

Theoretical relevance of orderliness and novelty.
Orderliness: measure derived from analitys cales,
procedure, interpretation of correlation.
Attention to novel stimuli: questionnaire item
correlation for girls, relation to yeasaying,
'sex role identity’ correlation for boys, discussion
of sex differences.
The combination of orderliness and novelty: methodo-
logical problems, interpretation of multiple
correlations.
Delay of gratification: theory, procedure, conf-
irmation of hypothesis for girls only.
Anxiety and Achievement: facilitating and debil-
itating test anxiety, need for achievement, sex
differences, school achievement.
Sex: different correlates, mean differences,
possible explanations.
Heading: differences in reading habits, theoretical
interpretation.
Independent inquiries in religion: accidental
discovery, theoretical relevance, possible
significance for education.

Chapter 6. Conclusion 99.

Agreement with Berlyne's theory: assumptions, scoring
categories and Berlyne's description, face validity,
Berlyne's theory explanations of findings in pilot
studies and correlates of the test, theoretical
difficulty of orderliness correlation.
Evaluation of Berlyne's theory: tension reduction model
modified, consistent explanations, logical and
empirical inadequacy of modifications to tension
reduction theory.
McReynolds' theoretical model: advantage over Berlyne's
theory, disagreement with cognitive process theory.
Advantages of cognitive process theory: two facets of
curiosity, dynamic aspects, flexibility of analytical
tool, predictive power.
Problems and possibilities: situational variability,
need for experimental studies, real life correlates,
implications for educational research, future of
motivation theory.

Appendix 4. Curiosity, Anality and Delay of Gratification: 139
Identification of variables,
Data summary and analysis,
Interpretation of factors.

Appendix 6. Curiosity in the Classroom: 161
School Mathematics Study Group Project,
Teachers comments on high and low
curiosity scorers,
Discussion of social implications.

Curiosity has become a matter of central concern in psychology. The
survival of motivation theory in general, and of tension reduction
models in particular, could well depend on the results of research in
curiosity. Evidence from studies of exploration, ludic behaviour and
other apparently stimulus seeking activities has raised again the
question of whether purposive behaviour is always to be explained as
escape from an aversive drive state. Psychologists have also been
called upon to contribute to the recent widespread debate on basic
requirements of an adequate educational system, and have had to
consider the possibility that acquisition of knowledge might not be
facilitated as much by pragmatic considerations as by intrinsic reward.
The demand by contemporary society for more scientists and the increasing
prestige of creative artists have directed attention towards
a greater understanding of the psychology of innovation.

If we are to enter the field directly at the level of human
cognitive processes one of the first tasks must be to develop an
operational definition of curiosity. Some progress can be made
along the lines of animal psychology by experimental manipulation
of stimulus conditions. For example, curiosity might be defined as
an intervening variable when there is an approach response to a novel
stimulus. But with human subjects verbal behaviour gives much more
direct access to the relevant properties of the organism. To decide

whether curiosity is present or absent or present in different degrees
we can take a sample of a person's cognitive processes from his verbal
[Re 'his:' a note made in 2016. I hope the 21st century reader will excuse the
use of male pronouns in the inclusive sense. This was written in the early
1960s before conventions changed and I have not edited this usage. In
this and other respects the work remains as it was originally. DB]
behaviour. The method of taking the sample and describing i t s content
will provide an operational definition which is a measure of individual
differences. Having obtained such a measure we can proceed to
investigate the determinants and effects of curiosity either by experimental
manipulation or by correlation of naturally occurring variations.

While the psychometric task must have high p r i o r i t y i t cannot be
undertaken independently of theory construction. We are faced with a
task which is different from that of building a test to predict
behaviour i n a given situation, for then we can be empiricists completely
and include any item which increases predictive роггег; but if we think
that curiosity i s an important variable we already have an implicit
conceptual definition and a set of expectations about its correlates.
The value of the measure will depend upon its construct validity:
that is, it must be related i n a predictable fashion to other relevent
variables, even though the predictions are made initially from a
theory which might include some hypotheses that will not stand up to
later investigations in which the test is used.

The parallel but interdependent tasks of test building and theory
construction take place against a background of significant developments
in experimental psychology and personality research. The
basic methodological problems associated with analysis of at least
some types of verbal material have already been solved and we are
not faced with the impossible task of establishing construct validity

in a theoretical vacuum. However, apart from the recent work of
Berlyne, the field of human curiosity is almost unexplored. Even
Berlyne's contributions owe much to animal studies. Whether
experimental observations of other species will eventually result in
a fundamental biological theory which would contribute significantly
to an account of adult human intellectual behaviour is an open
question. Before any relevance can be demonstrated, it will be
necessary to develop methods of measuring the human qualities and
describing their interrelations. Biological research and animal
behaviour theories can be a source of hypotheses to be tested with
human subjects provided that we can identify the variables correctly
in the human context. Assuming that this identification can be made,
the literature of experimental psychology should provide us with some
guide to the operational definition and its correlates.

Although the experimental study of exploratory behaviour might
be said to have begun with Nissen's work in 1930, there are scattered
references to exploratory and similar behaviour as far back as 1899
when Small (76) noted 'premonitions of curiosity' in the infant rat.
Small's report that exploratory behaviour appeared to be fully
developed at the age of three weeks was a forerunner of such developmental
studies as Meyers (58) 1962 finding of a critical period of
facilitation at 6 to 10 days; and his conclusion that the observed
behaviour was not merely a reflexion of hunger foreshadowed much

later research on the interaction between exploration and other so
called drive states — e.g. Jencks and Porter (-48) reported in 1963
that a novelty sequence of goal objects had as high an incentive value
as the highest incentive goal objects.

Nissen's finding (69) that rats would cross an electrified grid
to explore a novel area in the form of a Dashiell maze was preceded by
Dashiell's suggestion (28) in 1925 of an instinct of curiosity to
explain behaviour in the maze and by Lashleyfs reference (49) in 1912
to an ejqploratory impulse in connection with visual discrimination.
Tolman's cognitive theory generally and his interpretation of
alternation behaviour at the choice point (78) in 1925 also laid
foundations for the study of curiosity.

After 1930 the volume of research increased greatly. Heathers
(46) made a review of the literature in 1940 and by 1958 Barnett (6)
was able to cite 125 references in a critical analysis of exploratory
behaviour. Other reviews were published about the same time by
Butler (15, 17), Dember and Earl (29), Dember and Fowler (39), and
Glanzer (42). These were followed by further reviews and theoretical
analyses by Berlyne (10), and by Fiske and Maddi (Ed.)(33) including
White's (84) reconsideration of motivation. Against this background
there is little to be gained by making an exhaustive review of the
field and it will suit our purposes to note several theoretically
important issues which have received wide attention.

The alternation behaviour of the rat in a Y maze was one point
of conflict between response as opposed to stimulus oriented theories.

The question was whether alternation was to be explained by reactive
inhibition following Hullian principles of behaviour, according to
which, if a lef t-turning response is made, a certain amount of left-
turning inhibition is generated which renders the right-turning response
temporarily predominant. Glanzer (40, 41, 42) and others (80, 81) put
forward a hypothesis of stimulus satiation to account for the same
behaviour. At about the same time Montgomery and his co-workers
carried out a series of experiments which demonstrated the need for an
explanation in terms of the positive incentive value of novel stimuli
rather than avoidance of the alternative stimulus or response (61, 62,
63, 64-, 65, 66). (Berlyne and Slater came to similar conclusions, 12).
Montgomery claimed that i t was not only necessary to assume an
exploratory drive but that reinforcement of exploratory behaviour was
based on drive increase and not drive reduction.

I n spite of the progress made in recent years the drive reduction
controversy still continues. After making an extensive review i n 1959,
White (84) referred to 'the new freedom produced by two decades of
research on animal drives.' He said, 'We are no longer obliged to look
for a source of energy external to the nervous system, for a consummatory
climax, or for a fixed connection between reinforcement and tension
reduction.' The following year Berlyne published the first book to
appear i n the f i e l d of exploratory behaviour and related phenomena (10)
and offered a modified version of drive reduction theory. He remarked,
'It would be ironical if, after all, the study of curiosity provided
this type of theory with its firmest stronghold, and if the study of the

role of cariosity in learning as a whole served to buttress its claim
to far reaching justification . There can, however, be little doubt
that more satisfactory formulations will come to light before long and
replace it.' It is clear that such a formulation will have to possess
some of the characteristics of both White's 'concept of competence'
and Berlyne's theory of 'conceptual conflict '.

White and Berlyne (and the whole body of contemporary writers) are
in agreement that if the drive hypothesis is to be retained it can no
longer be stated primarily in terms of tissue needs. If exploration
is admitted to the class of drives attention is directed to central
processes as the locus of motivation. Much the same has happened as
a result of research into the more traditional exemplars of drive such
as hunger and sex. For example, Hardy's recent contribution of an
appetitional theory of sexual motivation (44) is based on cognitive
expectancy theory, Lewinian field theory, and modern hedonic or
affective theory. Hardy claims that 'the overwhelming proportion of
the variance in human sexual motivation and behaviour is not explicable
in terms of some biological need or tension, however conceived.'
Such emphases come out more strongly when human behaviour is studied
than in animal psychology, but it is typical of the conclusions drawn
from the rapidly accumulating evidence of experimental psychology
generally.

The nature and function of expectancies in purposive behaviour
was the point of a long controversy which seems now to have teminated
(53, 79). The Hullian rg (fractional anticipatory goal response) and
its stimulus counterpart Sg were devised to meet some of the challenges

-7-
of Tolman's cognitive theory. It now appears that what started out as
a sharp distinction between easily defined objective stimulus or
response elements and less accessible cognitive elements no longer
exists. The need is now for direct measurement and analysis of
mediational processes, which can be achieved by both physiological
research on the central nervous system and analysis of human verbal or
perceptual behaviour in terns of a cognitive theory.

There have been many significant developments towards a cognitive
theory of motivation including Piaget's account of intellectual
processes (70), Festinger's theory of cognitive dissonance (32),
Bruner's discussion of perceptual readiness (14-), Gardiner's research
into cognitive control principles (recently summarized by Silverman, 75),
Woodworth's behaviour primacy theory (86), McReynolds cognitive model
of curiosity (56). Not all of these are concerned with motivation as
such, but all refer to the determining effects of central processes on
stimulus selection.

One of the continuing problems is whether curiosity can be regarded
as a unitary concept. Is it a congeries of many things brought
together because they do not fit in the previous categories of
motivation? These many things do have in common the apparent directing
of behaviour in a stimulus seeking and selecting sequence; that is,
instead of reacting to a passively accepted stimulus, the animal goes
out to act upon the environment thereby increasing the variety of
stimuli received. The question is whether the same fundamental process
underlies all such behaviour — including its covert parallel in human

intellectual behaviour. For example, should Harlow's manipulation
motive be included? Harlow claimed (45) that manipulation and
exploration are distinct on the basis of the considerable persistence
of the former. However, since Butler and Harlow (19) have shown
that the same persistence can be a property of exploratory behaviour,
the basis for distinction no longer exists. White suggested the term
effectance motivation to cover the energizing and directing properties
of the central processes underlying all such behaviour, but he would
probably regard curiosity as only one aspect of this general tendency.
Berlyne has expounded his theory of conceptual conflict to cover the
whole field . It could be that all stimulus seeking behaviour should
be defined as one class or there may be profit in making some distinctions
at the behavioural level while seeking to define a common motivational
condition, but whatever definition of curiosity we adopt the
indication is that an explanation of curiosity will have wide ranging
implications.

The effect of admitting curiosity or exploration to the theory
class of drives is to raise some fundamental questions about the
nature of a drive. Because of the broad class of behaviours effected,
together with changing conceptions of drive deriving from other sources,
it threatens to change the whole of motivation theory. Quite apart
from whether we consider tension-reduction necessary for reinforcement,
the concept of a drive generally presupposes some mechanism for
increased activity following deprivation and a corresponding slackening
of effort with satiation. Many investigators have studied the drive

properties of exploration in this respect. Habituation of exploratory
behaviour (e.g. 27, 24) could be due to either response inhibition or
stimulus satiation. Berlyne (8) and Butler (16) found evidence of
satiation being related to previous exposure or deprivation. More
recently Bolles (13) has concluded that there is an effect from prior
but not current deprivation and that exploration is much more dependent
on environmental stimulus conditions than the animal's deprivation
conditions. Montgomery and Zimbardo (68) found no differences due to
deprivation and concluded that exploration was based on exteroceptive
stimulation and not escape from boredom. Charlesworth and Thompson
(23) came to exactly the same conclusion. Glickman with rats (43) and
Butler and Alexander (18) with monkeys found evidence of a tendency to
seek an optimal level of stimulation. In his review, Barnett (6)
concluded that environmental change, influencing the exteroceptors,
increases exploration and that deprivation (a) lowers the threshold of
response to external change and (b) in itself may increase not only
gross locomoter activity in general but exploration proper. There is
no simple solution to the problem, but it is certain that in this
respect also exploration does not fit the 'orthodox' model.

With respect to the effect of deprivation the problem is further
complicated by interaction between exploration and other drive states.
It has been amply demonstrated that food deprivation will increase
exploratory behaviour (1, 31, 72), however Zimbardo and Montgomery (87)
found a decrease which they attributed to conflict between exploration
and incipient consumetory responses. In one experiment involving food

deprivation Wehling and Prokasy (82) found no exploratory behaviour
in terms of choosing a condition of uncertainty for the high drive
group. They claimed this was contrary to Berlyne's theory, but
Berlyne replied (11) that the main point he had been trying to make was
that extrinsic exploratory responses are reinforced not by reduction of
uncertainty but by reduction of conflict which could explain their
results. Novel stimulation can induce fear (67, 5) which can function
as a competing drive, and exploratory behaviour can sometimes be explained
as escape or avoidance (71, 83), but these raise no difficulties
for Berlyne's theory and do not discount the positive findings.
Curiosity can be either pure or mixed — serving its own ends or subserving
another motive or being served by another motive.

If exploration is to be considered a drive the question arises of
whether it is primary or secondary, or if this distinction is hard to
maintain, the extent to which it might be subject to modification by
experience over a long period. Some studies have shown hereditary
differences in exploratory behaviour (77, 21). Modification due to
early training is suggested by a few investigations (52, 58, 88) but
the results are not yet veiy convincing. The genesis of the drive if
it is a drive remains largely a matter of conjecture.

Experimental studies leave no doubt that a broad class of novel
stimuli have an incentive value. Animals will overcome noxious
conditions and learn new responses to gain this incentive. If it is
postulated that such behaviour is due to a motive or drive stimulus
property of the organism, the result is to effect a basic change in the

theory of motivation. There is evidence from other fields to support
a revision of theory in the direction of a more central or cognitive
emphasis. Although it might not be possible to maintain a unitary
concept of curiosity, there is good reason to expect that the fundamental
principles of motivation which apply to it can be applied to a wide
variety of behaviour. Tension reduction theory can only be retained
in a highly modified form if at all. The influences of long term and
recent past experience are not yet clear, but we do know that a large
proportion of the variance in curiosity behaviour is due to present
environmental conditions and therefore the internal conditions for this
interaction must be very common properties.

Freud never explored the general problem of human curiosity for
its own sake. Nevertheless, his incidental comments are quite
important not only because of the general influence of psychoanalytic
theory but also because his insights suggest some interesting
possibilities for future theory construction.

Aronoff has reviewed Freud's conception of the origin of curiosity
(3) and found two quite different themes in his writings. The first
and most widely known suggestion is that adult intellectual curiosity
is a sublimated form of scoptophilia originating in the anal—sadistic
stage of development. The second suggestion placed curiosity in the
realm of an ego coping mechanism which he ascribed primarily to the
child's need to find an explanation for the arrival of a sibling.
However these two themes are not as distinct as Aronoff suggests.

The first theme is clearly illustrated by his comment in the
General Introduction that 'impulses with a passive aim are connected
with the eroto-genetic zone of the rectal orifice, at this period very
important; the impulses of scoptophilia (gazing) and curiosity are
powerfully active ....' (35). In the same work he says, 'Infantile
sexual curiosity begins very early, sometime before the third year.
It is not connected with the difference between the sexes This
curiosity is for the most part aroused by egotistic dread of the
arrival of another child.’ In the 'Three Essays on the Theory of
Sexuality' (36) the two themes are found together: 'This instinct
cannot be counted among the elementary instinctual components, nor can
it be classed as exclusively belonging to sexuality. Its activity
corresponds on the one hand to a sublimated manner of obtaining mastery,
while on the other hand it makes use of the energy of scoptophilia.'

Many people have never accepted Freud's 'final solution' to the
problem of the instincts and have preferred the earlier eros and ego
classification. Without going into the more recent ego instinct
theories which also point in this direction, it would seem reasonable
to regard curiosity as belonging primarily to the ego's capacity to
obtain mastery over the environment and to disregard the supposed
energy source in eros. This would fit in well with developments in
experimental psychology where exploration is being accepted as an
independent primary drive. However we might be missing something
very important if we do this. It was sado-masochism which, more than
anything else, persuaded Freud to regard the ego instincts as deriving

from eros. With this in mind, the affective amibivalence of curiosity
('My curiosity is killing me') might provide us with a clue to the way
of bringing together White's 'concept of competence' and Berlyne's theory
of 'conceptual conflict'.

In his final formulation (37) Freud defined eros in terms of a basic
tendency to seek the unity of all things. Curiosity is concerned with
seeking unity in the face of disunity and variety. It necessarily includes
a conflict between signals received in communication with the
environment or between the symbols of past and present experience.
To be curious one must be able to face and overcome the disruptive
effects of novelty. There is a masochistic pleasure implied in the
positive valence of disruptive novelty. Freud's suggestion that
curiosity originates in the anal-sadistic stage brings the pleasure
component into juxtaposition with the characteristics of ordered control.
The link with ordered control is at the point where sadomasochistic
elements are supposed to become differentiated from the
libidinal complex.

Freud was undoubtedly a genius in tracing the vagaries of symbolic
associations, but he never carried out experiments on the effects of
stimulating erogenous zones. If we place a large question-mark over
his theory of somatic origins, we are free to regard his findings as
relating to the dynamics of a person's cognitive map. In this instance,
the so-called anal characteristics can be regarded as a group of
associated processes without necessarily having a somatic location
reference except in the CNS. To do this is much the same as to

disregard or deny the somatic origins part of drive theory, but the
result is different. The result of this shift of emphasis in experimental
psychology is to make environmental stimuli the primary source
of variance and to define curiosity as a seeking of novel stimuli.
Freud's bringing together of environmental mastery and anal characteristics
directs our attention to the mechanisms of cognitive control
which create, maintain and resolve conceptual conflicts. The
suggestion is that we should look for two contrasting facets of
curiosity. The desire to experience disorder and the desire to order
experience are compounded in a cognitive style with tantalizingly
ambivalent affective properties.

We have seen that there is a need for an operational definition
of human curiosity which would enable us to give direct attention to
theoretically important mediating processes. Verbal behaviour is the
obvious point of entry. This is especially so if we are to proceed
with cognitive theory. Furthermore, a quantitative assessment of
verbal behaviour should provide us with a method which is not restricted
to laboratory conditions. This is important because so much
curiosity behaviour takes place in a relatively free environment. We
need a test which will enable us to correlate curiosity with naturally
occurring events as well as the somewhat artificial indicies of controlled
experiments and personality tests.

Quantitative thematic apperception tests of various forms have
been developed by McClelland and others (55, 4). Although based on

analysis of fantasy material these tests have a very high degree of
objectivity. It is possible to program a computer to obtain test
scores from written stories. The usefulness of these tests depends
upon the demonstrated possibility of discovering, in imaginative
verbal productions, certain typical thought forms which occur in a
wide variety of situations in real life.

The TAT is often regarded as a projective technique and
assumptions are made about the subject attributing his own motives
to fantasy figures. This might be useful for some purposes, but it
is not necessary to make any such assumptions. We can regard the
verbal material simply as a sample of a person's cognitive processes.
That it is a product of fantasy is relevant only to its generality
being increased by freedom from specific situational demands. This
last point is always open to the charge that the test might have no
generality at all and relate only to fantasy, but that is an empirical
matter to be decided by the presence or lack of meaningful correlates.

If curiosity is to be understood as a cognitive style, and if
there are consistent individual differences which can be evident in a
sample of verbal behaviour, then the thematic apperception test method
is admirably suited to the task of measuring curiosity. To achieve
our goal it will be necessary to define the sample by selection of
suitable picture stimuli and administration conditions, and to discover
the scoring criteria. Theoretical considerations enter into the test
building process — especially in regard to scoring criteria. We
have the advantage of experimental and psychoanalytic theories as well

as common language and literature to guide us. To demonstrate the
validity of the test we can set it within a network of correlations
which are consistent with our theory. Initially the test is modified
in the light of the theory until we have sufficient evidence of
validity. Then the test becomes a fixed operational definition and we
can proceed to examine the important theoretical problems surrounding
curiosity.

The test consists of a set of pictures (Appendix l ) , administration
instruction (Appendix 2 ) , and a scoring manual (Chapter 3 ).
Children were asked to write short stories after looking at pictures
projected onto a screen at the front of a classroom. The stories
were then scored for the presence of curiosity imagery. In the final
form of the test a story is written for each of four pictures. The
range of scores for each story is 0 -4 and the individual's test score
is the sum across four stories.

The administration instructions present no problem as the
requirements of the subjects are the same as with the TAT measures of
n Achievement, n Affiliation and n Power (Atkinson, 4). We took over
the same instructions as had been used with tests of this kind
previously — although slight changes occurred from time to time according
to the test situation and the need to administer other tests to the
same subjects or through slight differences in language between the
United States and Australia.

Selection of suitable pictures could not be undertaken independently
of deciding upon scoring criteria. We needed the picture stimuli
to obtain stories which could be used in the development of the scoring
manual. At the same time we needed scoring criteria in order to decide
which pictures produced curiosity imagery in a sufficient proportion of
stories to make a test of individual differences possible. This little

problem is illustrative of many which occurred in the early stages of
developing the test. In order to break this circularity we must
permit certain untested assumptions to enter. One is that we can
recognize curiosity cues in a picture even if this can be done only in
some cases and with low reliability. When we have some picture stimuli
which produce the responses we are seeking we can use correlations with
the responses to make a better selection. Similarly we must make
initial assumptions about scoring criteria. We must have some idea
of what categories of imagery to look for in stories before we can
relate imagery to picture stimuli or other conditions. Both these
assumptions presuppose the acceptance of a conceptual definition of
curiosity.

The conceptual definition with which we begin a psychometric task
is not always stated lest it be confused with the final definition
which must be in operational terms. One could argue that not only
is the conceptual definition irrelevant but the whole developmental
process could be disregarded in evaluating any test. A test is a tool
and its value depends on what can be done with it. Utility is the
only worthwhile test of validity [Cronbach]. However there is value in saying
where we began and in recounting step by step exactly what was done;
the conceptual definition and other assumptions of the investigator,
together with correlations obtained in developmental stages, can be
suggestive of the relations which the test might have with other
measures. The validity of the test is suggested but not demonstrated
by its method of construction.

'Curiosity’ is a word which has been used by many people for
many centuries. Its varied uses are documented in dictionaries and
similar reference works. A search of these led to a closer examination
of some now obsolete senses. Of particular interest was the
association between 'curiosity’, 'accuracy', and 'curate' through the
Old French 'cure' meaning 'care'. Historical associations in the use
of a word could suggest some psychological equivalence or common factor.
The following quotation from Chaucer describes a man who seems to have
the characteristic which concerns us.

'A clerk ther was of Oxsnford also,
That unto logik hadde longe y-go.
But al be that he was a philosophre,
Yet hadde he but litel gold in cofre;
But al that he mighte of his friendes hente
On bokes and on lerninge he it spente,
And bisily gan for the soules preye
Of hem that yaf him wher-with to scoleye.
….
Of studie took he most cure and most hede.
….
And gladly wolde he lerne and gladly teche.’
(Cant. Tales: the prologue, lines 285-309, Skeats edition)

This early close connection of care, carefulness, and attention
with curiosity, together with its obsolete sense of a scientific or
artistic interest, suggested that a somewhat wider definition than is
appropriate for contemporary usage would be an advantage if we wished
to discover psychologically interesting correlates. I was also
impressed by the ambivalence of feeling people had towards what they

called curiosity. While Chaucer used it in apparent admiration of
his clerk of Oxenford, Wycliffe condemmed it as dampening faith.
Mathematics, science, art, heresy and witchcraft seemed related in this
word.

The following conceptual definition was used as a starting point
in the selection of pictures and the first tentative scoring criteria.

1. A desire (need, wish, want, etc.) or inclination to know
or learn especially about novel or strange things.

2. A feeling of interest leading one to make inquiries or
perform some acts in order to gain understanding.

3. The application of careful attention to a learning task
the outcome of which has no immediately obvious pragmatic
value.

4. A scientific or artistic interest which is intrinsic to
the person and oriented to understanding rather than to
expression.

5. A desire to make trial or experience of anything novel:
to see what will happen; to find out what it would be
like; to see whether it will work.

Putting together something like the Curiosity TAT is a work of
synthesis. Analysis is helpful in both selecting elements and in
combining them; but it presupposes that we already have some conc-
eption of what will be relevant, where to look, what the bits will
look like, and at least a vague image of the finished product. The

less one has of these the less use one can make of an analysis. My
first and continuing task was to develop these expectancies.
I had previous experience with the TAT including some clinical
training and I had scored over one thousand stories for n Achievement,
n Power and n Affiliation. With a set derived from this experience,
the conceptual definitions of the dictionaries and their associations,
and hypotheses advanced in psychological literature, I read and re-read
TAT stories collected by others for other purposes.

(a) Selection of pictures. Thirty-six pictures were rated by six
judges independently for presence of curiosity cues. Three of the
judges had previous experience in the development of tests of this type.
The conceptual definition of curiosity given above was read to the
judges who were asked to rate the pictures for the amount of imagery
consistent with that definition they would expect to be found in
stories written in response to each of them. It was assumed that if
a picture showed a person performing a certain type of behaviour or
it could be so interpreted, then subjects with a motive to perfom that
behaviour would be likely to write stories in which the motive is
evident.

Eighteen pictures were selected on the basis of the judges' combined
ratings. One further picture was added because of the imagery
found in stories which had been written to it in a similar previous
study. The pictures v/ere divided into five sets and each administered
to about thirty seventh-grade children (a total of 152 subjects).

Those pictures which were most effective in producing curiosity imagery
(as it was then understood) were later selected for use in the second
pilot study. However the judges' ratings were also used for another
purpose in providing the first tentative evidence of validity of
scoring criteria for curiosity imagery.

(b) The first scoring manual. At this stage I was working under the
assumption that curiosity is a motive and that the evidence for it in
fantasy productions would be analogous to the imagery of need for
achievement as described by McClelland et al (55). That is, I expected
the evidence of curiosity to consist in part of a person in the story
being in a state of need, performing an instrumental act with the
purpose of meeting that need, and associating some kind of affect with
attainment of an appropriate goal object. Accordingly, I took the
n Achievement scoring manual in outline and fed into it analogous
material collected i n Phases 1 and 2 . The sub-categories were excluded
and the conceptual definition appended. Some general considera-
tions were added as instruction s to the scorer and the imagery
categories re-arranged to provide for detailed classification within
the major categories . The scoring manual used in the first pilot
study i s found i n Appendix 3. It i s markedly different from the final
form in Chapter 3 and was essentially a heuristic devise rather than a
measuring instrument.

The strategy of discovery rather than the strategy of proof is
Implicit in the following instructions : - 'Once a scoring definition
has been firmly established it is imperative that the scorer apply the

criteria strictly and assume the absence of imagery whenever he is in
doubt no matter how narrow or arbitrary the definition may appear.
In the early stages of constructing a coding system, exactly the reverse
applies. It is advisable to use the widest possible d e f i n i t i o n and to
assume the presence of imagery whenever one is i n doubt. This i s no
excuse f o r vagueness....the scorer should note the reason why he scored
each instance'.

The stories were coded f or presence or absence of curiosity imagery
by two scorers independently. Although the instructions were not
designed to produce high scoring reliability at this stage, the overall
agreement was 81 per cent (69 per cent f o r presence and 86 per cent f or
absence of imagery).

The percentage of stories coded by each scorer as containing
curiosity imagezy was determined for each picture and the pictures
ranked i n order of their ‘percent imagery’ f o r both scorers separately.
The rank difference correlation s between percent imagery found by the
scorers and the judges' combined ratings of picture curiosity cues were
significant for both scorers (Rho = .64- and .55, N = 18, P < .01).
Interscorer agreement on the ranking was also significant (Rho = .75)
and when the percent imagery was averaged for the two scorers the
correlation with the judges ratings was .66. These results are
valuable because the only common f a c t o r , operationally , in the judges'
ratings and the зсогегз' coding was the conceptual definition which
was read to the judges and used in writing the scoring manual. The
finding of this predicted stimulus-response relationship between picture

cues and imagery in children's stories was the first small evidence of
construct validity. Furthermore, if this is evidence of validity, the
pictures and the scoring categories which are most consistent with this
relationship will provide us with the means of modifying the test so as
to focus more clearly on the common personality factor in our subjects
which made this correlation possible and which we expect to be curiosity.

(c) Subsequent modifications to the test. On a common sense level
it might be held that it does not make sense to say that Jack is more
curious than Bill unless we say what they are more curious about.
Curiosity is a function of the interaction between a person and a
situation. We have found that when different people look at the same
picture they tend to write somewhat similar stories. That is, the
pictures can be ranked in order of the percentage of stories written to
them which contain curiosity imagery. There are reliable situational
differences; so the interaction is not entirely idiographic. But
the attempt to construct a test presupposes that there is a further
limitation on the interaction. That is, there must be consistency of
behaviour of individuals across situations.

The two pictures with the highest percent imagery in each set were
chosen for a test of interpicture reliability. For neither scorer was
there a significant correlation between imagery responses for the two
pictures in any of the five pairs. However, for both scorers there
was a positive trend (for people who gave curiosity imagery on one to
give it on the other and for those who did not give it on one not to do
so on the other) in four of the five instances with a different deviant

case for each scorer. We must а1sо take into account that there is a
built-in bias against inter-picture reliability in any TAT due among
other things to the deliberate attempt of subjects to vary themes
from one story to another in response to the instructions about writing
interesting and original compositions. In this study the correlation
between percent imagery to a given picture and that to the picture immedi-
ately following it was -.42. If in spite of this negative bias we
find some consistency within subjects, then we have a reasonable
indication that the same variable is being sampled by the different
items. We could also note that low inter-item correlations are a
psychometric advantage so long as the total correlates with external
criteria. However, the high degree of situational specificity in-
dicated that the validity of the test would be effected by the sampling
of situations by picture selection.

It now appeared that pictures selected for future use should
(1) vary qualitatively so as to sample different types of situations
In which curiosity behaviour might be imagined, (2) give a percentage
of imagery in stories which would enable reliable discrimination between
individuals—say 30 to 70 percent, and (3) produce stories which could
be coded reliably. The eleven pictures which were selected for later
use satisfied these criteria. All had obtained from30 to 66percent
imagery, and interscorer agreement was found to be significant for all
pictures with greater than 20 percent imagery.

Some modifications were made to the scoring manual after (1) a
study of interrelations among the scoring categories and between the

the categories and the five senses of the conceptual definition, (2)
discovery of the most typical categories of imagery for high cue and
high percent imagery pictures, (3) discussion of coding difficulties
between the scorers.

The scorers had recorded for each story the categories of imagery
and the senses of the conceptual definition illustrated by it. It
was therefore possible to find out whether any category of imagery
tended to be correlated with any other category or definitional sense.
If a person writes a story containing imagery scored under a given
criterion, is any other curiosity imagery in his stories likely to be
scored under the same or under different criteria? If two stories
of the same person illustrate different categories of imagery, is there
evidence that the categories tend to cluster in any way, and does the
scoring of any one category tend to reduce the probability of any
other particular category being scored in stories written by the same
person? The built-in bias against interpicture reliability makes it
difficult to answer these questions. However, it was possible to
exclude one definitional sense ('Careful attention to a learning
task....’) as being inconsistent with the others. When the wording
of this definitional sense was used to define a class of instrumental
acts consistent with other senses of the definition it was found to
correlate with other forms of imagery.

When we compared imagery in stories written to high cue and high
percent imagery pictures with stories written to the other pictures it
was found that stories coded as including 'cognitive instrumental

activity’, 'careful attention to a learning task', and a type of
affective excitement which could not be classified as either positive
or negative were more common among the high scoring group.
The most important difficulty which appeared from discussion
between the scorers was whether to score for curiosity in stories
where the imagery in question could also be scored as instrumental in
an achievement motive theme especially in response to pictures of
'scientists' or 'inventors'. The manual had included a prohibition
against scoring for curiosity when the evidence in a story formed part
of a behaviour sequence terminating in the attainment of the goal of a
different motive. It was decided to relax this prohibition to allow
for the scoring of curiosity except when evidence of the other motive
occurred in the story prior to any evidence of curiosity. At a later
stage this restriction was removed entirely.

Altogether there were only slight changes. Their significance
lies in the general tendency to give greater emphasis to what might be
called the area of intellectual curiosity in which cognitive acts tend
to be rewarding in themselves. It should be noted that although these
changes have theoretical significance they did not result from considerations
of theory but from attempts to increase internal consistency.

At the same time as I was making decisions about scoring criteria
I was developing the expectancies which provided the bases for those
decisions. It was all part of the same process, but it was not quite

circular for new information was being added all the time. Some modifications
were the result of data contingencies (e.g. those made at the
conclusion of the first pilot study) but others derived from hypotheses
formed on the basis of information gathered quite apart from the experimental
program. In this way the investigator is to some extent his own
instrument and what he does with this instrument is at least as important
a part of the story of development as how he analyses experimental
data.

This phase is a continuation of phases 1 and 2. It consisted
mainly of background reading apart from psychological literature and
of examining TAT stories in the light of impressions gained. From the
beginning I had been interested in the part which curiosity played in
the development of knowledge and I hoped that the test would measure a
variable which would help to explain such innovations. The attempt
to increase internal consistency in the first pilot study had further
oriented the test in this direction.

We can think of knowledge as an ordering of symbols and of new
knowledge as a new ordering. Symbolic manipulation takes many
different forms, e.g. witchcraft, science and religious doctrine. A
year earlier I had done a cross-cultural study of primitive theories of
disease, witchcraft and family sleeping arrangements. I now read
several books relating to witchcraft in Europe from the fourteenth to
the seventeenth centuries, the history of mathematics from earliest
times, the history of modern science (especially the sixteenth and
seventeenth centuries), early Christian heresies (particularly

gnosticism), and the principal figures in the Reformation. I was
aided in some of this by my previous training in theology and
philosophy. None of this reading can be considered very thorough and
it continued through the remaining phases of the investigation. I
cannot say exactly how these things helped me, but a few points which
seemed important were:- the coincidence of displacement and projection
with compulsivity in witchcraft, the socially disruptive
effects of striving after and gaining new knowledge, and the necessity
for relative security before such striving can take place. These considerations
led to hypotheses which were tested in later investigations.
They also suggested a radical change in the scoring system.
It seemed that curiosity behaviour could take place only when
gratification could be delayed for a sufficient time for appropriate
instrumental acts to take place. Further, the drive-cue-response-reward
schema appeared inadequate because the apparent instrumental acts
could not be distinguished from the goal response. We could describe
them as instrumental in the sense that they resulted in the attainment
of a goal of integrating new information with a symbolic ordering of
prior experience; but once this symbolic manipulation has taken place
there is nothing more to be done. This might suggest classification
of curiosity as an aversive motive, but it is inconsistent with aversive
behaviour for the subject to remain in the situation of tension performing
acts which are not the most efficient (in terms of time and
energy) in removing a noxious stimulus.

the first pilot study and to adopt a new one consisting of only three
simple categories of imagery : (1) Arousal stimulus (internal or external),
(2) Intra-psychic response (excitement, wonder, fascination
etc.) and (3) Instrumental acts. The new manual included categories
of imagery provided for in the version of the old manual which resulted
from the first pilot study but all those categories relating to goal
attainment were left aside.

In an experimental group of junior high school children in a Boston
suburb several pictures of strange animals (kangaroo, wombat, platypus,
echidna) were shown and the children were given a little information
about each before being asked to write down any questions they had
about these animals. The control group was instructed to write a
short essay showing what they knew about beavers. Both groups were
then administered the curiosity TAT using pictures selected in the first
pilot study. When the stories were scored for presence or absence of
imagery using the simple three category scoring system mentioned above
the experimental group showed a slightly higher incidence of curiosity
imagery in their stories than did the control group.

The result of this attempt to produce curiosity imagery in TAT
stories by experimentally arousing the subjects is shown in Table 1.
The critical ratio of the difference between the means for the
experimental and control groups in the last column of the table is
2.20 (p = .014 using one tail of the nomal distribution). Although

Number of Stories
Without            With                                                                 Mean Curiosity
Curiosity           Curiosity           Total                Percent             Imagery Stories
Imagery            Imagery                                    Imagery            per subject

Experimental
Group (N=81)
110      214                  324                  66                    2.66
Control Group (N=69)
119      157                  276                  57                    2.28
Total (N=150)
229      371                  600                  62                    2.47

the effect is small it is sufficient evidence of validity to provide a .
basis for further selection of scoring criteria.

The need at this stage was to refine the scoring manual so as to
discriminate between subcategories of imagery and then discover which
types of imagery accounted for the difference between the experimental
and control groups. The total sample was divided into two sections
before scoring. The first section of 78 subjects (39 from each group
scored in random order) was scored by recording the evidence for each
of the three categories of imagery defined by the revised manual.
The evidence was then examined and subcategories defined. The second
of 72 subjects (42 experimental group and 30 control group) was scored
according to the same principal categories but with the appropriate
imagery more clearly described by the definition of subcategories.
The percentage of stories coded as containing curiosity imagery was
slightly higher in the second section, but the difference between
experimental and control groups was approximately the same. The data
in Table 1 are from the whole sample of 150 subjects. Further
analysis was applied only to the second section.

We could explain the outcome of the experiment by saying that
the treatment of the experimental group subjects made them more curious
than the control group and that they therefore wrote somewhat different
stories. However, it is unlikely that the treatment resulted in in-
Creased curiosity for all or even a majority of the experimental group
subjects. In order to discover the detailed effects of the arousal
conditions it would be useful to know which subjects in particular were

more curious. The questions written by the experimental group subjects
after they had seen the pictures of strange animals provided us with an
independent criterion of high curiosity.

Simply asking questions upon instruction need not indicate
curiosity. After taking into account the number of categories of
imagery scored, each story was given a score on a 0 - 4 scale, and these
were added to give a score for each subject. The correlation between
subjects' total scores and the number of questions asked was r = -.05.
(N=42). Inspection of the questions revealed that some indicated a
greater degree of prior thought about the subject matter than did
others. ‘How big are they' is clearly a different type of question
from 'Why do they make burrows under the water?' When questions
beginning with 'Why' were excluded the correlation between curiosity
and the number of questions asked was significantly negative (r = -.33).
Children who asked at least one question beginning with 'Why' had a
higher mean curiosity score than those who asked only other kinds of
questions (means 5.5 and 3.5, t = 3.0, P < .01). It was then found
that certain subcategories of curiosity imagery occurred disproportionately
more often in stories written by children who asked questions
beginning with 'Why'. For example, the intrapsychic response subcategory
of 'wonder' was often found in these stories, and it was not
always found in association with an arousal stimulus specifically
described as strange, unusual, etc. so it was then included as a new
major category. Further examination of the differences between
stories told by 'why-askers' and others suggested some categories of

imagery which had not previously been recognized. For example, the
cue-response sequence category (see manual) was defined after finding
that certain types of instrumental acts and intra-psychic responses
of excitement followed the introduction of novel stimuli in the story.
Thus important modifications could be made to the scoring manual.

Similarly, changes were made in the scoring system by discovering
what was most typical of the experimental group compared with the
control group and of high scorers in general compared with low scorers.
Rules for scoring each category of imagery were then written down to
make the third revision of the scoring manual which was then substantially
the same as the final form given in Chapter 3.

It would be tiresome to spell out how each scoring criterion was
discovered. The method was briefly as follows. Firstly, I selected
those subjects who had imagery coded on only one story, and taking each
individually, I compared his story with others written to the same
picture by subjects who scored on more than one story. In making
these comparisons I tried to find differences, if any, between the
evidence for the category scored in the 'inconsistent' case and evidence
for the same categoiy in the 'consistent' cases. Secondly, I made
the same kind of comparisons after selecting those who had scored
only under one Category of imagery. Suggested differences were then
checked by comparing experimental and control groups. If I found a
case in which the evidence for scoring the category in question did
not differ substantially from those instances in which it occurred
together with other forms of imagery, I looked at other stories written

by the same subject and at its context in the story where it was
scored in order to find other forms of evidence which had not been
coded. When any possible new forms were suggested I looked for them
in the stories of high scorers. Similarly I examined the context and
a person's other stories for contrary evidence in cases where it seemed
inconsistent to score for imagery. Thus, for example, I found that
the variance of scores was increased and difference between the experimental
and control groups made greater if I did not score for
imagery when a fear reaction dominated a possible approach response or
when the seeking of information placed a person in a dependency
relationship to an authority figure — that is, when these were present
the person was unlikely to give additional evidence of curiosity and
when they were absent his score was likely to be high. These explorations
were carried out with the second section of the sample.
I then reviewed the evidence from the other section and the first pilot
study.

When the revised manual was used for a new set of stories the
scoring reliability was .91 (correlation between person scores, sum
across four stories, for two scorers, N = 30). It was found later
that undergraduate students with no knowledge of the theory behind
the test could learn to score with reliabilities of this order in
about one hour.

Example stories were added to the manual later and some slight
changes made to increase clarity but there were no further changes
of theoretical significance.

It can be said that the scoring criteria for curiosity were
selected so as to maximize correlations of imagery with (1) judge's
ratings of picture cues, (2) experimental arousal treatment of the
subjects prior to testing, and (3) the types of questions asked by
children in response to a novel stimulus. These correlations were
discovered when the test was in various stages of development. They
are not correlations of the test in its final form. They were an
important part of the general strategy of progressive modification
to increase internal consistency and correlations with apparently
relevant variables. However they are only part of the strategy and
it would be a mistake to conclude that these three correlates are the
external validity criteria of the test.

I avoided the single criterion method of validation for several
reasons: except perhaps for young children, curiosity is probably
dependent to a large extent upon fairly idiographic interactions
between people and situations and therefore the more specific the
criterion situation the more the test will apply only to particular
people; the more specific the validity criterion the less the chance
of finding out something we did not know before; if there is a general
trait (or motive) of curiosity it is likely to be some aspect of a
person's approach to situations in general or at least to be evident
in responses to a large variety of stimulus conditions.
The test was developed as a research tool and not for the purpose
of predicting any predetermined class of behaviour. It bears the

name ‘curiosity’ because that is the word I started with, that is
what it looks like, and the things it is related to are (at least some
of them) things we expect to be related to curiosity as we normally
think of it. No more can be said of a name than that it is readily
understandable - that hearing it arouses approximately the same expectancies
in different people and that these expectancies are not
unreasonably disappointed. There is sufficient evidence to ensure
that not too many people will be disappointed. The evidence to which
I refer is mainly from later investigations but it includes the correlation
with picture cues, asking 'Why' and the arousal effect
found in the pilot studies. That is, the use of the test, both in
its developmental stages and later, provides the evidence for its
validity by setting it within a network of empirical relationships.

Curiosity is defined operationally by the following scoring criteria
when applied to imaginative verbal productions written under appropriate
conditions. If the pictures are different from those we have
used, or if the test instructions or general administration conditions
are changed, there is no guarantee that these criteria will be valid.
All criteria should be applied literally, explicitly and, whenever
there is cause for doubt, conservatively.

Every care must be taken not to be influenced against scoring a
statement for curiosity because of the context in which it occurs.
We assume that curiosity can be part of an almost infinite variety of
behaviour sequences.

The criteria should be applied whether or not they seem to be
sensible in any particular instance. At the same time, it is quite
impossible to set down criteria which will apply unambiguously to all
possible imagery. There is some room for interpretation of the
criteria, but this must always be done with caution and the evidence
in the stories should always be taken at its face value. The
criteria are written in terms of statements which appear in TAT
stories and not in terms of what might be inferred from them — i.e.,
the question to ask is not "Is this statement evidence of curiosity?",
but rather "Is this statement the same kind of statement as those

which we know to be evidence of curiosity under this particular
criterion?" The task is to decide whether or not a piece of imagery
belongs to one of several classes defined by the manual and to assign
it to the class of unrelated imagery if the evidence is not clear.
If this general strategy is followed there will be very little disagreement
between scorers. The scoring reliability should be .90 or
better.

Abstract comment (e.g., "curiosity is a valuable asset") is not
scored. All criteria apply only to explicit description of objects
and people, of behaviour and experience, in the story. Some particular
object must be described or at least one person performing a certain
act or experiencing some subjective state. The behaviour and experience
which are scored can be that of any person in the story.
Each category of imagery is scored only once for each story.
That is, the scorer decides whether a story meets the criteria for each
category. Initially, he*
[*Please excuse apparently single gender pronouns. This was written in 1961 before conventions changed.]

pronouns might have to consider each category in turn,
but with a little practice the decisions can be made while reading a
story only once or twice.

No piece of evidence can be scored for more than one category.
When a category is scored the scorer must be able to note exactly
which words in the story he took as evidence, and then further categories
can be scored only if other words can be taken as evidence
for them.

Sometimes evidence appears which satisfies the criteria for more
than one category of imagery, but only one can be scored. The rule

is to choose so as to maximize the total score. This is usuallyaccomplished
automatically by giving the major imagery criteria precedence
in the order in which they appear in the manual and by prefering
minor imagery criteria whenever one major category has been
scored. More specifically, the first task is to decide whether
any of the major categories can be scored and then whether any minor
criteria are satisfied. If no major categories can be scored there
is no problem; and if only one is scorable then it must be scored
regardless of whether scoring it excludes minor categories. If
minor categories can be scored and there is also a choice between
major categories, choose the major category which does not require
evidence that would be needed to score a minor category and then
score all possible minor categories before scoring additional major
categories. (In this way, for example, we might prefer Co to CR
and then score Ov as well thus gaining three points instead of twcw)
Should it happen, in any way, that some evidence allows a choice
between categories one of which could also be scored from other
evidence, then choose the category which cannot be scored otherwise.
In general, if there is a choice, maximize the total score.

There are five principal categories of curiosity imagery:
wonder-interest; perceptual investigatory acts; cognitive acts
instrumental in problem solving; exploratory role behaviour; and
a cue-response sequence. These provide the criteria for deciding
whether curiosity imagery is present or absent. The other minor

criteria provide (together with the major criteria) the basis for
increasing the score and can be considered subcategories which if
scored alone give only Incomplete evidence of curiosity imagery.

(Wl) Wonder-Interest. Any statement of the kind that any person in
the story is wondering about, interested in, curious about, fascinated
with any object, idea or event which is the focus of present attention
can be scored for curiosity imagery. The subjective state of wonder
etc. must be described specifically and it must have an object.
The object is usually something strange or novel but it does not have
to be so described — it is only necessary that its presence be referred
to once. Long term, interests are not scored under this
criterion.

1• An apparently middle aged man is skin diving. He is not
on any particular assignment due to the fact he has no lungs to
breathe by. The tankers in the foreground are mysterious and
arouse his curiosity. He wants to explore the wreckage but
hasent the oxygen supply to do so. I presume he will accumulate
the equipment and finish his exploration.

2. The old man was looking at the book with fascination. In
his days there weren't so many words. He trying to figure out
the words and their meanings. He probably is thinking to himself
how things changed since he was a boy. He probably
realizes what he's missed and how the generation is learning
more and more things. To him the book seem strange and full
of new things. He never learned these things in school.

3. Will It Work.
As Unika the scientst looks at his new discovery he is
thinking if it will work. If the electron in too metals can

turn back the time in a man's brain. He is also triing to
find a Guiney pig. He is thinking of his nurse who is very
old and who he hates. He tries it. It works the nurse in
back, back into the past of younger years. Dr.Unika finds out
many things one of which are so unbeleivable. The nurses
mother has some Black magic in her so the nurse also has some
but she has forgotten it and now the doctor knows. Knows all
Knows how to make the dead come back to life.

(Pe) Perceptual instrumental acts. Any description of perceptual
investigatory behaviour, which has the general goal of finding out
something, can be scored without its purpose being specifically
stated. Perceptual investigatory acts are usually either orienting
responses or instances of careful attention. Orienting responses
are distinguished from other perceptual events principally by more
active verbs: e.g., "looking at" rather than "seeing", or "listening"
rather than'hearing", unless there is direct evidence of an active
focusing of attention such as "trying to hear what is going on in the
next room". Distraction from a task because of a small change in
the environment or a sudden change of attention are usually orienting
responses. Careful attention, besides often being specially described
as careful, is found when attention is maintained over a considerable
time and when verbs like "to examine" and "to observe" are
used:

Doctor Murray, when presented with an very unusual and
beautiful sea-shell was absolutely stunned. He just took it in
his hands, and sat at his desk and examined it for about three
hours. But, since he could not make out as to what type it
could be, he stood up and giving one last look at it went over
to his wastepaper basket and threw it in.
Never again did Dr. Murray have to worry about what it was
or where it came from.

5. The man is searching. Looking for an answer to the
question—What's under the sea? He's searchings bring him to
an old antic tank. His curiosity makes him want to explore
this object. Diving deeper and closer to it he looks at it
с1озеr. He finds an interesting object tinit and brings it
above to show to the people in his boat. He finds out this
he is clutching in his hand gold. Consequently, everyone is
eager to еx1оге and see if more gold can be found. So,
searching is done some more.

6. (see below) "...observing a piece of work-Sculpture....to
observe it more carefully."

(Co) Cognitive instrumental acts. There is a large variety of evidence
for this category of curiosity imagery. In all cases there must be
some physical or mental object upon which attention is focused. There
must be something to be investigated or some problem to be solved.
"Thinking" in a problem solving situation is a common instance, but
it must be instrumental or directed towards making known something
which was not known before. "Studying" is not scored unless the
object of study is mentioned (e.g. "He is studying the habits of these
insects.") "Reading" poses a special problem. It is often necessary
to examine the context to discover whether or not it is instrumental.
Reading" for fun" is scored, but reading as a duty is not. Any kind
of hypothesis building is scored (e.g. "He thought, 'That might be a
message from the lost ship.*? In general any cognitive act which is
investigatory or problem solving will be evidence for curiosity imagery.

6. A man is observing a piece of work – Sculpture on his study.
He found this and went to his study to observe it more carefully.
He is thinking of what it is made of. He wants to find out the
source of it. He is going to go to one of his books and look up
some information on its category. He will find out that it is
unusual land will try to learn as much as he can about it and then

7. An old man is reading a very interesting story in the park
About the depretion. After he was done reading he went home and
Planted a vegetable garden and in a few weeks he had many vegetables
to appease his apetite.

(EB) Exploratory role behaviour. To score this category the role of
A person in the story should be identified (either explicitly or
culturally) as exploratory and the description elaborated with details
of role behaviour. The mere fact that a person is said to be explor-
ing or in the role of an investigator (such as "scientist," "detective"
"archeologist") is not sufficient unless exploratory behaviour is
included in the story. A person acting out an exploratory role need
not be a professional worker in that role. He could be taking it for
a short time "just for fun" (e.g. amateur skin divers might sometimes
be described as explorers). Stories of inventors are scored if some one
is engaged in specific acts of investigation whether or not these
acts are instrumental to the success of the invention. Description
of a person's interests might contain evidence of an exploratory role.
Statements such as "to find out" or "let's see if" can indicate an
exploratory interest. Sometimes exploration is instrumental to thegoal of collecting unusual objects and stories of this kind are scored
provided t h a t some detail of exploratory behaviour is included.

8. Under water epedation
There was a frogman under water. It is a calm day for there
were no curunts and everything was serene. But what was in the two
tunnels. Was it just two large pipes with nothing but water in,
or was there some hidden secret; some strange being in side. The
man will go inside, into the unknown, where he will explore, what
is to happen? will he come out, and if he does in how may pieces.
It is men like this who love adventure who will do the daring.

9. The army frogman is 50 fathoms beneath the sea he is looking
for something, he is Peter Smith the famed demolition expert.
He finds the abandoned, and "sunten" ship in the main salon of
this ship is a bomb that he has to find.

10. Russ Johnson, a geoligist had been away in Africa for 3 years,
looking for the remains of an old tempel. During these years,
away from his family and his friends, and away from the University,
and his colegus, Russ often "wondered, was it worth it." Now
that he is back, and has brought back a cup, one of the richest
treasures every brought back from Africa, Mr.Johnson, facing his
fellow collgues and geoligist, says "yes it was worth it."
(Note: Although the role behaviour is not consistent with the role
name, we assume an appropriate exploratory role is designated in-
spite of the author's limited vocabulary.)

ICR) Cue-Response sequence. There must be two separate statements in
the stoiy (although they might occur in the same sentence) which refer
to a cue and to a response.
The cue can be any event which is specifically described as strange,
unusual, or novel and can be either environmental or intrapsychic,
except that sudden changes in the environment are often clearly unexpected
and can be scored without being specifically described as
unexpected. Sometimes statements like "He was struck by the idea
that..." occur and these are scored. Reference to objects which are
culturally defined as strange etc. is satisfactory evidence for a cue.
A need, wish, want or desire to know or learn can also function as a

cue. There are two broad classes of responses which complete the
sequence: one is overt, the other covert. The first is composed of
instrumental acts (other than perceptual or cognitive acts) such as
manipulation of objects, asking questions, bodily movements towards
but not away from the cue stimulus, any type of experimentation, or
any attempt to remove uncertainty (caused by the introduction of the
cue) through the gathering of more information. Covert responses
include excitement, surprise, need to know or find out, or the arousal
of a long term interest in a specific class of objects or ideas due
to the introduction of the cue. It is important to distinguish
between a general state of excitement or surprise and a fear reaction
(which is a contrary indication). The crucial difference is between
approach and avoidance orientations and can usually be determined
from the nature of subsequent activity.

The important point in scoring this category of imagery is that
two sets of evidence must be identified. Both cue and response must
be present in the story. The order in which the evidence occurs is
of no importance provided that there is a stimulus-response time
sequence. The two sets of evidence need not be entirely independent:
the unexpectedness of an event can sometimes be inferred in part from
the nature of the response and the instrumentality of response might
be inferred partly from the novelty of the stimulus. However, there
roust always be some independent evidence of both cue and response.

11. He could be a college professor discussing something new at
his private office with someone else who might know more about

that object in his hand: that he might not know about. After
he gets his information he will right up about and tell it to
his class or give a lecture on it and compare it with something
else a lot more familiar.

12. Joe Smith was a expert skin diver, but one day he ran into
some trouble. While diving for a mysterious anicent lost
treasure box, supposely on the ship Lulu Bell, which went down
in 1784-, he came upon a giant squid (you know what I mean) and
it attacked him. Fighting with all his might was useless
against this sea monster, until he managed to take out a flair
gun which squids hated the flash from it. With all his remaining
strength he shoot it four times and the squid disappeared.
Joe then after a few days rest returned to the bottom and
started to search again. By nightfall he found is treasure
and lived happily ever after.

13. An underwater diver who does it for his hobby stumbles
across a wreck ship or something and he goes to have а look.
While he is looking he could find something vary interesting
and spend a little time fooling around with. After he gets
home he would call the coast guard and tell them all about
and ask if he could keep the part he wanted.

Satisfaction of any of the major imagery criteria is sufficient
to determine the presence of curiosity imagery, provided that none of
the contrary indications described below is found. Presence or
absence of curiosity imagery is the most important distinction which
can be made with the scoring system. However, we are able to dis-
tinguish further various degrees of curiosity evident in TAT stories.
Each story can be scored on a five point scale by combining the scoring
of major categories ,minor categories and contrary indications.
The method of combining categories is not one of simple addition and
it is necessary to maintain the distinction between major and minor
categories in order to obtain a total, or scale, score

There are three minor categories of imagery which are really subcategories
in that the criteria for them comprise partial definitions
of some major categories. The category Ix (intrapsychic response of
excitement etc.) can be scored from evidence which would provide part
of the basis for scoring curiosity imagery CR. Similarly, Ov (overt
instrumental acts) and Si (strange or interesting stimulus) have
criteria which are identical with part of the definitions of exploratory
role behaviour (EB) and cue-response sequence (CR). If no major
category is scored but the criteria for one or more of the minor
categories are satisfied we have incomplete evidence of curiosity
(or "doubtful imagery"j but note that this does not provide a compromise
for the scorer who is unable to decide whether imagery is
present or absent.) On the other hand, if a major category is scored
and there remains additional evidence of curiosity the total score can
be increased by scoring either other major categories or minor
categories.

The following minor imagery criteria should be applied in preference
to those of the major categories which they define in part
whenever the story contains independent evidence sufficient to score
any one of the major categories. This is part of the general principle
of maximizing the total score.

(Ix) Tnt.ra-psychic response of excitement. Criteria for this
category include those referring to covert events in the Cue-response
sequence category. The distinction between stimulus and response is

somewhat arbitrary here. Arousal of a need or desire to know might
be a response but it also acts as a stimulus. Descriptions of
behaviour which are culturally equivalent to specific descriptions of
a subjective state of excitement can be scored under this criterion —
e.g. "Ah...Let's see," or "Look! John. Look! Look what's here!"
However, one should not infer excitement, surprise etc. from an
environmental event. In general, we include any statement of excitement
or surprise, or need, wish, want etc, to know or learn, or the
arousal of a long term or exploratory interest in a specific class
of ideas or objects.

14. Howard's grandfather was an old man, but not a sad one. He
knew of all the wonderful places in the world and he would tell
Howard all about them. This would amaze Howard and one day he
asked his grandfather how he knew about all these wonderful things
in the world.
grandfather Smith was happy to see that Howard was interested ,
so he took him into a small room in the attic of the Smith's house.
In this room were shelves and shelves of books. Howard was
anxious to learn about these books and of the stories they contained,
but because he wasn't old enough to read his grandfather
read to him And this was how they both spent many afternoons
until one sad day Howard's old grandfather died. With him died
the wonderful stories he would read. Howard found his grandfather
again three years later when he had learned how to the
books and stories that they both loved so much.

15. "What can be wrong with the formula, said the scientist.
"I've tried everything," he said. "Ah; but wait. I haven't
tried everything. My one more chance. Let's see Marvelous,
terrific. It works! I've done it! Now everybody will be
able to wash there dishes with an electric dish wahser.

1. (see above) "He wants to explore..."
6. (see above) "He wants to find out the source of it."

those for an environmental cue in the cue-response sequence category.
Description of an object, event, person or idea as strange, interesting,
unexpected, peculiar, novel, curious, etc. or reference to an object
which is culturally defined as having those properties should be scored
Si. If something or someone is said to be interesting or curious
(in the sense of peculiar or strange) we score Si, but if someone is
interested in or feels curious about such an object we score WI. This
Si category deals with the attributes of things whereas WI is concerned
with qualities of subjective states, and one should not be inferred
from the other.

16. A friend of mind, who is an Arciologest, gave a lecture on
what he found in the Holy Lands, He brought with him a vessel
which he 3aid was over thousand years old. We looked it over
as he went on to tell that it came from a Town, which had many
such articles. He gave us more infomation and then left the
lecture hall.

2. (see above) "...strange and full of new things..."
3. (see above) "...things...so unbelievable..."
5. (see above) "...an interesting object..."
8. (see above) "...some hidden secretj some strange being..."

(Ov) Overt instrumental acts. Instrumental acts other than those which
are perceptual or cognitive and which have the general purpose of finding
out something are scored for this category. Acts which would be
part of exploratory role behaviour (EB) if the role were designated
or overt instrumental responses which would complete a cue-response

-51-

sequence (CR) i f the cue were described are included and the scoring ,
Criteria for those acts under EB and CR can be applied here.
Example stories:
17. He could be a man like Tomas Edison discovering the light
and getting it to work. First he would study about the problems
if he could and then...prints of it and take notes while he is
experimenting and the picture could finally show what he looked
like after his work had been successfully completed after a
long hard time.
18. I picked up the piece of rock that was on the mantel I
guess that's what it was. I stood their holding the object
for the longest time. I think all together I looked at the
object for five days. My family thought I was crazy wasting
my time like that. My picture was in all the papers. I
finally realized what it was. It was a rock after all!
1. (see above) "...he will accumulate the equipment."
5. (see above) "Diving deeper and closer to it..."
14. (see above) "...he asked his grandfather how he knew..."

Contrary Indications
Some stories which contain curiosity imagery also have evidence
of disruptive anxiety* associated with the curiosity theme and would
be invalidly scored as evidence of curiosity if these contrary in-
dications were not taken into account. They are:
1. A fear reaction to the introduction of a novel stimulus or
to the situation in which it occurs provided that the fear I not over-
* It will be noted that the contrary indications are rather extreme
instances of various types of avoidance responses motivated by
anxiety. Just how far we could move from these extremes toward a
more general scoring of anxiety is a matter for further empirical
study.

-52-

come. It could be specifically described or it could be an avoidance
response which is usually indicative of fea, but is not scored if
there is a later approach.
2. An outcome to the story which contains punishment imagery such
as death or destruction of something valuable to a person who has performed
curiosity behaviour in the story.
3.An affective tone which is predominantly negative—despair,
disappointment, guilt, hopelessness. There must be no statements of
"positive" affect.
4. An instrumental act which involves an inquiring person ina
dependency relationship to another person such that the act is an
avoidance of insecurity. This usually consists of a child asking a
question for the purpose of gaining attention or approval of an
authority figure before carrying out exploratory behaviour.
We are not concerned with the number of contrary indications in
a story, but only with whether or not any are present.

Example stories:
19. Dr. J. P. Bridgewater, alious Dr Jenkle was once a mad
scientist who was studying how to shrink human bodies to the
size of a mouse. He worked hard in his study trying for
years. But the day he was going to try it on someone the
police caught him and sent him to an mental institution.
Where he's still trying.
20. Way back in 1801Tony a man of 21 lived a happy normal
life he was always curious as a boy and when he heard the
another man was trying to invent a telephone he imagined he
could to. He proceeded and failed many times during he next
year meanwhile the other man had given up. When he finally
perfected it he showed it to his best friend Roger. Roger
was amazed for he know Tony would be highly regarded in their

-53-

town. On the way to the patent office Roger pushed Tony into
the path of an oncoming buggy. On killing him Roger had been
regarded as the inventor and every time he looks as the picture
taken of him at work he knows he will be punished.
(Note: Story 19 is scored Co ("...studying how to shrink human bodies
..."). Story 20 is scored CR "always curious...heard that another man
was trying invent...imagined he could to"—arousal of long term
interest, and Ix "Roger was amazed... "; both actors are associated
with curiosity imagery and both are subject to punishment.)

Total (or scale) score
There is a maximum score of four (4) for each story and the score
is not obtained by simple addition. It may be read from Table 2 and
is found by applying the following principles:
(a) If no imagery is scored the total is zero (0) whether or not
any contrary indications are scored.
(b) Contrary indications reduce to one (1) the total score for
stories containing major categories of imagery and to zero (0) the
total for those with only minor categories.
(c) Minor categories alone, regardless of how many are scored,
are incomplete evidence and give a total of one (1).
(d) Any one, but only one, major category contributes two(2)
points to the total score and designates the presence of curiosity
imagery.
(e) Other major and minor categories contribute one (1) addition-
al point each up to a maximum total of four (4).
The overall total (test score) is given by summing the scale
scores of several stories. The test usually consists of four stories,
with test scores ranging from 0 to16, but this may be varied within
the limits of acceptable reliability.

-54-

TABLE 2
Curiosity TAT Scale Score
Heading for table columns.......Number Major Categories Scored
...............................................0          1          2          3-5
Number of                    0          0,0       2,1       3,1       4,1
Minor Categories          1          1,0       3,1       4,1       4,1
Scored                         2,3       1,0       4,1       4,1      4,1

If any contrary indications are scored the total score
a story is given by the second figure in each cell of the
table; otherwise, the first figure is the total score for
the story. The total for the test is obtained by adding
the scores for all stories.
The stories are reproduced as they were written without correction of spelling or grammatical errors.

This theory of individual differences in curiosity relates to the
interaction between cognition and emotion. To the extent that
affective terms are involved in postulates relating to purposeful
behaviour sequences it can be considered motivational. But no
curiosity motive is suggested. Curiosity behaviour does share
some characteristics in common with what is generally considered
motivated behaviour; but on closer examination it appears that the
language of orthodox motivation theory becomes largely irrelevant.
The most outstanding characteristic of curiosity is that the
means-ends relationships of instrumental behaviour do not apply.
If the highly curious person is pragmatic at all the pay-off he seeks
is intrinsic to the behaviour itself. We can say that curiosity
behaviour is instrumental to the removal of uncertainties. Yet it
is not so much the aim of removing uncertainties but the means or
strategy of dealing with them which distinguishes the highly curious
person. To the extent that it can be understood by reference to
the kinds of instrumental acts employed in the removal of uncertainties,
curiosity can be considered a habit or perhaps more
generally a style of behaviour. The theory presented in this

Category system. We find it easier to understand human behaviour
if we assume that people have a cognitive map of the world - that
everybody carries around with him some kind of symbolic representa-
tion of the world in which he lives. This cognitive map is, in
part, a summary of experience. It is not like a tape recording of
events in the past, but rather a simplified grouping of experiences.
We might speak of these groupings as concepts (rules for the use of
symbols) or perhaps as categories organized in a set or system more
or less corresponding t o events which have been experienced.

Signal. A stimulus in the most general sense might be considered
a signal . We include covert events of thought and emotion as well
as those happenings i n the environment which activate sensory
receptors. However, mere activation of receptors is not sufficient
for a stimulus to become a signal. A signal must at some time become
the object of attention - there must be some interaction between peri-
pheral and central processes.

Coding operating. Signals are coded by allocation to some place in
t h e category system. We "make sense" out of present experience by re -
ferring it to some representation of past experience. This is the
coding operation.

Efficiency of a category system. There are as many different category
systems as there are organisms capable of producing them. Sometimes
the differences are small and sometimes great. One of the major
sources of difference is the variety of environments. Just as two
people (with similar constitution) who have experienced similar environments
would be expected to have similar category systems, so a single
individual who remains in a relatively constant environment will have a
relatively constant category system - relative, that is, to one who has
a greater variety of experience. Such people will suffer few misunderstandings
and have few surprises. The probability of a signal which
cannot be coded - or for which there is no relevant summary grouping,
concept, or category - is rather small in such cases, and we can say
the category system is highly efficient. The efficiency of a category
system is a function of the interaction between a signal and a category
system and depends upon whether or not there is a category corresponding
to a particular signal. It should be noted that a maximally efficient
category system might be highly undifferentiated or it might be
composed of narrow, discrete categories in a complex organization, just
so long as the signal can be coded without alteration to the system.

Redundancy (or certainty) of a signal. Signals vary in their compatibility
with a particular category system. Signals similar to
those previously coded will be largely redundant. A maximally
efficient category system presupposes a maximally redundant signal
and vice versa. An uncertain signal is one which is incompatible
with a particular category system.

Difficulty of coding. The coding operation functions with various
degrees of difficulty depending upon the efficiency of the category
system or the redundancy of the signal. Difficulty varies inversely
with redundancy and efficiency. Whenever any degree of difficulty
is present the coding operation will involve a corresponding degree
of one or both of two processes of modification - assimilation and
accommodation.

Accommodation of a category system. Whenever any signal which is
not entirely redundant is coded the category system will be changed
unless the signal is modified to make it redundant. Accommodation
is the modification of a category system in the process of coding a
signal (see Piaget, 70).
Assimilation of a signal. Difficulty of coding may result in a
signal being modified so that it becomes more compatible with a
category system. A signal may be assimilated to a category system
and/or a category system may be accommodated to a signal, but one
or the other must take place if a coding operation is to be completed
whenever there is any difficulty of coding. Assimilation
and accommodation are complementary - the more there is of one, the

Rejection of a signal. Sometimes the difficulty of coding will be
so great that the coding operation will not be completed – the ргосезз
will be brought to an end by the signal being rejected. Rejection
could be conceived of as an extreme form of assimilation in which the
signal is so modified that it has no discriminable characteristics
and ceases to be an object of attention before it is allocated toa
place in the category system. There is also a form of accommodation
which is akin to rejection. A signal can be pigeonholed as it were
in a place of its own without any modification of previous categories.
However such a process is accommodation in the sense that the category
System as a whole is changed.

Individual differences in the coding operation. Some people will
experience coding difficulties of greater intensity and more frequently
than others. Difficulty of coding will also vary within a person
from time to time. Some situations are more likely than others to be
the source of uncertain signals, and some people are more likely than
others to be in such situations. In any particular situation differ-
ent people will have different probabilities of coding difficulties.
Category systems vary in efficiency regarding a particular signal and
also in the probability of their being inefficient given a different
signal. For any given degree of difficulty there will be differences
in the degree to which assimilation or accommodation takes place as
well as in the probability of the signal being rejected. Sometimes

assimilation will be greater than accommodation and sometimes accom-
modation will dominate. In many such respects we might expect some
consistency within people across a large number of situations. Such
consistencies which differentiate one person from another are often
termed personality traits, sometimes motives or perhaps we might
think of them as highly general habits or cognitive styles in this
case. To a large extent curiosity can be understood as one of these
cognitive styles.

To this point no 'dynamic' properties have been considered. We
have noted several types of individual differences in cognitive styles
in the process of coding a signal. These differences derive from
different probabilities of uncertainty and different strategies in
dealing within certainty. In the following section we shall be cons-
idering the matter in more affective terms and the possibility of
making explanations by tension reduction will be apparent unless we
first accept an alternative account of the relationship between conceptual
conflict and learning.

Uncertainty and conceptual conflict. Berlyne (10,11) holds that
curiosity is not to be explained by reference to uncertainty but to
conflict which is a product of uncertainty and the absolute strength
of competing responses. It is not easy to translate his stimulus response
terminology into the terms of cognitive theory without making
a fundamental change in the proposition, but he is certainly correct

in insisting that prior learning is relevant not only at the point of
determining the degree of uncertainty but also in regard to the difficulty
of removing uncertainty.

Two equally uncertain signals could require greatly different
degrees of accommodation for the coding operation to be completed.
For example, suppose we are walking in the bush and hear something
move in the grass at the side of the track — we could be very uncertain
as to whether it was a mammal, a bird, or a reptile. On the other
hand, suppose we have an eighteenth century view of the animal kingdom
and we discover a platypus which both lays eggs and suckles its young
as well as having fur like a cat and a bill like a duck and even a
poison gland like a snake. Completion of the coding operation in the
first case could be a simple process resulting in only a small degree
of accommodation relating to the circumstances under which a certain
animal is seen, but in the second case accommodation must be very extensive
involving a radical revision of the prior category system.
Conceptual conflict is a function of the uncertainty of the signal
and the amount of accommodation which would be required for the signal
to be coded without assimilation. Conflict also has affective
characteristics, but, it will be suggested below, these are due to the
nature of accommodation.

Accommodation and learning. A category system represents previous
learning and accommodation which results in a new system is learning.
However, there is a little problem here. Immediately an uncertain
signal is perceived there is some change in the category system as a

whole but the coding operation has still to be completed, although
in a sense, perception is not complete until the coding operation and
with it any necessary accommodation is completed. But there is always
some amount of disordering before a new ordering is achieved and this
initial disorganization is a type of change in the category system
which could not properly be called learning. Learning takes place
when the new information is integrated with the old. This integration
might be more or1еззadequate and in extreme cases might be so in-
adequate as to correspond closely to the initial stage of disorder.
When there is an inadequate integration conceptual conflict will not
be completely resolved and the conditions for further accommodation
will continue even though some has taken place and the coding opera-
tion has come to an end at least temporarily. Accommodation is a
reordering of experience and presupposes some prior change of disorder
which is not accommodation but conceptual conflict and which is a
necessary part of the coding operation when learning takes place.

Conflict, tension reduction and learning. This is not the place to
write a complete general theory of learning and motivation, but a
theory which will account for curiosity must include some basic prop-
ositions about the place of tension reduction in learning. The
crucial point to be made here is that while resolution of conceptual
conflict might reasonably be regarded as tension reduction it in no
sense implies restoration of an equilibrium. Indeed, the definition
of conceptual conflict in terms of the requirement of accommodation
implies the impossibility of restoring any prior state of affairs.

Conceptual conflict can only be resolved by the development of a new
Ordering. (Removal of uncertainty by assimilation or rejection can
be regarded as a defence against conflict, but when the alternative
of accommodation is ruled out conceptual conflict as distinct from
uncertainty has scarcely developed.) If resolution of conceptual
conflict is regarded as tension reduction then this theory implies
that learning is necessary for tension reduction and not vice versa.

In more general terms, the following assumptions are made:
1. There is an inherent tendency towards systematic integration of
the cognitive map — that is, to replace disorder by order. 2. An
uncertain signal tends to create disorder, but disorder also results
from a variety of other causes including deprivation of stimuli.
3. Achievement of a new ordering always results in the integration
of new signals — although it might be that the new signal consists
only of perceiving or conceiving of a relationship between previous
signals. 4. For every state of disorder there is a signal or class
of signals which is necessary for reintegration. (In a different
language, there is a goal object for every motive.) 5. Nothing else
is needed for learning to take place.

Curiosity is a cognitive style or strategy of learning in which
conceptual conflicts are created, maintained and resolved. The
result of the strategy is not equilibrium but growth or development
of the cognitive map. In this sense, the learning which takes place
by curiosity does not depend upon tension reduction and it could just
as easily be said to depend upon tension production. This is not to

say that tension reduction theory cannot account for much in curiosity
behaviour. We shall see that it can when we come to review Berlyne's
theory in Chapter6. It is only maintained here that it is not necessary
to make any assumptions about tension reduction. Further, it
will be apparent when we consider some of the components of adult human
curiosity evident in the cognitive strategies of highly curious people,
that some of the skills employed in the strategy are learned and could
be accounted for [by] tension reduction theory. However, it does appear
that adherence to a tension reduction model creates grave problems
which are not encountered in cognitive theory; and it could well be
that an adequate account of curiosity will eventually mean the disappear-
ance of motivation as a technical term in psychology.

Pleasure and excitement. It is assumed, following Hebb (47), that
pleasure is not a matter of receiving a particular class of signals
which have the quality of giving pleasure, but is fundamentally a
directed growth or development in organization. That certain classes
of signals are usually pleasure inducing is due to the fact that they
correspond to certain common classes of disorganization and to their
integration resulting in a new organization. Thus need satisfaction
is deemed pleasurable. But there can be a great variety of disorders
in the cognitive map requiring a great variety of very specific signals
for integration to take place. In this sense, curiosity is not to be
regarded as another need to be added to hunger, thirst, sex, etc., but
(if we must speak of needs) as a process which includes satisfaction
of a great variety of uncommon needs. The pleasurable integration of

a signal with a category system is assumed to be a general principle
of learning, but it is from the process of integration and not the
attendant quality of pleasure that explanatory principles are derived.

Just as integration is pleasurable so disintegration is un-
pleasurable. (In the 'Outline' Freud (37) suggested that eros and
the death instinct were two basic mechanisms of integration and dis-
integration.) The function of conceptual conflict in curiosity brings
both pleasure and unpleasure, so that the affect associated with it is
an ambivalent kind of excitement.

It is convenient to distinguish between two sources of individual
differences in cognitive styles relative to curiosity. It is one
thing to characterise a curious person by the way in which he handles
an uncertain signal, but it is another problem to explain why he experiences
more uncertain signals than one who is less curious. Wondering,
looking, thinking and overt investigatory instrumental acts are
directly relevant to the coding operation. Exploratory behaviour in
which the individual creates situations of uncertainty or goes out
seeking them requires other considerations. These two aspects of
curiosity are very closely related, and the theory deals with both of
them in such a way that the explanation of one is essential to the
explanation of the other, but it is convenient to treat them separately
for the moment.

efficient category system is very unlikely to be curious. Coding
difficulties will seldom arise. Most signals will be redundant.
This, of course, applies to typical behaviour in a very familiar
situation - it is a description of a central tendency in coding operations
for a given class of signals. But given this tendency let us
consider what will happen if the class of signals is enlarged - if a
'novel' stimulus is introduced. For some people the change in signal
characteristics will receive little attention for at least one of
several reasons. The category system might be so undifferentiated
that it is just as compatible with the new signal as with the old and
the signal will be coded without difficulty; or if there is some
difficulty the signal could be immediately assimilated to the old
categories; or the category system might be quickly accommodated to
the signal; or the signal could be rejected. The highly curious
person will do none of these things.

Being curious implies that one has a sufficiently highly differentiated
map of the world to make changes in signal characteristics
likely to create coding difficulties. (This is one respect, apart
from exploratory behaviour, in which the curious person is more likely
to experience uncertain signals.) Curiosity also includes a 'liking
for stimuli' — a willingness to accept and not reject signals. This
combination of openness and discrimination will at least result in
novel stimuli becoming the focus of attention. There will be an
attempt to code the signal, but it will be a difficult operation.
Curiosity, in its oldest literal meaning and in theory must

include 'careful attention'. To confuse it with a gobbling up of
stimuli denies the part played by an orderly cognitive map. The sink
has a sieve. To be carious a person must be open, but also dis-
criminating.These two primary conditions place limits on the kind
of coding operations he is likely to perform. He wants to order
experience as well as to experience disorder and, indeed, to experience
disorder presupposes some prior order.

The interaction between order and novelty can be illustrated by
the following remarks of a scientific explorer. Writing of the
nuisance value of travel adventure to an anthropologist, Claude Levi
Strauss (51) declared "The truths that we travel so far to seek are of
value only when we have scraped them clean of all this fungus". Note
the compulsive orderliness theme of his metaphor. He continues, "It
may be that we shall have spent six months of travel, privation and
sickening physical weariness merely in order to record - in a few days,
it may be, or even a few hours - an unpublished myth, a new marriage-
rule, or a complete list of names of clans". Apparently the goals he
seeks are not so much novelty but the ordering of novelty; he seeks
to record something previously unrecorded, to discover some new rule or
to complete a symbolic system. Certainly, he places considerable
emphasis on novelty and perhaps we need not accept the implication
that the trials along the way are irrelevant to his evaluation of the

outcome. We see here some evidence of the masochism we shall suppose
to be associated with the establishment of curiosity behaviour
sequences. But to Levi-Strauss at least, novelty has value only as
it is related to order.

Concern with both the signal and his category system makes the
curious person loathe to assimilate or accommodate while he strives
to complete a difficult coding operation. These conditions create
conceptual conflict or tension which must be suffered if one is to
be curious.

To inhabit the no-man's-land between chaos and cosmos is to take
risks. One's view of the world is placed in jeopardy and functional
blindness is an inviting defence. To remain at the frontier with
eyes open requires a certain equanimity if not an appetite for danger.
What distinguishes the curious person above all else is that he
remains in that situation performing acts which are instrumental to
removing uncertainty but which at the same time delay the resolution
of his conflict.

Creation and resolution of conceptual conflicts. Berlyne (10) has
some interesting remarks in connection with the delaying nature of
curiosity type behaviour.
Ultimately, the only effective way to deal with ambiguity
is to procure additional information through exploratory
and epistemic behaviour, but this means that the conflict
must be faced and borne for a while. The conflict can
be relieved more quickly by turning from, or refusing to
attend to, the troublesome stimuli, but it will then be
apt to recur. The alternative is analogous to that
offered by other drive states. Physiological needs can

be gratified impulsively, but at the cost of baleful consequences,
or gratification can be postponed in accordance
with Freud's 'reality principle', until it can be undertaken
more safely.

The obsessive-compulsive patient is, as is well known,
given to brooding, wondering and doubting, to pettifogging
distinctions and classificatory schemes — all
of which amount to a morbid travesty of the epistemic
behaviour that we shall be taking up...

It will be noted that Berlyne came very close to the theory we
are considering. He recognized the relationship between a curiosity
and a highly differentiated cognitive map of the world but dismissed
it as 'abnormal'. He pointed out the secondary process characteristics
of curiosity behaviour but made it a less desirable alternative
under the influence of the reality principle and forced upon
the person by the 'baleful consequences' of immediate gratification.
He did not see 'compulsive gratification' as a reasonable alternative
to 'impulsive gratification'. In Berlyne's system working to remove
uncertainty should be no more satisfying than simply waiting the same
length of time. Indeed there is no suggestion in his theory that
curiosity behaviour can be satisfying in itself - it is necessary only
in order to remove a noxious stimulus . Curiosity would then be a
maximally expedient escape response, whereas we had thought it to be
characterized by an approach orientation which is not only a means to

The question arises of whether the wondering, doubting and pettifogging
distinctions of the compulsive person are not enjoyed or even
sought out by the highly curious person. Is the curious person one
who is able to tolerate cognitive dissonance as a necessary evil or
is he one who behaves in such a way as to generate conceptual conflicts?

At first sight exploratory behaviour poses a special problem.
Curiosity has been characterized as a preferred cognitive style of
dealing with coding difficulties, but is it not a different thing to
explain the seeking of novelty? We have dealt with the reaction to
a situation in which coding difficulties have arisen, but what of the
proactive characteristics which have forced people to postulate a
curiosity motive or an exploratory drive? Why should the curious
person be in situations where coding difficulties are likely to arise?
To begin, it is necessary to see that exploratory behaviour is
not fundamentally different from other forms of curiosity. Even given
the situation in which a coding difficulty has occurred we still have
to explain why the subject stays in that situation. We have said
that the curious person delays resolution of his conceptual conflict,
but why should he place his cognitive map in jeopardy? If we can
explain why a person seeks novelty we shall know why he accepts it
when he finds it. It is not a matter of what he does in a given
situation as compared with why he is in that situation. 'The
situation' here is a phenomenological term — that is, the situation
is what the subject perceives it to be. Coding difficulties might

sometimes result from moving about geographically but the highly
curious person will be likely to create them by the way he perceives
his environment generally» If he is not in a situation appropriate
to the cognitive process called curiosity he will restructure his
perceived environment so that he is. Exploration is only one way
in which he generates conceptual conflicts; or, we could say, he is
always exploring. It would almost appear as though curiosity were
a matter of liking conflict.

Hebb (47) has suggested that 'pleasure is not the activity of
particular structures in the nervous system, and not even a particular
pattern of cerebral organization, but fundamentally a directed
growth or development in cerebral organization. It is thus necessarily
a transient state of affairs in which conflict is being
reduced, an incipient disorganization being dissipated, or a new
synthesis in assembly action being achieved'. His terminology is
different but he makes the same point. He adds '...some degree
of conflict is stimulating and necessary to the maintenance of
normal responsiveness to the environment'.

As we saw when considering the role of conceptual conflict and
tension reduction, Hebb has given a valuable lead in stressing growth
or development. This could be interpreted to support the idea that
curiosity is a matter of liking conflict because conflict is a
necessary part of the growth process. On the other hand, it could
be pointed out that as a new organization is developed conflict is
reduced so that the aim could be conflict reduction. However,

reduction of conflict involving growth is very different from avoiding
conflict and thus avoiding change. It could be that some people have
learned techniques for increasing conflict by an optimal amount so as
to enjoy reduction of tension . But if reduction of tension is not
in itself the aim but only a quality of the process of integration the
picture changes. Even given the basic propositions of stimulusresponse
reinforcement theory some interesting suggestions can be made
about the development of individual differences in curiosity if we
consider how learning takes place under conditions of reinforcement
which are both positive and negative.

Responses which are punished are less likely to occur in the
future. This is a simple generalization and often applies. However,
it is not a general law. There are conditions in which the opposite
is found; punishment sometimes results in an increased probability of
the response. A response tendency has been found to strengthen or
become fixated under the following conditions of punishment :-

1. Punishment is mild. The response which is punished
continues to be emitted. The escape response is
identical with the punished response.
2. Fear is the discriminative stimulus for the response.
3. Positive reward is used to maintain a response which

All these conditions can be present in the creation and resolution
of conceptual conflicts. If such conflicts are a noxious state of
affairs then the behaviour which produced them would constitute punished
responses. If the tension is not too severe some of these responses
could continue to occur and eventually become fixated. As the response
tendency becomes stronger it would tend to carry through a behaviour
sequence in spite of increased tension which would in turn provide
further reinforcement. Fear is a primitive response to the unknown
and would therefore constitute one of discriminative cues to the
presence of a coding difficulty. Whatever response occurred would
tend to be reinforced — it might be a continuation of the cognitive
process or it might not, but if it is then the probability of future
curiosity is made greater. Finally, if some degree of accommodation
is pleasant in itself, there is a positive reward to maintain curiosity
behaviour while the unpredictable degree of tension involved in coding
operations is periodically punishing to a high degree. The imposition
of negative reinforcement on the continually rewarded response would
increase the probability of that response. If all these conditions
are common to curiosity it is reasonable to expect that strong habits
could be formed. The effect of these habits would be to increase,
prolong and resolve conceptual conflict. Subjects who have learned
these habits would give the impression to a casual observer that they
'seek out' a noxious state of affairs. They might appear to be
seeking novelty or to be motivated by curiosity. However there is no

need to talk of a motive here. To do so says no more than that they
are being curious.

This discussion of the place of learning in the development of
curiosity by no means implies that it is a result of purely operant
conditioning. Some respondents undoubtedly enter in. The orienting
response, for example, can be viewed as an instinctive component.
Some such stimulus-specific responses must be present before there is
ongoing activity within which more complex sequences are developed.
Postulation of learning by reinforcement from slightly aversive
stimuli presupposes some prior response tendencies of a different
origin. Neither is it implied in the foregoing discussion that such
primitive antecedents cannot include motivational conditions, provided
that we accept White's suggestion (84) of effectance motivation or
some such formulation which postulates that the need is neurogenic and
satisfied by a great variety of stimuli according to the function of
conceptual conflict. I am not concerned to explain the genesis of
curiosity in general but to suggest conditions for development of
individual differences in curiosity so as to facilitate our understanding
of its function in adult human behaviour.

Habits of acting so as to increase, prolong and resolve conceptual
conflicts would constitute a learned strategy of learning.
Every time one of these acts is performed it is instrumental in
giving attention to additional information. Whether this is by way
of further intake from the environment or by covert symbolic manipulation,
it could provide the means of resolving the conflict or it

could increase the tension. This playful gamble, accompanied by
ambivalent expectancies, might become a source of considerable excitement
— a source of 'compulsive gratification'. Conversely, a
strategy of avoidance will develop whenever the negative component is
too great.

The antithesis between curiosity and avoidance behaviour is
vividly exemplified by Frankl's report (34) of the intense curiosity
experienced when confronted with what appeared to be certain death in
a concentration camp. Death is obviously an unknown and since it was
certain no avoidance behaviour was possible. Curiosity was then the
only adaptive strategy. On the other hand, had there been any hope
of escape avoidance behaviour would have excluded curiosity. Integration
of an unavoidable death signal with a category system of life
must surely be the most difficult coding operation. The only alternative
to completing it is disintegration — that is, death — while
completion signifies continuing integrity — that is, life. When
the ambivalent expectancies involve the ultimates of life and death,
curiosity cannot be anything but intense.

In summary, the highly curious person is one who has developed
a preferred strategy of dealing with uncertainties which include instrumental
acts having at the same time the potential of both increasing
and decreasing coding difficulty. This cognitive style is
associated with an ambivalent expectation of excitement which terminates
in a pleasurable integration of a signal with a category system.
In the process of coding the signal, assimilation and accommodation

are delayed through attention given to the signal and care taken not
to disrupt the category system. His aim is to order experience by
placing his cognitive map in jeopardy. By contrast, the person
with low curiosity has learned an avoidance strategy. The highly
curious person will generally have a highly differentiated category
system and will also tend to structure his perceived environment so
as to increase the probability of coding difficulties. The means by
which this structuring takes place are also the instrumental acts
which result in satisfying solutions to coding difficulties. These
acts maintain a balance between assimilation and accommodation, but
the ultimate purpose is accommodation. The strategy can be viewed
as instrumental or purposive, but there are no goal responses to be
distinguished from instrumental acts and it can therefore be regarded
as intrinsically rewarding.

After the TAT measure of curiosity had been developed and basic
theoretical considerations worked out, several explanatory investigations
were made with the purpose of discovering whether the correlates
of the test were consistent with the theory. Certain predictions
were fulfilled. However, these studies must be regarded as primarily
exploratory not only because there are no previous studies in this
area but also because we sought to discover unpredicted correlates as
well as to confirm hypotheses. Together with the evidence obtained
in the developmental studies they demonstrate that the test has a
useful construct validity. In future investigations it should be
possible to proceed on the assumption that the test is a means of
gathering data which will enable the acceptance or rejection of
theoretical propositions regarding curiosity.

The studies in which the results reported here were obtained
were sometimes concerned with issues not directly relevant to the
present task of showing that the test is consistent with the theory.
I have selected only the most relevant aspects of these investigations
without, of course, omitting any findings contrary to those discussed
here.
The theory we have been considering supposes curiosity to be a
preferred strategy or cognitive style of dealing with uncertainties.
These uncertainties create conflicts which arise from and are resolved

by a concern with ordering experience and an openness to novel stimuli.
The general traits of orderliness and attention to novelty are relevant
to curiosity only as they are brought together in the performance of
appropriate instrumental acts. However, since the preferred strategy
includes these general tendencies, we can expect positive correlations
of curiosity with favourable attitudes to both orderliness and novelty.
If we can find evidence of these relationships, then the basic presuppositions
of the theory will have some support.

A measure of orderliness was derived from a questionnaire
developed at the Harvard Psychological Clinic Annex under the direction
of Henry A. Murray and used by Couch and Keniston (26). These scales
had previously been used with male college students. Our subjects
were thirteen-years-old boys and girls. The wording of the items was
carefully examined for intelligibility to children of this age, and
where necessary changed or simplified. The questionnaire as we used
it is found in Appendix 5. All items are statements containing anal
metaphores or statements theoretically relevant to the general area
of anality. Responses were on a six point scale from strongly agree
to strongly disagree.

Couch and Keniston had defined three anality scales by factor
analysis of the questionnaire items. Two of the scales named

‘socialized anality’ and 'rejective anality' which have no relevance to
the present discussion were included in our study for other purposes.
The third factor, which was bipolar and might be called 'orderliness-
messiness' was made into two scales by selection of items from the
questionnaire according to the following criteria. After obtaining
separate totals for the positive and negative items of the factor, all
item-total correlations were examined for boys and girls separately,
and (1) items which had low or negative correlations with the total
scale score were eliminated, (2) items which showed inconsistency
between boys and girls were also discarded, and (3) other items in the
questionnaire which had not previously been included in the particular
scale were added if they had high correlations with the scale total and
low correlations with other scales. The two scales of 'orderliness'
and 'messiness' thus obtained were included in our later correlation
and factor analysis study together with the Agreement Response Set
scale (modified by the same procedure) and some individual items not
included in any scales but of interest in themselves. For composition
of the scales, correlation and factor matricies obtained with the modified
scales and some interpretations of factors see Appendix 4.

The subjects were 60 eighth-grade children from predominantly
middle-class backgrounds, attending a public school in a Boston suburb,
and included a large proportion of Jewish children. The curiosity
TAT and a delay of gratification measure were administered on the same
day and the 55 item anality questionnaire two weeks later. All the
other measures used in the analysis were available from previous

testing . Because of incomplete data three subjects were excluded,
and there remained 29 boys and 28 girls.

One class had done the curiosity test in similar circumstances
but with different pictures six months previously. The test-retest
reliability was found to be.71.Scoring reliability fort he
curiosity TAT was .91 for two independent scorers.
The correlation between curiosity and orderliness was.31 for
boys and .30 for girls, and the correlation for all subjects is
statistically significant at the .05 level (or a the.01level if
we allow a one tail test). Due to the small size of the correlation
coefficient this result alone is not very convincing bu it should be
interpreted in the light of (1) the fact that it was predicted,
(2) the amount of measurement error (perhaps 50percent of the
variance) in the curiosity scores, (3) the length of the orderliness
scale of only nine attitude items, and (4) the content of the items
in the orderliness scale.

Items in the orderliness scale which had the highest corr-
elations, individually, with curiosity are shown in Table3. A
person behaving predominantly in the manner described by these items
would normally be regarded as highly compulsive. However, the
population of responses being sampled here is probably very large
and could be expected to include many adaptive strategies as well
as some which might disrupt normal functioning. It is reasonable
to expect that only a limited realistic orderliness and not the whole
general field of compulsive behaviour is compatible with curiosity.

I am a great one for picking up .23 .43
things and putting them where
they belong (30).

The first item in the table gives a behavioural paradigm of the
coding operation discussed in the previous chapter, and the other
references to system and doing things carefully also illustrate
points made in the theory. We should note that the orderliness
scale is from an attitude questionnaire and is not a behavioural
measure. The responses being sampled are part of the subjects'
symbolic behaviour or cognitive strategies. The correlation
confirms our expectation that orderliness is an attribute of the
cognitive strategy of the highly curious person.

If concern with orderliness is one side of the curiosity coin
then openness to novel stimuli or sensitivity to arousal is the
other side. Evidence to support the theory at the point of the
relationship of curiosity with this second facet was obtained in
the same study as confirmed the relationship with orderliness.
One of the items in the attitude questionnaire was 'Novelty
has great appeal to me'. In our sample it was found to be
independent of the anality scales for girls, although for boys it
could have been included as part of Agreement Response Set scale
as it was in Couch and Keniston's work with college males. This
item was included in our analysis as a variable of interest in
itself, with the hypothesis that for girls at least it should be

positively correlated with curiosity . The correlation between
curiosity and this measure of sensitivity to stimuli was .42
(z = 2.24, P = .013) for girls while for bоуз it was zero.
The lack of a correlation in the case of boys requires some
attention . One of the aspects of the yeasayer's personality , as
described by Couch and Keniston, i s the desire for novel stimulation,
the acceptance of unusual stimuli , whether internal or external . So
we might expect some relation between curiosity and yeasaying, if it
were not obscured by other elements in the yeasaying syndrome. In
fact , the correlations of agreement response set with curiosity were
- .02 for boys and -.05 for girls, while the novelty item was corr-
elated .53 with yeasaying for boys but - .07 for girls. It was only
when the appeal of novelty was independent of acquiscence set that any
correlation with curiosity was found. At this stage we can only guess
as to why this should be so, but it is probably due to the naysaying
aspects of stimulus selection. The curious person's openness to
stimuli is not indiscriminate.

We had another rather unusual type of measure which can be inter-
preted as tapping the general area of sensitivity to arousal. In a
previous study with older high school children , I noticed that some of
the subjects had written comments after the word 'sex' when filling
out the identification blank on the front of the curiosity test book-
let . These comments seemed related to the subjects' curiosity scores.
I also noticed that sometimes they answered 'boy' (or 'girl' ) and
sometimes 'male' (or 'female' ) . Recalling the psychoanalytic

hypothesis that intellectual curiosity is sublimated sexual voyeurism
it seemed at least possible that the response ‘male’ as opposed to
'boy' might be characteristic of the highly curious. The assumption
here was that the response 'male', especially for a thirteen-year-old,
indicates a higher degree of intellectualized sex interest and that
not giving this response could be due in part to repression of sexual
ideation and thus of curiosity. Our measure, named 'sex role
identity', was simply the number of times the subjects wrote 'male'
(or 'female') when giving identification data on the two testing
occasions — that is, we had a three point scale.

The correlation between sex role identity and curiosity was .39
(z = 2.10, p = .018) for boys and -.06 for girls. Perhaps we have
here some confirmation of the psychoanalytic hypothesis for male
subjects — and we might even have some evidence to support the general
suspicion that Freud did not really understand women. In any case we
can assume that sexual development in early adolescence can provide
an important source of novel stimuli. Acceptance of these internal
stimuli can be expected to be as much a part of curiosity as attention
to novelty in the environment.

The sex differences we have found so far suggest some interesting
possibilities. The sex role identity variable seems to play the
same part for boys as does attention to novelty in general for girls.
This would be consistent with some fundamental differences in intellectual
functioning and in sexuality between men and women. It would
seem that women are more inclined to take their cues from the environ-

ment than are men — as is demonstrated in studies of field dependence
(85). On the other hand, sexuality in the female is generally supposed
to be much less specifically genital, especially during the
early years after puberty, than it is in the male. Our results are
based on far too little evidence to test these well accepted propositions
but they do fit in the same pattern.

We can conclude that the evidence in favour of curiosity being
related to attention to novelty is slight but reasonably convincing
when we take into account known sex differences. The correlations
obtained are statistically significant and, indeed, quite high when
we consider how short and presumably unreliable the measures were.
Furthermore, these correlations have not been selected as the only
favourable instances; they represent the only attempts we have made
to test this part of the theory, and the results were as predicted.

Ideally, we should use a factorial design and analysis of
variance to study the interaction between orderliness and novelty in
the effect on curiosity. This cannot be done with the data from the
investigation in which the primary relationships were established
because (1) there are too few boys with high scores on sex role
identity and high scores on orderliness, (2) there are too few cases
in general, and (3) the scores are from measures of general attitudes
and not an interaction in a particular situation. The third objec-
tion might not rule out the possibility of a significant interaction
altogether, but the theory relates to coding difficulties which arise

from the incompatibility of a particular signal with a particular
category system. The sense in which the stimulus is novel is given
by a particular ordering of prior experience. Even when we have a
coincidence of high scores for orderliness and novelty from these
scales, there is no certainty that there is, in fact, a coincidence
of orderliness with those novel stimuli whose acceptance is indicated
by the scale score. We might expect that statistically there
would be more such coincidences with high scorers than with low
scorers, but a better test would be obtained with a more highly controlled
experimental study.

The best we can do with the present data is to show that a better
prediction of curiosity is obtained by combining the two facets,
novelty and orderliness, than is possible with either separately. The
multiple correlations of curiosity with orderliness and the appropriate
measures of attention to novelty for boys and girls respectively
were .64- and .53. In the case of the boys the correlation between
the two independent variables was significantly negative (r = -.39)
which meant that the predictive power of the combination was much
better than with either separately. (The proportion of variance
accounted for by the best combination of orderliness and sex role
identity is significantly greater than is accounted for by sex role
identity alone : F = 10.99, P < .001). For the girls, the two
independent variables are uncorrelated (r = -.03) and thus we have
again good conditions (but not as good as with the boys) for increased
predictive power. In a conservative test of the significance of the

increase in predictive power, we compare the variance accounted for by
the combination with the variance accounted for by the highest correlate,
novelty, and find F = 3.48 and P = .08 if we do not employ a
directional hypothesis. Taking into account that this relationship
had been predicted and that there is one degree of freedom in the
greater estimate of variance we might use a one tail t test and find
we have reached the .05 level; but, to quote Mosteller and Bush, when
this is necessary the results are 'not ecstatically convincing'.
However, the increase is statistically significant when we add novelty
to orderliness (rather than the other way around) and there is not a
great deal of difference in their two independent correlations. So
we might say that, for both sexes, the evidence is favourable to the
theory, but that more information is needed. At least we can conclude
that the lack of positive correlation between what we have considered our
two independent variables should enable a better prediction of the
criterion than would be possible with either alone; and it is certainly
true that the correlation of neither with curiosity can be accounted
for by the other.

One characteristic of the highly curious person, according to the
theory, is that he is able to maintain cognitive dissonance and delay
the satisfaction of reducing tension while performing appropriate instrumental
acts. The secondary process dominance which we have
assumed would entail a high degree of impulse control — at least in
certain situations. This does not necessarily imply that delay of

gratification, as a general trait, should be correlated with curiosity.
The delay in question relates only to the resolution of conceptual conflict.
However, in view of the results obtained by Mischel and others
(59, 60), it was thought worthwhile to test the hypothesis that
curiosity is correlated with delay of gratification as measured by a
series of simple choices.

By offering children a choice such as 'Would you rather have a
small candy bar now or a large candy bar later?' Mischel and others
have found delay related positively to social responsibility, to in-
telligence, to age, and to moderate, realistic time conception of
past and future events. Such choice behaviour has been regarded as
a simple model or paradigm of delay of gratification situations in
real life. At the time when the curiosity test was administered we
showed the children two candy bars, one large and one small. We
told-them they could have the small one now or they could wait until
the next time we came — in a week or two — and have the large one
instead. Two similar choices had been offered in individual testing
situations previously. The children received the rewards they chose.
We then had a score ranging from 0 to 3 indicating the number of immediate
or delay choices made. Eight of the 60 subjects chose the
delayed reward on all three occasions, 19 chose to delay twice, 27
once and 6 always chose the immediate reward. The means were 1.6
immediate choices for boys and 1.4- for girls with standard deviations
of 0.93 and 0.73 respectively. Raw scores were used in the analysis.
The hypothesis received some confirmation in the case of girls

(г = . 36, p =. 03) but not with boys (r = .08). Delay of gratific-
ation is correlated (.37) with orderliness for girls but not for boys,
and thus appears part of the curiosity syndrome for one sex only.
Whether this is a true sex difference or an artifact of the methods of
measurement remains to be discovered.

Two more variables in respect of which we found an indication of
sex differences were the Alpert-Haber Facilitating TestAnxiety Scale
(2) and McClelland's n Achievement TAT measure (55,4). Although we
made no predictions, these results are consisten twith our explanation
of sex differences in attention to novelty. It seems to be a matter
of the relevance of internal as opposed to external cues. Thus with
boys curiosity would appear to be increased by the internalized standards
of n Achievement but to be reduced under external pressure
which leads to higher performance for some people, while the opposite
holds for girls. For a summary of these and other results see Table
4 and Appendix 4. Debilitating Test Anxiety is uncorrelated with
Curiosity for boys (r = .04) while there is a significant negative corr-
elation (r= -.42) with female subjects. Those girls who react to
external demands for achievement with increased efficiency will have
their performance facilitated in part by curiosity, while those whose
anxiety reduces efficiency will at the same time be less curious.
This is borne out further by the high loadings of actual school
Achievement and 'extreme response set' on the girls' curiosity factor.

Curiosity (13)                                                   .69       .68
Mathematics Achievement (19)             .32       .65                   .35       .45
Average School Achievement (20)                    .31       .55                   .27       .45
Intelligence (18)                                                .36       .34                   .13       .29
Sex Role Identity (28)                                       .54       .02                   .39       -.06
Orderliness (3)                                     .25       .42                   .31       .30
‘I love to read a lot' (10)                                   .21       .44                   .17       .40
Extreme Response Set (6)                                .15       .51                   .12       .30
'Novelty has great appeal to me' (8)                  .07       .54                   .00       .42
n Achievement (21)                                          .36       -.44                  .21       -.21
v Education (23)                                               .12       .37                   .10       .12
Behavioural Delay Choices (14)                        -.03      .39                   .08       .36
Facilitating Test Anxiety (25)                            -.30      .29                   -.22      .23
Debilitating Test Anxiety (26)                .04       -.66                  .03       -.42

1 Including all variables with factor loadings greater than plus or
minus .30 in the Beswick-Metzner Curiosity-Delay study details of
which are given in Appendix 4. The factors referred to are
Factor V2 for boys and Factor I 3 for girls. Boys V2 was obtained
by rotating to obtain a single curiosity factor although for other
purposes of the study a better solution gave curiosity loadings on
two factors. For present purposes the criteria of rotation are
simply to explore the correlates of curiosity.

2.Numbers in parentheses refer to identification and description
of variables in Appendix 4.

‘Extreme response set’ is interpreted as a measure of trust or confidence
and was obtained by counting the number of times subjects
responded 'strongly agree' or 'strongly disagree' on the attitude
questionnaire (the correlation between 'strongly agree' and 'strongly
disagree' being .57). The suggestion is that girls with a sufficient
basic security will use curiosity to raise their level of performance
under stress; but boys, whose curiosity is more directly related to
internal cues, will seek new knowledge better when left to their own
resources.

The several sex differences referred to above indicate that
curiosity functions somewhat differently for boys and girls in that
it’s correlates vary with the sex of the subject. In addition, there
is a mean difference in curiosity scores. There has been a tendency
for girls to score higher than boys in all investigations. In the
study from which the above results were taken the means were 6.10 and
7.61 for 29 boys and 28 girls respectively (C.R.=1.65,P=.10).
For 53 boys and 52 girls of the same age group in Australia the means
Were 5.26 and 7.08 with a critical ratio of 2.06 and P=.04.
Similar differences were found in the pilot studies and two other
investigations.

It is a puzzle to know why girls should score higher than boys.
The higher loading of school achievement on the girls curiosity factor
(TableA) suggests that it could be due to greater verbal skill which
might artificially inflate the scoring from written stories. In the

Australian sample the girls had better grades than the boys in English
and this might seem to account for the difference, but the hypothesis
is not supported when we examine the correlations between performance
in English and curiosity scores. These correlations were -.15 for
girls and .21 for boys. Since the difference between the means for
the two groups takes up only a small proportion of the total variance,
and there is no significant correlation within the groups, we must conclude
that there is some other explanation of the sex difference in
curiosity.

Another explanation might be that girls are more compulsive neat
tidy and orderly in their work and personal appearance. If this attribute
were found in their cognitive strategies we would expect them
to have more conceptual conflicts and thus be more curious. However,
although the correlation between curiosity and orderliness is found
for both boys and girls there is no mean difference in their orderliness
scores. Thus, although we might assume that girls are more compulsive
in their overt behaviour, and we have found orderliness to be characteristic
of the highly curious person, we have no indication that sex differences
in curiosity can be accounted for by measured differences in
orderliness at the cognitive level.

The most likely explanation which appears at present is that girls,
in adolescence, are attentive to a greater variety of novel stimuli.
We have no independent evidence that there is a sex difference at this
point, but the assumption that females are more environmentally
oriented was shown above to be consistent with some differences in the

We found that the attitude item ‘I love to read a lot' was
correlated with curiosity (Table 4). In one of the pilot studies,
when teachers were being asked to describe any differences they could
discern between extreme groups on curiosity scores, the school
librarian remarked that those in the high scoring group tended to
make more use of the library. (The teachers did not know which was
the high scoring group or the criteria of selection). Cattell (22)
has named one of his factors of human motivation 'exploration
(curiosity)' although he has not used a direct measure of individual
differences in curiosity. One of the strongest representatives of
this factor is an indirect measure of the desire to read books, newspapers
and magazines. Reading is one of the 'cognitive instrumental
acts' which define a scoring category in the test manual (Chapter3).
People with high curiosity should therefore have different reading
habits from those with low curiosity.

To test the hypothesis that reading habits differ according to
scores on the curiosity test, high and low curiosity groups were compared
in the Australian sample. This sample consisted of 53 boys and
52 girls with an average age of thirteen years in the class which соr-
responds to eighth grade in American schools. High scoring and low
scoring groups on the curiosity test were selected for both boys and
girls from each of three classrooms. There were three to five

children in each group and several teachers were asked to write down
any difference they noticed between the groups without knowing why
they were placed in those groups. Details of these comparisons can
be seen in Appendix 6. The number of books which had been borrowed
from the school library was ascertained for 39 of the selected
students. The 19 subjects in the high curiosity group borrowed an
average of 37.7 books while the 20 subjects with low curiosity scores
borrowed an average of 23.9 books. The difference between the means
is statistically significant (t = 2.38 and P = .01 for a one tail test
with 37 degrees of freedom). The groups had been equated, by the
method of selection, for intelligence, average school achievement and
distribution of boys and girls. Our conclusion must be that the
curiosity test discriminates between children with different reading
habits.

When interpreting the results of this investigation, it is important
to note that the measure of amount of reading relates to
reading for pleasure. It does not include the use of reference books
for the study of school subjects. The use of the library as measured
by the number of books borrowed is purely voluntary. The lowest
number of books borrowed was zero and the highest in the selected group
was 71. It might also be relevant that reading for pleasure is a much
more popular diversion in Australia than it is in the United States,
so that the generality of the finding might be limited by cultural
differences. However, the generality of the finding is supported by
the fact that the hypothesis was formulated because of results with

American subjects and confirmed wilth the Australian sample. A cor-
relation between curiosity and reading for pleasure is to be expected
as an instance of exploratory behaviour in which a person both increases
and resolves conceptual conflicts by structuring his perceived
environment so as to give attention to novel stimuli.

This finding is particularly significant in that it gives evidence
of a behavioural correlate of the test. The results regarding orderliness
and arousal stimuli were valuable in showing that the fundamental
propositions of the theory are consistent with the correlates of the
test, but they were obtained from questionnairemeasures. The validity
of the test and theory is increased by the results not being restricted
to psychological test behaviour. Furthermore, according to the theory,
an important aspect of curiosity is the acquisition of certain investigatory
habits. These habits provide the instrumental acts which prolong
and remove conceptual conflict arising from the interaction
between an orderly cognitive map and novel stimuli. The theory is
therefore supported not only at the point of the conditions which give
rise to conceptual conflicts but in the argument regarding the manner
in which the curious person prefers to deal with such conflicts. The
curiosity behaviour of reading for pleasure is one of the instrumental
acts which play such an important part in the theory, and it illustrates
the point that such acts not only remove but also increase uncertainty,
while they are at the same time both a means to an end and
an end in themselves.

A further behavioural correlate of the curiosity test scores was
discovered accidently. In retrospect one can see that it could easily
have been predicted from some of the initial expectations which guided
development of the test after reading in the history of religion. One
of the earliest descriptions of curiosity as a psychologically significant
variable is given by St. Augustine of Hippo (74). It was noted
in Chapter 2 that as far as Christianity is concerned curiosity tended
to be most evident in association with heresy. If this is a general
tendency then, in so far as heresy is the result of the formation of
hypotheses independently of accepted doctrine, we might expect a correlation
of curiosity with aspects of religious behaviour not conditioned
by social factors. Due to the social nature of Christianity
it is highly probable that such behaviour would be unorthodox, but it
need not always be so. Given a social situation in which orthodoxy
is not the norm, independent inquiries could lead a person to discover
and accept the orthodox.

This finding is partly autobiographical. As a parish minister I
have often wondered why certain teenagers whose parents have no connection
with the Church suddenly appear regularly in the congregation.
I have searched for some common social factor and found none. It
never occurred to me that they might come initially out of curiosity.
I was surprised to find that there is strong evidence that they do.
While looking over the names of subjects in a sample from the local

high school which I had tested for curiosity a year earlier I noticed
the names of eight whom I had previously classified as independent
inquirers as far as Church affiliation is concerned. All of them had
curiosity scores in the top third of the distribution and two were
among the top five in a sample of 105. The odds are approximately
6,500 to 1 against all eight being in the same third of the distribution
if they were selected at random. We must conclude that
what the test measures is systematically related to their religious
behaviour.

There are two possible explanations which are not mutually
exclusive. One is a direct casual relation between curiosity and
independent seeking of knowledge and experience in general and especially
in those fields which include systematic ideation such as
religion. Both the experience of disorder and the ordering of experience,
which are theoretically hallmarks of curiosity, are evident
in the affiliation of a person to a religious system. The other
explanation is that due to my research interest in curiosity and
(hopefully) to my understanding of the processes involved I could
have presented the subject matter of Christianity in a manner which
appealed especially to the more curious. This could very well have
been the case even though I did not consciously use my knowledge of
curiosity in teaching or preaching. My contact with the subjects
concerned was not restricted to Church services for I taught them at
the school and in some groups at the Church. It would appear that
the strong correlation between curiosity and independent inquiries

in religion could be due to teaching techniques influenced by the
psychology of curiosity as well as to the nature of curiosity and
religion. Whatever the explanation may be, the correlation itself
is a prima face case for the validity of the test; reasonable explanations
are consistent with the theory, and there is a suggestion
that objective knowledge of curiosity gained by using the test can
be employed in education.

Although correlations with relevant variables guided the selection of
scoring criteria and other decisions in the test construction process
there was no single criterion of validity. The aim was construct
validity. The findings discussed in the previous chapters set the
test within a network of theoretically acceptable correlations. The
relevance of the correlates depends, of course, on the theory of
curiosity. The correlates we have found are consistent with, and
were largely predicted from, the theory presented in Chapter 4.This
theory was developed concurrently with the test and was influenced by
findings in the pilot studies. Construct validity would be increased
by consistency with an independent theoretical formulation to the
extent that such a theory is valid.

Berlyne's theory (10) was published at about the time when
definition of the scoring categories was completed. It is by far the
most comprehensive available. The principal difference with the
theory presented here derives from his attempt to produce a biologic-
ally fundamental theory beginning with simple processes studied in
the laboratory and elaborated by a modified drive reduction theory to
account for human epistemic behaviour, whereas we began with cognition.

-100-
and had no commitment to drive redaction theory. It will be argued
below that Berlyne's theory is defective in at least one important
respect, but it is generally descriptive of the curiosity process and
the test should be largely consistent with it.

If we examine the scoring categories in the light of Berlyne's
theory the test appears to measure what Berlyne is talking about.
While it is not necessary to make such an assumption, it will aid
exposition to assume that definitions of curiosity imagery in the
manual are at the same time descriptions of curiosity processes in the
common experience of the person who wrote the story. That is, if we
regard the imagery as a sample of a person's cognitive processes we
can apply Berlyne's theory directly to descriptions of the imagery by
which the manual defines curiosity.

The first of the major imagery criteria is a direct reference to
conceptual conflict. A person 'wondering about' some 'object, idea
or event which is the focus of present attention' is in a state of
arousal caused by the arousal potential of the object of attention
‘which is usually something strange or novel but does not have to be
so described’ (because the subjective state is direct evidence of its
arousal potential.) The other four major categories of imagery include
indirect evidence of conceptual conflict which can be inferred
from behaviour correlated with arousal and its reduction.

Orientation behaviour supplies basic evidence for Berlyne's concept
of arousal and it is completely included in the imagery category

of ‘Perceptual Instrumental Acts.’ Similarly, 'Cognitive Instrumental
Acts' correspond to the epistemic behaviour classes of epistemic
thinking and epistemic observation except that part of this latter might
be included with orientation behaviour in the previous category,
Berlyne's third class of epistemic behaviour, consultation, is included
in the 'Cue-Response Sequence' category except for reading which is
taken as cognitive instrumental behaviour in certain conditions. All
of the responses defined as perceptual or cognitive instrumental acts
are what Berlyne would regard as behaviour instrumental in reducing
conceptual conflict, while the appropriate stimulus conditions are
indicated by the requirement that there must be some object of
attention - 'something to be investigated or some problem to be
solved,' All such behaviour is part of the stimulus selection
process, typified by simple perceptual attention, and facilitated by
arousal. The stimuli selected (internally or externally) have the
general purpose of reducing arousal by resolving conceptual conflicts,
or more precisely they restore an optimal influx of arousal potential.
In retrospect it seems that the less obviously cognitive an act
is the more evidence is required before it can be classified as
curiosity behaviour. The two categories of imagery which include
references to overt behaviour require independent evidence of the
arousal potential of stimulus conditions before the instrumentality
of the response can be inferred. Exploratory behaviour is a very
broad category for Berlyne and only a very limited class of it is
included in the scoring category 'Exploratory Role Behaviour' in

which the exploratory nature of overt responses is confirmed by description
of an intrapsychic condition of exploratory interest or a
role name which indirectly designates such an interest. That is,
the arousal conditions and the instrumental behaviour are both included
in the scoring criteria. Similar confirmatory evidence is
required for the remaining category of imagery.

Perhaps because of the less specifically human nature of the
behaviour imagery the 'Cue-Response Sequence' category was the last
to be discovered, and it may be also for this reason that it most
clearly exemplifies Berlyne's understanding of the curiosity process.
This category might be better named 'S-O-R' because there are three
sets of conditions:- 1. A stimulus with high arousal potential
('interesting, unexpected or novel event'). 2. Covert events
('excitement or surprise, or need ... to know, or arousal of ...
interest') which might be regarded as either a response to the
arousing stimulus or as a stimulus for ensuing overt behaviour and
correspond to the drive conditions of arousal in Berlyne's theory.
3. Overt investigatory acts including 'any attempt to remove
uncertainty (caused by introduction of the cue) through the gathering
of more information.' This is the sequence which Berlyne believes
operates generally. If any two points in it are identified the
scoring manual defines the evidence as curiosity imagery, while the
three sets of conditions separately are defined by the minor imagery
criteria.

out only in the last category of imagery it can be readily inferred in
the others. Berlyne's attempt to find a psychological equivalence
between adult human intellectual curiosity and the simplest behavioural
processes of stimulus selection would appear to be supported by the
equivalence of these diverse categories of imagery in the test (although
the whole of the evidence is from cognitive processes and is only
symbolically representative of overt behaviour.) Support for Berlyne's
emphasis on intrapsychic arousal conditions seems implied by the dis-
covery made while constructing the test that, whereas the presence and
context of positive and negative affect is usually important in TAT
scoring, no evidence of curiosity was found by paying attention to
statements of affect except to a neutral or ambivalent kind of intra-
psychic excitement which could not be easily classified but indicated
an association between curiosity and some very general state of
arousal. Further support is given by the 'contrary indications' of
the scoring manual which refer to conditions Berlyne describes as
antithetical to epistemic behaviour. On the face of it there is a
very close correspondence between Berlyne's description of the
curiosity process and what the test appears to measure. Since the
scoring criteria were developed empirically and without regard to
Berlyne's theoretical formulation such close agreement must addsome
credence to both
.
If the test has face validity in terms of Berlyne's theory, do
the correlates indicate that it functions according to the Conflict.
Arousal and Curiosity schema? Two correlations obtained in the

pilot studies (Chapter2) indicate associations with conceptual conf-
lict and intrapsychic arousal, asking questions beginning with 'Why'
is fair evidence of conceptual conflict; and the 'arousal experiment’
using pictures of strange animals is consistent with Berlyne's conc-
eption of arousal potential. Neither is there any inconsistency when
we consider the later correlates; what is remarkable is that Berlyne
has offered very little which would have enabled us to predict the
most interesting correlations.

If 'sex role identity' and 'liking of novelty' are correctly
Interpreted as 'sensitivity to arousal' Berlyne's theory easily exp-
lains the correlation with curiosity given that such sensitivity does
vary between individuals. Although his postulates are derived almost
entirely from experiments involving manipulation of stimulus conditions
without controlled observation of independent variations in arousal,
Berlyne does note some evidence of individual differences. He
suggests that they may be due to the degree to which the cortex
exercises dominance over the reticular arousal system. He also says
'We can expect personality factors, cultural factors, learning, and
physiological states all to play their parts in determining the level
at which arousal tonus is maintained. Consequently, the rate of
arousal potential that is optimal can be presumed to vary widely from
individual to individual and from occasion to occasion'. The corr-
elation would be explained by certain people maintaining a relatively
high level of arousal tonus and thus requiring a high rate of arousal
potential which would be supplied by frequent exploratory behaviour

unless they were habitually within a very stressful situation.
When interpreting the correlation with 'orderliness' the crucial
question is the point at which concern with orderliness enters into
the curiosity process. In both Berlyne's theory of drive reduction
and the cognitive process theory, orderliness as a personality trait
could be expected to increase intellectual curiosity by providing
conditions more likely to create conceptual conflicts. However, a
difference in predictions follows from the theories if we make
certain assumptions when considering what happens after arousal has
occurred. In Berlyne's theory orderliness at this point should be
regarded primarily as a concern to reduce arousal, while in cognitive
process theory its principal function is to maintain the conflict.
If these assumptions are correct, drive reduction theory
would predict a negative correlation between curiosity and orderliness
at this point in the process because an escape response should
later become an avoidance strategy. So, on this understanding,
Berlyne's theory could only be made to explain the results by giving
attention to conditions creating conflict rather than to tension
reduction — scarcely a happy thought for drive reduction theorists.
Furthermore, it would be difficult to sustain this explanation in the
light of the finding of an independence or disjunction between orderliness
and sensitivity to arousal. However Berlyne's theory would
permit acceptance of the assumption that the function of orderliness
is to maintain conflict at least in some individuals or circumstances.
His theory could then explain the results by appeal to individual

differences in arousal tonus, which might require in some a mechanism
for frequently increasing arousal independently of 'sensitivity to
arousal.' Alternatively, recourse could be had to the possibility
that 'a temporary increase in arousal may be sought for the sake of
the drop in arousal that follows it. 'The orderliness correlation
pushes Berlyne's theory to its limits and it is hard to see how it
could have been predicted even though it can be explained by making
certain assumptions.

Berlyne lists reading as an example of consultative epistemic
Behaviour and making independent inquiries in religion could fall
into the same category. There are no theoretical problems with these
correlates. If they are regarded as examples of curiosity behaviour
they ought to be correlated with a test of curiosity. The findings
regarding delay of gratification ,n Achievement, test anxiety, and
sex differences, have greater theoretical relevance. The correlation
with the first was directly predicted by Berlyne in the passage quoted
above in Chapter *? N Achievement and Facilitating Test Anxiety are
easily identified sources of arousal which could increase epistemic
behaviour as an effective means of reducing arousal. Debilitating
Test Anxiety can be regarded as a source of arousal not sufficiently
subject to control for curiosity behaviour. The sex differences
raise other problems not relevant to the present task.

We can conclude that the test could have been constructed to
Measure curiosity as a personality trait which functions according to
the theory of D. E. Berlyne. The content and the correlates are cons-
istent with his theory. However we might question whether a drive

reduction theory is likely to stand up to much more investigation.
The point of greatest difficulty for the theory with our findings is
also the weakest point on other grounds.

In reference to drive reduction theory Berlyne states (page 164)
'It is true that this type of theory has often been coupled with a
picture of the animal organism as a being whose whole behaviour is
ultimately a collection of devices for keeping stimulation down to a
minimum (Freud 1915, Miller and Dollard 1941)* It is a picture that
has been severely battered by the growing body of knowledge about exploration.’
He then proceeds to argue that the shattering of this
picture associated with drive reduction theory need not mean abandonment
of the theory. 'The shattering of a picture may accompany a
comparatively minor change in the wording of a postulate. Nor is the
fact that these theories, as they now stand, fail to predict the facts
about stimulus selection a sufficient reason for rejecting them out of
hand.' Berlyne shows that much of the evidence can be accounted for
by drive reduction theory without modification if we make certain
reasonable assumptions about the equivalence of drive and arousal
states; and he suggests some 'minor' changes to account for the more
proactive characteristics of stimulus selection in exploratory
behaviour. The value of his theory depends upon the adequacy of
these modifications.

Stimulus seeking or epistemic behaviour which has the function
of supplying the organism with more information can be explained by

drive reduction theory in Berlyne's terms:- Additional information
means a reduction in uncertainty and consequently resolution of conceptual
conflict (which is a product of uncertainty and the absolute
strength of competing responses) and reduction of arousal or drive.
Such reactive behaviour often occurs in response to the introduction
of a novel stimulus. A stimulus having information value might be
considered novel in another sense, but what is meant here is one
having arousal potential due to the conflicting response tendencies
which arise from it. There is one case in which no use is made of
the function of conflict: boredom is regarded as an aversive state
having 'disorganizing effects like those of conflict.' In another
case arousal might increase prior to the introduction of conflict
inducing stimuli: anticipatory arousal, in response to warning
signals, prepares the organism to cope with an impending situation of
uncertainty. Thus much behaviour which is commonly explained as
being motivated by curiosity is accounted for in one way or another
as escape from an aversive condition of arousal. However this does
not imply that a rise in arousal may never be sought.

Berlyne lists 'three reasons why stimulation with relatively high
arousal value should be sought out:

1. The situation may move the influx of arousal potential back
towards its optimum when it has fallen below, thus reducing boredom.
This will apply principally to diversive exploration.

2. Continued exposure to the stimulation may reduce arousal. This
will apply principally to specific inspective exploration and to

'3. A temporary increase in arousal may be sought for the sake of the
drop in arousal that follows it. This will apply principally to other
cases of specific inquisitive exploration.' (10, page194).

None of these three reasons will bear examination.The first must
be examined in the light of his 'hypothesis that
the torments of boredom are associated with an upsurge of arousal.'
Either, this assumed 'endogenous rise in arousal' is a noxious condition
while the incoming stimuli have low arousal potential and becomes a
desirable state when influx of arousal potential is raised or, the in-
coming stimuli provide escape from boredom by reducing arousal. What
this amounts to is that stimuli with high arousal potential might
either increase or decrease arousal. Yet it seems that Berlyne wants
to have it both ways for his second reason for seeking 'stimulation
with relatively high arousal value' is that 'continued exposure to the
stimulation may reduce arousal. 'How can this be so?

Berlyne discusses at length a variety of stimulus properties
(e.g. intensity, novelty, surprisingness, complexity) and cites
evidence that they have the effect of increasing arousal. He concludes
the discussion with a definition: 'We shall henceforth refer
to all these properties of incoming stimuli with power to affect
arousal as arousal potential.'(page179). There can be no doubt
that 'to affect' here means' to produce'or 'to bring about'; and
all the evidence is of the power of stimuli with these properties to
increase arousal. But later(page193) he introduces a different
idea: 'The properties of stimuli that we have grouped together as

arousal potential are ones whose intensification seems, on the whole
to entail a rise in arousal. But they seem also to overlap with
properties that enable the cortex to moderate arousal. The relation
between arousal and arousal potential can, therefore hardly be a
straightforward one.' Indeed it is hardly straightforward, but what
explanatory power can such a concept have? Does it explain anything
to say that stimulus properties which have the power to increase
arousal are sometimes sought in order to decrease arousal? It might
be true but how are we to decide which stimuli have arousal potential
when arousal might either increase or decrease in response to them,
and how are we to know whether they are sought for their increasing or
decreasing effects? A theory of this nature can explain anything.
If the second of Berlyne's 'three reasons' is not a logical contradiction
of other propositions in the theory it is an insurance policy
against possible empirical contradiction. The premium on such insurance
policies is always the same — irrelevancy.

The crucial point in relation to Berlyne's third 'reason' is
whether it is possible to separate the onset of arousal from its offset
in reference to its reward value. If the two can never be
separated then it is not possible to know whether increased arousal
is sought for its own sake. However he refers to an experiment by
W. W. Roberts (73) and remarks on the findings 'that the reward value
of the onset seems to be independent of that of the offset.' Here
increased arousal was sought for its own sake even though reduction
was also a positive goal. Such an event cannot be explained by

reference to optimal influx of arousal potential because if the
stimulation is sought in order to increase influx of arousal potential
there can be no reason for avoidance of the same stimulus under the
same conditions. There are no grounds for assuming that the re-
inforcing effects of the arousal jag are due only to its closing
stages and Roberts' experiment is evidence to the contrary. It may
still be argued that there are conditions under which arousal is
sought only for the sake of its reduction but it is difficult to
imagine how this could be demonstrated experimentally except along the
lines of Roberts' experiment. Since Berlyne does not modify the temporary
increase hypothesis in the light of Roberts' findings we must
assume that it is another insurance policy. Alternatively, his
citing of the experiment with the comment 'We ought, however, to
mention another type of neural mechanism that might underline the
arousal jag' might indicate a possible change in the theory. Such a
change, providing for the seeking of stimuli independently of subsequent
reduction in arousal and without appeal to optimal arousal
potential, would make it no longer a drive reduction theory.

Of course, the concept of an 'optimal influx of arousal potential'
is in itself a significant modification of simple drive reduction
theory. However it is a hypothetical construct of little value when
it cannot be operationally defined due to the ambiguity of the
definition of arousal potential in relation to the level of arousal.
If the relevant stimulus properties could be defined in some way independently
of arousal, then the relation between stimulus seeking

and arousal could be usefully examined whether or not arousal is conceived
as a drive state. Berlyne has made some progress in this
direction, but his ultimate concern to account for all behaviour in
terms of the arousal concept within a drive reduction theory as little
modified as possible has led him into an impossible position.
In summary, Berlyne's theory is useful except at the point where
drive reduction theory comes under greatest strain due to the evidence
from studies of exploration. The modifications he has made to drive
reduction theory are logically and empirically inadequate.

McReynolds (56) has made a useful distinction between 'Novelty-adjustive
behaviour' and 'Novelty-seeking behaviour'. It would appear
that Berlyne's theory can adequately account for the former but not
the latter. In order to account for both McReynolds suggests a
theoretical model very similar in outline to the cognitive process
theory which guided the later stages of this investigation. 'It is
assumed that in the course of their commerce with the environment,
higher animals develop a cognitive structure which represents for them
the nature of the world....When new experiences occur a certain amount
of reorganization would be required.... It is postulated that animals
have an inherent tendency to assimilate new input, i.e. to carry
through such cognitive restructuring as necessary to "fit" new percepts
congruently into the over-all cognitive structure.'

The 'inherent tendency' McReynolds postulates could be regarded
as a drive to learn. He does not spell out how the process takes

place but it would appear to imply the intrinsically rewarding nature
of what would otherwise be regarded as instrumental acts. This is
what we had supposed.

McReynolds accounts for individual differences by his concept of
'perceptualization rate (P.R.)':- 'A high P.R. would imply considerable
elaboration of cognitive organization per time unit....it is postulated
that an animal has an optimum P.R., a rate which it prefers.1 He
notes that this proposal is similar to Berlyne's 'optimal influx of
arousal potential' and Berlyne cites his earlier work. We have noted
the ambiguity in Berlyne's definition but the same criticism does not
apply to McReynolds'formulation. Berlyne's definition is in terms of
stimulus properties which have certain effects. McReynolds’ is a direct
reference to a preferred cognitive style. However we must take issue
with McReynolds on another point. If a high P.R. preference is to
explain a high degree of curiosity then it follows that the highly curious
person seeks a rapid and extensive accommodation of his cognitive
map. The correlation with 'orderliness' and the theory from which
it was predicted place the emphasis not on the extent or speed of
accommodation but on a minimum of change in a category system in the
process of coding a signal even though the highly curious person seeks
also to place his cognitive map in jeopardy. It is this interaction
between openness and discrimination which gives curiosity its
tantalizingly ambivalent character and which McReynolds has overlooked.

is that it provides a language framework within which to account for
intrapsychic variability as well as external stimulus effects. The
two facets of orderliness and novelty are not brought together in
theories developed from experimental studies. Freud, paying attention
to intrapsychic conditions, pointed the way forward while, at the same
time, he obscured the real issues by appealing to somatic drive states.
If the process is analysed in terms of the coding operation, attention
is focused upon the interaction between category system and signal in
such a way that the intrapsychic determinants are seen to be significant.

Since the theory deals with a process, and does not relate only to
a variable personality trait, it has dynamic aspects. This frees us
from any need to postulate a drive state. The function of conceptual
conflict in the coding operation could be said to provide the motivational
conditions, but when we look at it in detail the distinction
between motives and habits seems to disappear. The freedom gained in
this way pays dividends of both simplicity and flexibility. A sufficiently
detailed exposition of the theory provides a constructive
alternative to drive reduction models without limiting our ability to
account for behaviour which seems to support drive reduction theory.

That the theory has predictive power is demonstrated by the
findings reported in Chapter 5. Some additional evidence is discussed
in Appendix 6. One source of predictive power in a cognitive theory,
which is properly so called, is in the coincidence between its terms and
personality or behavioural charactaristics which can be directly described
and measured in human verbal behaviour. In some respects it

bears a resemblance to the old associationist theories and to Gestalt
psychology, but it has the potential advantage of operational
specificity.

The construct validity of the curiosity test is sufficiently well
demonstrated for future research on curiosity to include direct measurement
of what would otherwise be an assumed intervening variable.
Theoretically there is a great need for detailed attention to mediating
processes. There is ample experimental evidence that curiosity behaviour
is largely situationally determined, and our finding that the
test is sensitive to changes in administration conditions suggests
similar limitations on the idea that the test measures a personality
trait. Tet there is also evidence of generality or consistency in the
test-retest reliability and in the variety of theoretically meaningful
correlates. It is simply not possible to build a test of this kind
unless a person functions somewhat the same way in one situation as in
another, but this does not mean the test scores reflect only the more
general characteristics. The suggestion is that the test could be
used in experiments to study the changes in mediating processes which
are contingent upon external circumstances. If it is used to measure
individual differences and changes in curiosity it would be a valuable
addition to laboratory equipment. Most experiments on curiosity
depend upon manipulation of stimulus conditions to control curiosity
and have no check on whether the intervening variable has in fact been

held constant or systematically varied.
The findings relating curiosity to reading habits, religious
activity and classroom behaviour are very valuable in showing that
something is being measured which has effects in 'real life'. The
correlates are not restricted to responses on other psychological
tests. However much more evidence is needed on the behavioural correlates
and this could be a fruitful field for future investigation.
The relationship between curiosity and creativity is generally
assumed to be a positive one (e.g. 50). Two recent reports of correlations
between curiosity cognitive style elements and creativity
(39, 57) suggest some interesting lines of research. Garwood found
creativity to be positively correlated with 'composite personality
originality', 'cognitive flexibility' and 'dominance'. Mendelsohn
and Griswold interpreted their results as 'reflecting wider deployment
of attention and less screening out of "irrelevant" past experiences
by high creatives during problem solving.' There was no
concern with curiosity as such in these studies, but the results can
be interpreted by the cognitive process theory as relating components
of the curiosity strategy to creativity. However there are probably
some aspects of curiosity which are incompatible with creativity –

even though Samual Johnson observed that curiosity ‘is in great and
generous minds....both the first passion and the last.’ The point of
probable difficulty is in the highly curious person's tendency to
maintain conceptual conflict. To be creative he must produce a new
order. This growth in organization is of prime importance to the
highly curious, but the curious person will not be easily satisfied.
He will tend to seek only that order which is for him the best integration
of the disparate elements he is open enough to consider. This
could limit his production. It might also contribute to a higher
quality product. The problem for educational research is to discover
means of encouraging divergent thinking without losing the ability to
bring order out of chaos.
The significance of curiosity for motivation theory in general
demands more attention than it has been given. There is little
profit, at this stage, in trying to use the results of research in
curiosity to support one theory against another. Curiosity raises the
question of whether there is value in continuing to use 'motive' as a
technical term in psychology. This question must be answered before
we continue the various disputes of the past. The mechanism underlying
curiosity behaviour appears to be so primitive, and so capable
of sophisticated development, that it could point the way to a general
theory of learning and action which has no place for a class of
variables such as drives or motives. The greatest need is for hard
conceptual work, which might include gathering more information, but
which is primarily concerned with placing one's cognitive map in jeopardy
to the extent that is necessary to produce a new order out of the
old - and that is curiosity.

1. Adlerstein, A., & Fehrer, Elizabeth . The effect of food
deprivation on exploratory behaviour i n a complex maze.
J . сomр. physiol. Psychol. 1955, 48, 250-253.
2. Alpert, R., & Haber, R. N. Anxiety i n academic achievement
situation. J. abnorm. soc. Psychol.. 1960, 61, 207-215.
3. Aronoff, J. Freud's conception of the origin of curiosity.
J . Psvchol. 1962, 54, 39-45.
4. Atkinson, J. W. Motives in fantasy, action and society. Princeton:
Van Nostrand, 1958.
5. Barnett, S. A. Experiments on "neophobia" in wild and laboratory
rats. Brit. J. Psychol.. 1958, 49, 195-201.
6. Barnett, S. A. Exploratory behaviour. Brit. J. Psychol. 1958,
49, 289-310.
7 . Berlyne, D. E. An experimental study of human curiosity . Brit.
J . Psychol.. 1954, 45, 256-265.
8. Berlyne, D. E. The arousal and satiation of perceptual curiosity
in the rat . J. соmр. physiol. Psychol.. 1955, 48, 238-246.
9. Berlyne, D. E. Conflict and information theory variables as
determinants of human perceptual curiosity . J. exp. Psychol..
1957, 53, 399-404.
10. Berlyne, D. E. Conflict, arousal and curiosity. New York:
McGraw-Hill, 1960.
11. Berlyne, D. E. Note on food deprivation and extrinic exploratory
responses. Psychol. Rep. 1962, 11, 162.

12. Berlyne, D. E . , & Slater J. Perceptual curiosity , exploratory
behaviour and maze learning. J. соmр. physiol. Psychol. 1957,
50, 228-232.
13. Bolles, R. C., & de Lorge, J. Exploration in a Dashiell maze as
a function of prior deprivation , current deprivation, and sex.
Canad. J. Psychol. 1962, 16(3), 221-227.
14. Bruner, J . S. On perceptual readiness. Psychol. Rev. 1957,
64, 123-152.
15. Bu t l e r , R. A. Curiosity i n monkeys. Scient. Amer. 1954,
190, 70-75.
16. Butler, R. A. The effect of deprivation of visual incentives on
visual exploration motivation in monkeys. J. соmр. physiol. Psychol.,
1957, 50, 177-179.
17. Butler, R. A. Exploratory and r e l a t e d behaviour: a new trend in
animal research. J. indiv. Psychol. 1958, 14, 111-120.
18. B u t l e r , R. A., & Alexander, H. M. Daily patterns of visual exp-
loratory behaviour in the monkey. J. соrр. physiol . Psychol.
1955, 48, 247-249.
19. Butler, R. A., & Harlow, H. F. Persistence of visual exploration
in monkeys. J. сотр. physiol . Psychol.. 1954, 47, 258-263.
20. Caron, A. J . Curiosity, achievement and avoidant motivation as
determinants of espistemic behaviour. J. abnorm. soc. Psychol..
1963, 67, 535-549.
21. Carr, R. M., & Williams, C.D. Exploratory behaviour of three
Strains of rats. J. соmр. physiol. Psychol.. 1957, 50, 621-623.

22. Cattell , R. В. Personality and motivation structure
measurement. London: Harrap, 1958.
23. Charlesworth, W. R . , & Thompson, W. R. Effect of lack of visual
stimulus variation on exploratory behaviour in the adult white
rat . Psychol. Rep.. 1957, 3, 509-512.
2 4 . Clark, A. W. A reinvestigation of the habituation of exploratory
behaviour. Aust. J. Psychol.. 1958, 10, 151-162.
25. Church, R. M. The varied effects of punishment on behaviour.
Psychol. Rev.. 1963, 70, 369-402.
26. Couch, A., & Keniston, K. Yeasayers and naysayers: Agreeing
response set as a personality variable. J. abnorm. soc.
Psychol.. 1960, 60, 151-174.
27. Danziger, K., & Mainland, Margaret. The habituation of exploratory
behaviour. Aust. J. Psychol., 1954, 6, 39-51.
28. Dashiell, J . F. Aquantitative demonstration of animal drive.
J. соmр. physiol. Psychol., 1925, 5, 205-208.
29. Dember, W. N., & E a r l , R. W. Analysis of exploratory manipulatory
and curiosity behaviour. Psychol. Rev., 1957, 64, 91-96.
30. Dember, W. N., & Fowler, H. Spontaneous alternation behaviour,
Psychol. Bull. . 1958, 55, 412-428.
31. Fehrer, Elizabeth . The effects of hunger and familiarity of locale
on exploration . J. соmр. physiol. Psychol.. 1956, 49, 549-552.
32. Festinger, L. Theory of cognitive dissonance. Evanston, I l ls
Row, Peterson, 1957.

33. Fiske, D. W . & Maddi, S. R . (Ed.) Functions of varied experience.
Horaewood, 111: Dorsey, 1961.
34. Frankl, V. E . From death-camp to existentialism. Boston:
Beacon, 1959.
35. Freud, S. A general introduction to psychoanalysis. New York:
Boni & Liveright, 1920.
36. Freud, S. Three essa.vs on the theory of sexuality. London:
Imago, 194-9.
37. Freud, S. An outline of psycho-analysis. London: Hogarth, 1949.
38. Frick, F. C. Information theory, in Koch, S. (Ed.) Psychology:
a study of a science. Study 1, Vol. 2 . New York, McGraw-Hill,
1959.
39. Garwood, D. S., Personality factors related to creativity in young
scientists. J. abnorm. soc. Psychol.. 1964, 68, 413-419
40. Glanzer, M . The role of stimulus satiation in spontaneous
alternation. J. exp. Psychol.. 1953, 45, 387-393.
41. Glanzer, M . Stimulus satiation: an explanation of spontaneous
alteration and related phenomena. Psvchol. Rev., 1953, 60,257-268.
42. Glanzer, M . Curiosity, exploratory drive, and stimulus satiation.
Psychol. Bull.. 1958, 55, 302-315
43. Glickman, S. E . Effects of peripherial blindness on exploratory
behaviour in the hooded rat. Canad. J. Psychol.. 1958, 12, 45-51.
44. Hardy, K. R . An appetitional theory of sexual motivation.
Psychol. Rev.. 1964, 71, 1-18.
45. Harlow, H. F., Harlow, M . K., & Meyer, D. R . Learning motivated by
manipulation drive. J. exp. Psychol.. 1950, 40, 228-234.

46. Heathers, G. L . The avoidance of repetition of a maze reaction
i n the rat as a function of the time interval between trials .
J. Psvchol.. 1940, 10, 359-380.
47. Hebb, D. O. The organization of behaviour. New York: Wiley, 1949.
48. Jencks, В., & Porter, P. B. Need reduction and reinforcement:
Incentive values of various goal objects and novelty.
J. Psychol. 1963, 56(1), 79-83.
49. Lashley, K. S. Visual discrimination of size i n the albino rat.
J . an. Behav.. 1912, 2, 310-331.
50. Leuba, Clarence, A new look at curiosity and creativity.
J . higher Educ.. 1958, 29, 132-140.
51. Levi-Strauss, C. Tristes Tropiques. Encounter. 1961, 16, No.2.
52. Levine, S., Staats, S. R., Fromme, G. Studies on response by the
Rat to environmental change. Psychol. Rep., 1958, 4, 139-144.
53. Logan, F. A. The Hull-Spence approach, in Koch, S. (Ed.)
Psychology: A study of a Science. Study 1, Vol , 2 . , New York:
McGraw-Hill, 1959.
54. Maw, W. H., & Maw, E t h e l W. Selection of unbalanced and unusual
designs by children high in curiosity . Child Development. 1962,
33(4), 917-922.
55. McClelland, D. C., Atkinson, J. W., Clark, R. A., Lowell, E. L.
The Achievement Motive. New York: Appleton-Century-Crofts, 1953.
56. McReynolds, P. Exploratory behaviour: Atheoretical interpretation.
Psychol. Rep.. 1962, 11(2), 311-318

57. Mendelsohn, G. A., & Griswold, B.B. Differential use of
Incidental stimuli in problem solving as a function of
creativity . J. abnorm. soc. Psvchol.. 1964, 68, 431-436.
58. Meyers, W. J . Critical period for the facilitation of exp-
loratory behaviour by infantile experience. J. сотр.
physiol , Psychol.., 1962, 55, 1099-1101.
59. Mischel, W. Preference for delayed reinforcement and social
responsibility . J. abnorm. soc. Psvchol., 1961, 62, 1-70.
60. Mischel, W., & Metzner, R. Preference for delayed reward as
a function of age, intelligence , and length of delay interval.
J . abnorm. soc. Psychol.. 1962, 64, 425-431.
61. Montgomery, K. C. Spontaneous alternation as a function of time
between trials and amount of work. J. exp. Psvchol., 1951,
42, 82-93.
62. Montgomery, K. C. A test of two explanations of spontaneous
alternation . J. соmр. physiol . Psvchol.. 1952 , 45 , 287-293.
63. Montgomery, K. C. Exploratory behaviour as a function of
"similarity" of stimulus situation . J. соmр. p h y s i o l.
Psvchol.. 1953, 46, 129-133.
64. Montgomery, К. C. The effect of activity deprivation upon
exploratory behaviour. J. соmр. physiol . Psychol.. 1953, 46,
438-441.
65. Montgomery, К. C. The role of exploratory drive in learning.
J. соmр. physiol . Psychol., 1954, 47, 60-64.
66. Montgomery, К. C., & S e g a l l , M. Discrimination learning based upon
the exploratory drive . J.соmр. physiol. Psychol.. 1955. 48, 225-
228.

67. Montgomery, К. С. The relation between fear induced by novel
stimulation and exploratory drive. J. соmр. physiol . Psychol..
1955, 48, 254-260,
68. Montgomery, К. C., & Zimbardo, P. G. Effect of sensory and
behavioural deprivation upon exploratory behaviour in the rat.
Percept, mot. Skills . 1957, 7, 223-229.
69. Nissen, H. W. A study of exploratory behaviour in the white rat
by means of the obstruction method. J. genet. Psychol..
1930, 37, 361-376.
70. Piaget, J. The Psychology of intelligence. London: Routledge
& Kegan Paul, 1950.
71. Rahman, M. M. "Escape valence" i n the study of exploratory
behaviour of albino mice i n the laboratory. J. psvchol. Res..
1963, 7, 79-85.
72. Richards, W. J . , & L e s l i e , G. R. Food and water deprivation as
Influences on exploration . J. соmр. physiol. Psychol.. 1962,
55, 834-837.
73. Roberts, W. W. Both rewarding and punishing effects from
Stimulation of posterior hypothalamus with same electrode at same
intensity . J. соmр. physiol . Psvchol.. 1958, 51, 575-579.
74. Ryan, J. K. The confession of St. Augustine. New York:
Doubleday (Image), 1960.
75. Silverman, J. The problem of attention in research and theory in
schizophrenia. Psvchol. Rev.. 1964, 71, 352-379.
76. Small, W. S. Notes on the psychic development of the young white
rat . Amer. J. Psychol., 1899, 11, 80-100.

77. Thompson, W. R. & Kahn, A. Retroaction effects in the explor -
atory activity of "bright" and "dull" rats . Cand. J . Psvchol..
1955, 9, 173-182.78. Tolman, E. C. Purpose and recognition : the determiners of
animal learning. Psvchol. Rev.. 1925, 32, 285-297.
79« Tolman, E. C. Principles of purposive behaviour, in Koch,S. (Ed.)
Psychology: A study of a science Study 1, V o l . 2. New York:
McGraw-Hill, 1959.
80. Walker, E. L., Dember, W. N., Earl, R. W., & Karoly, A. J.
Choice alternation : I. Stimulus vs. place vs. response.
J . como. physiol . Psvchol.. 1955, 48, 19-23.
81. Walker, E. L . , Dember, W. N., Earl, R. W., Fawl, C. L . , & Karoly, A.J.
Choice alternation : I I I Response intensity vs. response dis-
criminability . J. соmр. physiol . Psvchol.. 1955, 48, 80-85.
82. Wehling, H. E . , & Prokasy, W. F. Role of food deprivation in the
acquisition of the observing response. Psychol. Rep. 1962,
10, 399-407.
83. Welker, W. I . "Free" versus "forced" exploration of a novel
situation by rats. Psvchol. Rep.. 1957, 3, 95-108.
84. White, R. W. Motivation reconsidered: the concept of competence.
Psychol. Rev.. 1959, 66, 297-333.
85. W i t k i n , H. A., Lewis, Helen В., Hertzman, M., Wachover, Karen,
Meissner, Pearl В., and Wapner, S. Personality through
perception. New York: Harper, 1954.

86. Woodworth, R. S. Dynamics of behaviour. New York: Holt, 1958.
87. Zimbardo, P. G., & Montgomery, K. C. The relative strengths
of consumatory responses in hunger, thirst , and exploratory
drive . J. comp. physiol . Psvchol.. 1957, 50, 504-508.
88. Zimbardo, P. G., & Montgomery, K. C. Effects of " free -
environment" rearing upon exploratory behaviour. Psychol. Rep..
1957, 3, 589-594

1. Italics indicate Berlyne's technical terms.

2. For example, Berlyne (7) and Caron (20). Maw and Maw (54) used
ratings by teachers, peers and the subjects themselves. Our results
(see Appendix 6) suggest that teachers might be able to discriminate
between extreme groups, but whether such ratings are valid is an
empirical question which can be answered by using the test.

1. For a thorough examination of this problem in motivation
theory see White (84).

2 . 'Coding' here is analogous to decoding in information theory.

3. Efficiency of a category system should not be confused with efficiency
of adaptation to the environment generally or with efficiency of
learning strategy. An efficient strategy of learning or adaptation to
the environment has a flexible quality by which an inefficient category
system is changed to an efficient one for the purpose of coding signals
not previously received. In this section we are using the term
•efficiency' in the sense in which it is used in information theory and
without the connotations it has in personality theory where flexibility
is regarded as a trait conducive to efficient adjustment. For the relevance
of information theory see Frick (33) and Berlyne (9, 10).

4. Whether we use motive terms, such as 'want', or habit terms is
largely irrelevant.

5. See evaluation in Chapter 6.

6. See evaluation of Berlyne's theory on this question in Chapter 6.

7. I am indebted to J.M.Carlsmith for an unpublished review. For a
general review of the varied effects of punishment, see Church (25).

1. This study was carried out in co-operation with Ralph Metzner.

www.beswick.info>psychres>Beswick Curiosity 1964