toys in the attic: A
Psychological Model 279 Part 3: A Psychological Model – Edward C Tolman 3.1 Introduction Part
III will attempt to place those concepts especially pertinent to
psychology, which have been developed in Parts I
and II, within a more strictly psychological
framework. In other words, a theoretical system of
psychology will be presented which contains, it is believed, all the
descriptive and theoretical constructs necessary for explaining and predicting
the action or behavior 1 of individual persons as this concept has
been developed above. And again it will
be emphasized that such action or behavior of individual human beings
practically always takes place in,
and is relative to, an
environment which typically contains not only "mere" physical objects
but also other persons, collectivities, culturally presented values, and
accumulated cultural resources. The general theoretical
organization to be used in the present analysis is one which today has received
relatively wide acceptance among psychologists. This organization distributes all descriptive and explanatory
constructs into: (1) independent
variables; (2) the
dependent variable of behavior or action; (3) postulated
intervening variables; and (4) postulated
causal connections between the three types of variables. Let us
consider each of these four headings. 1 The terms
"action" and "hehavior" will he usea interchangeaMy in this
section. I have previously used the
term "hehavior" (Tolman [22]) with practically all the connotations
which have been given here to the term "action." That is, a behavior, or an action, differs
from a mere movement or response in that a hehavior or action can he folly
identified and described in terms only of the organism-environment
rearrangement which it produces. A mere
movement or a response can, on the other hand, he identified in purely
intra-organism terms, e.g., as consisting of such-and-such muscle contractions
or glandular secretions. 280
A Psychological Model THE INDEPENDENT VARIABLES The independent variables are
the initiating causes of the individual’s action. The main types of independent variable which have been found by
psychologists to date are three: (1) Stimulus situation (i.e., the environmental
entities presented to the individual actor at the given moment). These environmental entities consist of
physical, social, and cultural objects and processes. Such environmental objects and processes can, of course, in the
last analysis affect the action of an individual only through the mediation of
sense perception and memory-trace arousal.
But the technical understanding of such neurophysiological processes
will not, as such, be our concern.
Rather, we shall assume that, for the level of causal determination with
which social science is concerned, one can assume, within reasonable limits
(without, that is, detailed psychophysiological, "mediational"
investigation), what potential "perceptions" 2 are
possible and probable as a result of the presence to the individual actor of a
certain stimulus situation (i.e., certain physical, social, and cultural
objects).3 Granting this
assumption to be valid, one general problem of psychology will be that of
discovering why, although certain specific types of social, cultural, and
physical objects are presented environmentally, the given actor on a given
occasion will react only to certain ones of these objects or will behave to
these latter only with certain degrees of distortion as compared with the
"normal," "standard," or "usual" ways of
behaving. (2) States Of drive arousal and/or drive satiation. The physiological entities of drive arousal
and drive satiation can today roughly be identified by a physiologist from
observations of the recency and/or intensity of just past "consummatory behaviors." 4 (By "consummatory behaviors" are to be understood such types of "terminal behavior"
as eating, drinking, copulating, sleeping, dominating, submitting, affiliating,
aggressing, avoiding, etc., which terminate spells
of activity.) But the real
definition of drives, as we are conceiving the term here, lies not in the
resultant readinesses or non-readinesses for such types of terminal
behavior but in precise statements concerning the
states of the underlying organs and tissues themselves. It is these organ and tissue states which
ultimately define "drives" as they are discussed here as one of the three
main types of independent variables.
The resultant readinesses or non-readinesses for types of consummatory
behavior will constitute, rather, what we shall define below as the "needs."
And such needs, as will also be seen below, are to be considered as a
type of intervening variable. (3) Such individual-difference-producing variables as
heredity, age, sex, and special
physiological conditions such as
those produced by drugs, endocrine disturbances, and the like. These variables are assumed to act directly
in determining those types of intervening variables which are called traits - either capacity traits or
temperamental traits. These trait
variables are assumed to interact with the other or "content" variables so as to enhance or
depress the magnitudes of the latter.
But any final and definitive assumptions as to the nature of the
interacting functions between "traits" and "contents" seem
to be a problem for the future. And in
the present essay the whole problem of traits will be largely ignored or at the
most only formally acknowledged. 2 The term
"perception" will he used throughout in a broad sense to cover
immediate memories and inferences as well as perceptions in the narrower
sense. 3 This fact of a great
consistency between the given environmentally presented entities and the
probable perceptions that will he "achieved" (irrespective of
tremendous variability in the mediational neurophysiological mechanics from
occasion to occasion) is, of course, the problem of "thing
constancy." The most fruitful and
searching analyses of this problem have, I believe, been made to date by
Brunswik [3]. 4 Such observations have
sometimes been called "maintenance-schedule" observations. Introduction
281 THE DEPENDENT VARIABLE OF BEHAVIOR (ACTION) This dependent variable is
conceived as consisting of responses which, from the point of view of a purely
physiological analysis, are merely combinations of verbal, skeletal, and
visceral reactions; but which from the point of view of the present action
schema are identified and defined not in terms of their underlying physiology
but in terms of their "action meanings." In other words, a given behavior or action is to be identified
and defined, in the last analysis, only in terms of
the ways in which it tends to manipulate or rearrange physical,
social, or cultural objects relative to the given actor. An actor "goes toward the light,"
"consumes food," "aggresses against a friend," "avoids
the shade," "puts on a coat," and so on. In other words, it
appears that the mediational problems of muscle contraction and of gland
secretion, like those of sense-perception, can for our purposes be largely ignored
because the action meanings tend to remain the same through a wide diversity of
alternative, physiological movement-details.5 INTERVENING VARIABLES Intervening
variables are postulated explanatory entities conceived to be connected by one
set of causal functions to the independent variables, on the one side, and by
another set of functions to the dependent variable of behavior, on the
other. A certain misunderstanding as to
what may thus be the
ultimate definition of such intervening variables must, however, be corrected. In a recent article
MacCorquodale and Meehl [15] have suggested that different schools of
psychology have defined intervening variables in two different ways. One school employs what these authors
have decided to call the initial, or pure, concept of intervening
variables. The pure concept assumes,
they say, that the whole character and meaning of an intervening variable is
given in and exhausted by its assumed functional (mathematical) relationships
to the causal independent variables on the one side, or to the caused dependent
behavior on the other. The second
type of theory assumes, they say, more full-blooded types of intervening
variable, which they designate as "hypothetical
constructs." This second
type of theory ascribes, they believe, ostensively definable, substantive
properties to intervening variables, which properties can, hypothetically at
least, eventually be given direct operational measurement. "Hypothetical constructs" thus are
defined by substantive properties which are separate from and more than the
mere functional relationships of such constructs to the independent variables
or to the dependent
behavior. 5 Again we are
indebted to Brunswik [3] for having emphasized these facts. Guthrie and Horton [6] have suggested the
use of the two terms act and action for the response defined in terms of the
environmental-actor rearrangements which it tends to produce and the term movement for the detailed physiological
character of the response. 282
A Psychological Model I agree with MacCorquodale and
Meehl that this distinction between the functional or mathematical
identification of intervening variables and a denotative, ostensive
identification of intervening variables is an important one. Further, I would also agree that this is a
distinction between two different approaches to the intervening variable which
has been slurred over in previous discussions.
MacCorquodale and Meehl have done us a service in bringing these two
approaches to light. On the other hand,
I do not agree with these authors that, corresponding to these two ways of
defining, there are really to be found two disparate classes of theory. I shall contend rather that actually all
current psychological schools - whatever their explicit (or lack of
explicit) statements - define intervening variables both by the assumed functional relations of
such intervening variables to the independent and/or to dependent variables and by the postulated, ostensive properties
also attributed to such intervening variables.
The differences between different theories seem to me to lie not in
whether or not they ascribe constitutive properties to their intervening
variables, but rather in the nature of such ascribed properties. In fact, three major
trends in current theories relative to such ascribed properties are to be
found. These may be called (1) the neurophysiologic~trend, (2) the
phenomenological trend, and (3) the trend toward a sui generis
model. Let me briefly summarize each of
these. (1) The theories which adopt the neurophysiological trend ascribe primarily
(either implicitly or explicitly) neurophysiological constitutive properties to
their interv£ning variables. Most of
the stimulus-response theories, whether of Hull [8] and his students or of Guthrie
[5] and his, in inventing their intervening variables, rely, I believe, on more
or less explicit assumptions to the effect that these intervening variables are
in the nature of afferent, efferent, or associative neural connections.6 6 Kohler [121, Krech
[9 and 101 and Hebb [71 also belong to the neurophysiological trend, but
their neurophysiological constructs involve more in the way of brain fields and less in the
way of insulated nerve channels. Hull [8] tends to deny that he
is using nenrophysiological constructs in thinking of his
intervening variables. Nevertheless, it seems to me that be is relying on them,
at least
implicitly if not explicitly, for his notions about how these intervening
variables act. Introduction
283 (2) Theories which
adopt the phenomenological trend
ascribe primarily introspectively derived, experiential characteristics to
their intervening variables. Much of
Freudian theory would seem to belong in this category. Gestalpsychology also seems to belong here
(as well as in the first category) since Gestalt psychologists postulate not
only brain-field events but also correlated phenomenological events. In fact, Gestalt theory seems to have begun
with phenomenological constructs and then to have moved back to correlated
physiological constructs as its more basic explanatory device. (3) Theories which exhibit the trend
toward a sui generis model invent
a set of explanatory structures and processes (hypothetical constructs) which
draw on analogies from whatever other disciplines - mathematics, physics,
mechanics, physiology, etc. - as may be deemed useful. Freud's water-reservoir concept of the
"libido," Lewin's "topological and vector" psychology, and
the theory to be presented in the following pages belong primarily in this
third category.7 The theory to be presented here
will then be quite frankly one which develops (by various analogies drawn from
simple physics and mechanics, from Lewin's "topological and vector"
psychology, and from common experience) a sui generis model. This model has its own (tentatively)
ascribed intervening constitutive structures and processes and its own variety
of interconnecting causal functions. The
justification for such a model is, of course, wholly pragmatic. Such a model can be defended only insofar as
it proves helpful in explaining and making understandable already observed
behavior and insofar as it also suggests new behaviors to be looked for. And any such model must, of course, be ready
to undergo variations and modifications to make it correspond better with new
empirical findings. Finally,
insofar as such a model holds up and continues to have pragmatic value, it must
be assumed that
eventually more and more precise and intelligible correlations will be
discovered between it and underlying neurophysiological structures and
processes - especially as more about the latter comes to be known and verified
by physiologists.8 7 A fourth type of
theory, of which that of Brunswik [31 and that of Skinner [21] are the most
distinguished examples, seeks to do without much in the way of intervening
rariables (Skinner does postulate a "reflex reserve") and to attempt
to develop at once direct empirically establishable functions between the
initial independent variables and the final dependent behavior. 8 It was this assumption that such a model is
eventually to be correlated with neurophysiology which led me in another place
[25] to describe it as a "pseudo-brain model." It may be noted further that the position I
am adopting here is similar to, but slightly different from that proposed by
Krech [9 and 10] - He has proposed a much more explicitly neurophysiological
model. But the substantive properties
which be ascribes to his neurophysiological "dynamic system" draw, I
believe, more from the findings concerning hehavior or action than from neurophysiology
as such. That is, Krech is proposing to
remake neurophysiology in the image of psychology, whereas I am proposing
merely to make for psychology a pragmatically useful model of its own - leaving
it for the future to discover the correlations between it, this pseudo-brain
model, and a true neurophysilogical model. £viuuel POSTULATED CAUSAL CONNECTIONS Before turning to the model
itself we must consider briefly the fourth category listed above that of the postulated types of causal
(functional) relationships between independent variables, intervening variables
(hypothetical constructs), and the dependent variable of behavior. It is obvious from the above discussion of
types of intervening variables that the types of postulated interconnecting
causal functions assumed by the different theories will be intimately connected
with the types of assumed intervening variables. To analyze these relationships for each of the three major types
of theoretical trends cited above would be too great a task to attempt
here. All I shall do is to develop my
own model and to indicate the sorts of interconnecting causal functions assumed
in it. Let us turn then to the model. 285 2 The Model A schematic outline of the model is presented in Fig. 1. The
three groups of independent variables are located at the left. The interventng variables, which in this
exact form are, of course, peculiar to the present
theory, are presented in the middle. And the dependent variable of behavior has
been placed at the right. The nature of
the independent variables and of the dependent behavior have already been
discussed so that the features of the model which need principal consideration
here are the intervening variables. These intervening variables
comprise six main yet closely interconnected items: need system (A) ;
belief-value matrix (B) ; immediate behavior space (C) ; locomotion within the
immediate behavior space (D); restructured behavior space resulting from
locomotions or from learning or from the psychodynamic mechanisms (E) ; and
capacity and temperamental traits (T), which are assumed to interact with the
other five and with their functional interconnections. The arrows in Fig. 1 represent
the postulated directions of causal determination (see IV). The solid arrows indicate the initial
directions of such causation. These
causations are assumed to result from the original values of the independent
variables before behavior or restructuring have taken place. The broken-line arrows represent subsequent
causations in reverse directions. That
is, the changes in the behavior space due to locomotions, learning, and the
psychodynamisms may result in changes in the belief-value matrix, which in turn
may produce changes in the need system.
This matter of causal determination is, however, complex and will become
only somewhat cleared up in the course of our further discussion of the
intervening variables. Finally, it is
to be noted that locomotion (represented by the arrow which is an intervening
variable to be distinguished from the actually observable behavior, is
conceived both to cause a restructuring of the behavior space and to produce
the externally observable dependent variable of behavior. 288
A Psychological Model Let us turn now to a more
full-bodied diagram. In Fig. 2 the picture has been elaborated for the
specific case of a hungry actor going to a particular restaurant and ordering
and eating a particular food.1
We may examine the intervening variables one at a time. NEED SYSTEM (A) Before we look at the details of the need system as here
pictured, let us consider first the concept of "need" itself. It is
to be distinguished from the concept of "physiological drive," which
is to be conceived as an independent variable.
In most previous psychological literature the tendency has been to use
the two terms interchangeably and the two concepts have not been clearly
distinguished. I propose, however, to
differentiate between them and to use the term "drive"
for an initiating physiological condition only and to use the
term "need" for a postulated
resultant, intervening, behavioral process to be defined in the last
analysis as a readiness to get to and to manipulate in consummatory fashion (or
to get from) certain other types of object.
Thus, for example, in referring to the hunger
"drive," I would be referring to the physiological conditions of the
digestive apparatus and of the blood stream, or whatever, which are presumably
the main factors in producing the hunger need; whereas in speaking of the
hunger "need" itself, I would refer, rather, to a readiness to go to
a standard food and to consume it and at the same time to go away from
nonfoods. Similarly, in referring, say, to a dominance "drive,"
I would be referring to some as yet unspecifiable neuro-physiological
condition, which when in force would predispose the actor to exhibit a
dominance "need"; that is, a readiness to go toward a standard goal
situation which would support the consummatory behavior of dominating another
individual or individuals. It should, however, be noted
that although I have chosen "drive" to designate the physiological
determiner and "need" to designate the resultant psychological or
behavioral readiness, some psychologists may prefer just the reverse
usage. They may prefer, that is, to use
"need" to indicate a physiological defect and "drive" for
the resultant behavioral disposition. I
am doing it the other way around simply because for me
"drive" has a more physiological connotation and "need" a
more psychological one. And it
may be noted that my use of "need" is the one which has been adopted
and given wide currency by Murray [17].
Again, it should
also be pointed out that this distinction between "drive" as an
essentially physiologically defined condition and "need" (though
having itself, of course, its own unique neurophysiological correlate) as an
essentially psychologically or behaviorally defined condition allows for the
fact that in certain situations the magnitude
of a given need may be determined more by the character of the given
stimulus situation than by the strength of any single univocally
correlated drive. Thus, for
example, the presence of an especially tempting food as part of the stimulus
situation may arouse the hunger need even though the hunger drive as a
determining physiological condition is, and remains, relatively weak. 1 In this diagram the
broken-line "back-acting" arrows and the trait variables have, for
the sake of simplicity, been omitted. Finally,
it may be noted that a given need, even when caused primarily by the arousal of
physiological drives, may characteristically be activated not by a single drive
but rather by some combination of drives.
Thus, for example, the hunger need itself (defined as the readiness to
go toward and to consume food) may in certain individuals be activated not
univocally by the hunger drive but also by the arousal of other drives such as
gregariousness, fear, or the like. For
by now it is a psychological commonplace that in certain
individuals an aroused gregariousness or an aroused fear drive may also express
themselves in the hunger need. Individuals so affected will be observed to
overeat in loneliness or in stress. But let us now look at the need system as a whole as
depicted in Fig. 2. It is represented by the circle labeled A and is shown as containing a number of
interconnected compartments. Each of
these compartments is intended to represent a single need and, when the need in
question is aroused, the compartment is conceived to contain both positive and
negative "electromagnetic" charges.
When the need is strong the compartment contains many charges; when the
need is weak it contains few charges.
In this particular diagram the two main needs depicted are a libido need, so labeled, and a hunger need,
labeled H. Two undesignated needs are
also indicated. The term
"libido" has been used without any preconception as to the nature of
this need libido, other than that it seems to be correlated primarily with some
physiological energy and that it seems to vary in average magnitude from
individual to individual and to go up and down in any one individual with such
factors as health, sickness, and time of day.
Further, it is postulated that this libido need has no specific goals of
its own, but through contact with all the other specific needs (whatever the
list of these may eventually turn out to be), it adds magnitudes to the
electromagnetic charges in these other specific needs. Thus, it is assumed that when the energy or
tension in the libido is great, there will be more tension than there otherwise
would be in all the other need compartments.
It is assumed, in short, that the energy or tension in the different
special compartments of the whole need system maintains some sort of an
interactive balance with the tension in the basic libido compartment. Look now at the hunger need,
labeled H. The amount of energy (number of plus and minus charges) in this
hunger need is conceived to be determined primarily by the strength of the
physiological hunger drive and in some degree by the character of the stimulus
situation (i.e., whether or not food objects are present) - For it will be
assumed that, if food stimulus objects are present, and especially if they are
particularly tempting, the strength of the hunger need will be increased over
what it otherwise would have been as a result of the mere hunger drive. Also the strength of the hunger need will be
affected, other things being equal, by the strength at the moment of the
general libido tension. 290
A Psychological Model Before we turn to the next
feature of the diagram, let us discuss briefly the positive and negative
charges, which are represented by the small plus and minus signs. The introduction of these charges as
constituting the nature of need tensions is an electromagnetic analogy which
should not at present be taken too seriously.
It seems useful, however, as a way of summarizing the fact that needs
express themselves both as a readiness to go to
certain types of objects (positive valuing), and as a concomitant readiness to
go away from
certain other types of objects (negative valuing). When the specific hunger need is high in tension, the actor may
be conceived to have both a strong positive readiness (many plus charges) for foods and strong negative readiness against nonfoods. An increase in the hunger tension within the hunger compartment
is to be represented by an increase in both positive and negative charges. The positive charges are then conveyed - see
arrows in the diagram - to one end of a belief-value matrix (or a number of
different belief-value matrices) and the complementary negative charges are
conveyed to the opposite end, or ends, of such a matrix or matrices. The further usefulness of this introduction
of the concept of positive and negative charges will become more obvious when
we turn now to a consideration of the belief- value
matrix.2 BELIEF-VALUE MATRIX (B) In the situation depicted in Fig. 2, the belief-value matrix chosen for
representation is a relatively banal one.
It is constituted by the cognitive categorizations, beliefs, and values
of an actor (obviously one in our culture) relative to hunger deprivation,
hunger gratification, foods, and restaurants.3 In understanding this matrix it will be
helpful if we consider it in successive steps.
Let us then turn to Fig. 3, which
represents this same matrix but with the value features left out and only the
categorizations (differentiations) and beliefs left in. 2 It should be noted that the concept of belief-value
matrices is practically identical with that of need-dispositions in the
preceding parts of this book. 3 It will be pointed
out below that any such simple single matrix as this will, of course, be only
one among a vast number of matrices (i.e., sets of categorizations, beliefs,
and values relative to other types of objects and other types of gratification
and deprivation) which must be assumed to be "built in" (some with
more permanency than others) into any developed personality. 4 The term
"image" is here used in a purely objective (noaphenomenological)
sense, to designate a categorized readiness to perceive, which the actor in
question is to be thought to possess as a result of past experiences with
instances of the given type of "object." It has already been indicated in the preceding parts of this book
that the term "object" is to be used to cover a number of
"modalities," such as the qualities of self and of others; the
performances of self and of others; the qualities and performances of
collectivities; cultural accumulation objects (physical artifacts, languages,
ideologies, industrial techniques, skills, etc.), as well as purely physical,
environmental things. The
Model 291 The small squares or "cells" in Fig. 3 represent "typed" (not concrete)
images of objects 4 which the individual possesses by virtue of the
differentiations and categorizations of the object world (specifically in this
case of restaurants and foods) which he has previously acquired. These typed images are conceived to be
arranged along functionally defined "generalization dimensions" - in
the present example, along a restaurant (or food-providing) generalization
dimension and along a food (or hunger-gratification-providing) generalization
dimension. The units of these generalization dimensions, if we had them, and
the relative Fig.
3 BELIEF-VALUE MATRIX re TYPES OF RESTAURANT AND TYPES OF FOOD
(Values omitted) locations
along these dimensions, stated in terms of such units, of the different typed
images would indicate the precise degrees of functional similarity between such
images. Our present lack of knowledge
concerning the nature of most such dimensions and their units and the relative
locations of the specific "typed" images along them is, I believe,
one of the reasons why psychology is at present so frustrating. In other words,
I would argue that we have to date little theoretical or empirical knowledge as
to what and how many such dimensions of generalization can be assumed and what
their units are for any given actor or even for most actors in any given
culture - save for simple, perceptual dimensions such as colors, shapes, sizes,
pitches, loudnesses, tastes, etc. 292
A Psychological Model By way of illustration of the
sort of further data which we need, I would point to a recent important
contribution toward the discovery of a really new and important generalization
dimension for children in our culture provided by the work of Sears and his
collaborators on doll play [20]. They
have discovered by working with nursery-school children that in expressing aggression, the children arrange their dolls in a
certain order along a dimension - "things
to be aggressed against."
If a child is high in need aggression (and there is little conflicting
need fear) the child will, in his play, express the most aggression against the
parent doll of the same sex, the next most aggression against the parent doll
of the opposite sex, the next most against the child doll of the same sex,
the next most against the child doll of the opposite sex, the next most against
the baby doll, and finally the least against the walls and furniture of the
doll house. This, then, is an
empirically discovered new type of generalization dimension - a generalization
dimension which, with regard to the specific ordering of the objects along it,
is obviously a function of our culture.
Children raised in other cultures with different family constellations
might well order their dolls for the purposes of aggression in a different
sequence. But many other such
generalization dimensions for all kinds of functional purposes, as well as for
aggression, must be discovered and made precise if psychology is finally to
achieve wide practical significance. Look next at the looped arrows
of Fig. 3 . These arrows (or, we might call
them "lassos") represent
what I call means-end beliefs 5 and are very important constitutive
features in the representation of any matrix. The double tail at the beginning
(left) of each arrow, or lasso, represents the generalization spread as regards
the kind of initial (terminus a quo)
typed images. That is, the spread and
shape of this forked tail indicate the range of initial types of objects or
situations along one dimension which will be accepted (and with
what degrees of readinesses) as appropriate means (or instigations) for
releasing a given type of behavior-act in order
to get to such and such a further terminus
ad quem type of object or situation. The cell or
cells caught by the loop (to the right of the fork) of a lasso indicate the
specific differentiated types of terminus ad
quem end-object or situation, which the actor
"believes" will be achieved by the type of behavior-act included in
the lasso. And the little arrow issuing
out of each generalization dimension indi-cates the particular type of
behavior-act which is thus believed to be involved. Specifically, the diagram as a whole represents, then, a
hypothetical case in which the actor in question has come to believe 6
that, if he is hungry (i.e., if he negatively values hunger deprivation and
positively values hunger gratification) and if he should respond by the act of
searching for and going to (indicated by the small straight arrow) certain
types of restaurants, and if those restaurants were responded to by the act of
spending (indicated by the zigzagged arrow) this would get him to certain types
of food, and that, if these certain types of food were responded to by the act
of eating (indicated by the corkscrew arrow), it would get him finally to
hunger gratification. 5 What I am
here calling "beliefs" are essentially what were called
"expectations" in the preceding parts of this book. I am using the
term "beliefs" to designate the readiness or potentialities for
expectations which the actor entertains, and am reserving the term
"expectations" to designate the concrete particularized instances of
such beliefs which result from the presence of a particular stimulus situation. Thus (see Fig. 2)
whereas I would locate "beliefs" in the belief-value matrix, I would
locate "expectations" or
particularizations of those beliefs in the behavior space. The
Model 293 But Fig.
3 presents, as has been said, only the cognitive
(categorizatioti and belief) side of a belief-value matrix. In actuality, the
categorizations and beliefs are usually accompanied by value concomitants. Thus, the individual will usually at one and
the same time not only believe that certain types of food, if eaten, will lead
to hunger gratification but also will have a positive value for hunger
gratification and a negative value for hunger deprivation. Similarly, he will have positive value for
certain restaurants and certain foods.
In other words, a complete diagram of any matrix will tend to contain
not only differentiations (categorizations) and beliefs but also values 7
– that is, goodnesses or badnesses deposited on the various cells of the
matrix. In Fig.
4, therefore, plus and minus value charges have been added. The magnitudes of the final positive
hunger-gratification value and the final negative deprivation value at any
given moment are conceived to be largely determined by the plus and minus
charges in the hunger-need compartment at that moment
(see connecting arrow in Fig. 2, issuing
from the hunger-need compartment, bifurcating, and running to both ends of the
matrix). But the magnitudes of these
values may also be determined in part by the character of the presented
stimulus situation. Some stimulus
situations will in themselves tend to activate a given matrix. 6 A "belief" is operationally defined as a
connection that makes a readiness to perceive and to behave in a certain way
relative to one type of object (as end) give rise to a readiness to perceive
and to behave in a certain way relative to certain other types of objects (as
means). What I am here calling a belief
is thus essentially what I have previously designated [22] as a "means-end
readiness." 7 It should he
emphasized that I am using the term "value" here in a more specific
and special sense than it is used in the other sections of this hook. I am using it to designate what types of
object or situation will in the given context of need-activation and belief
tend to be approached or to be avoided by the given actor. 294
A Psychological Model Next, given types of food are
represented as having different degrees of positive values insofar as they are
"believed" to lead on successfully to hunger gratification and away
from hunger deprivation. The
relative successes with which the different foods are thus believed to lead on
are represented by a generalization
gradient - that is, by the shape of the initial terminus a quo, or
forked, end of the lasso. The plus
value is brought back by the arrow (see the plus sign in the circle at the
point of the arrow). Then this
value is distributed to the specific types of food according to the shape of
the forked tail of the arrow. Further,
in order to make the illustration as general as possible, one type of food is
indicated in the diagram as actually having negative value because "it is
believed" by the actor actually to lead away from hunger gratification and
toward hunger deprivation.8
Similarly, the different types of restaurants are represented as having
different positive values because they are believed to lead on with different
degrees of success to a given valued type of food. Fiq.4 BELIEF-VALUE MATRIX a. Modal matrix for a culture. A further point to be noted, now, is that Fig. 4 may also be used to represent (for a very
simple area) a culturally and sociologically determined belief and value system
as shared by a community of individuals.
This belief and value system, this matrix, if correctly inferred by the
observer for a relatively homogeneous group of individuals, would be statable
in verbal propositions such as: (1) If X (the modal individual in this culture)
were positively valuing hunger gratification and correspondingly negatively
valuing hunger deprivation, he would value certain
types of food-to-be-eaten in a certain order by virtue of his beliefs
concerning their respective gratification-producing characters. (2) If X
values certain types of food, he would be ready to value
positively certain types of restaurant because of his belief that these
types of restaurant would lead to the certain types of food through the act of
spending. 8 It is to be noted
that the concept and term "cathexis" will refer to these relative
degrees to which the different foods are believed to lead on to hunger
gratification and away from hunger deprivation. The food
which through the behavior of eating is believed by the actor to lead most
readily to hunger gratification would be the most strongly
"cathected" food. The
Model 295 Further, insofar as
questionnaires and verbal reports can be considered reliable forms of data,
these propositions about the matrix could be relatively adequately inferred
from mere questionnaires or interviews.
Thus, for example, one could ask the subjects: (1) "What
are you ready to do when you haven't eaten for a considerable length of
time?" (2) "What
kinds of food do you like? Name six varieties of food in order of preference.
What do you like about each of these six?" (3) "For
each of these six foods what types of restaurant would you go to and in what
order? List all the considerations you
would take into account in choosing the one kind of restaurant or the
other." Suppose we now obtain from such
questionnaires (of course, far better constructed ones) statistically reliable answers indicating
that practically all the individuals of a given social status 9 in a
given culture give practically the same answers. We could then use these answers to define the modal belief- value
matrix relative to foods and restaurants for individuals of this status in this
culture. This would be the
statistically average or "modal" matrix relative to a given need and
a given type of environmental situation for the given group of similarly placed
individuals. Indeed, such a modal
matrix, or rather a whole collection of such modal matrices, shared by a whole
society and concerned with relatively basic needs and relatively general
features of the environment, is, I believe, what anthropologists have sometimes
called the ethos of a
culture. By conceiving such an ethos as
a very large and relatively general belief-value matrix system I am, I believe,
merely saying in my terms what has already been implied in Chapter III of Part II. b. A matrix equation or modal matrix for an individual. It must be noted, now, that a matrix for a
single individual may also be conceived as a "modal" affair. It may
be drawn, that is, to represent not the actual momentary absolute magnitudes of
the values, the beliefs, and the categorizations, but rather their
average or modal magnitudes relative to one another. In other words, a matrix
may be drawn to represent merely an equation in which specific absolute
magnitudes have not yet been substituted.
When a
matrix is so drawn, the absolute magnitudes on any specific occasion will be
arrived at by the substitution
in this matrix equation of the then-and-there absolute magnitudes of the
attached need or needs and of the specific stimulus situation. The formal or modal matrix is a
"mathematical equation" which states the functional interrelations
between the variables within it. The
absolute magnitudes to be given to these variables on a given occasion will be
determined by the substitution in this equation of specific need-gratification
and need-deprivation magnitudes resulting from the then-and-there aroused
physiological drives and front
the then-and-there presented stimulus situation. Let us turn now to C - the behavior space. 9 See Linton [141 for
this use of the term status. 296
A Psychological Model BEHAVIOR SPACE (C) Return to Fig. 2 and look at C. There are two main types of causal arrows
shown as determinative of the contents of C:
a causal arrow springing directly from the stimulus situation of the moment,
and three causal arrows coming down from the belief-value matrix. A behavior space is thus to be defined as a particularized complex of
perceptions (memories and inferences) as to objects and relations and the
"behaving self," evoked by the given environmental stimulus situation
and by a controlling and activated belief-value matrix (or perhaps several such
matrices). What is perceived (expected) is thus determined by what is presented
by the stimulus situation at the moment and by the store of categorizations,
beliefs, and values which the actor brings to the presented stimuli. Or, to put it another way, the immediate
behavior space is to be defined as an array of particular objects, in
such-and-such particular "direction" and "distance"
relations to one another, which are perceived by the actor at the given
moment. And some of these objects tend
to have positive or negative "valences" on them.10 Among such particular objects a very crucial
one is the actor's self (designated as the behaving self) which is also a part
of any such perceived array. The words "perceive"
and "perception" have been chosen as the most appropriate ones for
summarizing the behavioral character of a behavior space.11 Several provisos are, however, to be kept in
mind in this use of the terms. The first proviso is that
"perception" as here used covers not only perception in its narrow
meaning of strict sense-perception, in which physical stimuli for the
corresponding objects are all then and there present to the senses, but also
includes the perception of objects some of whose parts are, in common- sense terms, merely inferred or
remembered. Thus, for example, I
myself, as the actor, may often be considered to perceive" (i.e., to have
present in my immediate behavior space) not only the objects on my desk and
before my eyes at the moment, but also some of the familiar objects on the wall
behind my back or even such an object as the University Campanile outside my
window at the left. For I might well be
found to be immediately ready to behave toward all these other objects and not
merely toward those on my desk. 10 A belief-value
matrix contains "universals"; a hehavior space contains
"particulars." Thus a
"valence" is a particularization of a value. In other words, whereas
a value is deposited on the image of a type of ohject, a valence is deposited
on the perception of an instance of such a type of object. 11 In other places
[23 and 24] I have used the terms "apprehend" and
"apprehension" for this immediate character of the behavior
space. It is felt now, however, that
the good old words "perceive" and "perception" used in a general sense more nearly carry the desired
meaning. Further, this broader use
of the term "perception" includes whatever concrete particularized
instances of spatial, temporal, aesthetic, mathematical, or other such
relations as may be immediately given along with the objects themselves as ways
of getting from one object to another.
That is, a behavior space will contain not only particular objects but
also their particular spatial and temporal, or other, relations to one
another. Or, in other words, the
"medium" (i.e., the "directions" and "distances"
constitutive of a behavior space) may be not only spatial and temporal,
but also mechanical, aesthetic, mathematical, or the like. Thus, for example, when I behave on a
specific occasion in terms of the number system, the French language, simple
logical principles, simple aesthetic principles, or the like, these
particularizations as to how to get from one object to another are
to be conceived by the psychologist as at that moment "there" in my
behavior space along with the objects which they relate. A third proviso is that
"perception" (and hence the behavior space) may also include entities
of which the actor is not then and there consciously aware. Any concrete particular objects or relations
which govern the actor's immediate action are to be said to be in the behavior
space - that is, to be also "perceived," whether or not (in
introspective terms) the actor is then and there consciously aware of them.12 But let us look now in more
detail at the behavior space presented in Fig. 2. The entity surrounded by the two concentric
circles and labeled B.S. is
the "behaving self." This is
the actor's behaving self as perceived by him at the given moment and perceived
as located in a certain way with respect to other objects. The symbol a
represents a particular food of the type A
perceived as within the available environment.
And x is a particular
restaurant of the type X also
perceived as within the available environment.
The plus signs on a and
x represent positive
"valences." 13
These positive valences are represented as determined (see arrows) by
the positive values in the belief-value matrix on the corresponding typed
objects X and A.
It is to be observed next that there is a negative charge within the
behaving self. This negative charge is
labeled "need-push," which is conceived as evoked in the behaving
self by the negative charge in the hunger-deprivation compartment in the
belief-value matrix (see connecting arrow). That is, any need deprivation in a
matrix arouses in the behaving self a corresponding negative need-push. Further, given a positive valence and a
complementary negative need-push, there will result a field force tending to
push the behaving self toward the positive valence. Also, if there were a negative valence and a corresponding
negative need-push, there would result a field force tending to repel the
behaving self away from such a negatively valenced object. In the present case the behaving self is
impelled first to the restaurant x and
then to the food a. And this will tend to result in the actual
locomotion of the behaving self first to x
and then to a. 12 Just what the
further refinements may he, as far as the governance of behavior is concerned,
between those behavior-space contents which are consciously present and those
which are unconsciously present is a question which I shall touch upon below
but to which no completely adequate answer can, I believe, now he given,
because of the present inadequate state of our empirical knowledge. 13 Plus and minus
signs in a belief-value matrix (see above) indicate plus and minus
"values" for categorized types
of objects. Plus and minus signs on
objects in the behavior space indicate the resultant concrete pulling or
repelling properties of particular, perceived instances
of such types of object. The term
"valence" is used to distinguish these specifically located pullings
and repellings from the values. The
term "valence" has, of course, been borrowed from Lewia [13]. 298
A Psychological Model An important point to be
emphasized is that some such concept as that of a need-push (perhaps a better
term for it could be found) seems to be necessary. Thus the food need-push (corresponding to the activated strength
of hunger deprivation) may be relatively great in a given instance, although no
strongly valenced food or foods may be present in the immediately perceived
behavior space. In such a case, I would make a further assumption: to wit, that
such a need-push may evoke, by some process analogous to "electromagnetic
induction," positive charges in any regions of the behavior space which
the behaving self is not then in. As a
result of these "induced" positive charges the behaving self will be
attracted to such regions. When, however, the behaving self locomotes
to them, no actual food is perceived and hence no discharge
takes place. The food need-push remains
practically unreduced. I assume further
that thereupon some small portion of the negative electric charge of the
behaving self need-push spreads by "conduction" to this immediately
surrounding region which makes the region then somewhat negative; whereupon the
behaving self is propelled away from it to new regions. It would, in
short, be by some such assumptions that I would explain some of the restless,
exploratory behavior of a hungry animal who does not yet perceive actual
food in any region in the behavior space. Let us now consider still
another important feature of the behavior space. This concerns the fact, already noted, that
the "directions" and "distances" which are constitutive of
a behavior space may be other than spatial; they may be temporal,
mechanical, social, mathematical, and so forth. In other words, the "locomotion" of the behaving self
which would get it to or from a perceived object, may involve time
manipulations, mechanical manipulations, social manipulations, or mathematical
manipulations, as well as simple spatial manipulations (i.e., mere spatial "goings-to" or
"goings-from"). To what extent these different dimensions of locomotion have
to be conceptually separated out in the case of a complex behavior is not yet
known. It may turn out that for predicting
any save the simplest behavior one will have to draw different behavior spaces
lying in different locomotor planes (dimensions) and that the final locomotion
must be depicted as some resultant of all these according to some principles
analagous to those of descriptive geometry.
For the present, however, I shall leave such further complications aside. An actually, I personally tend to believe
that an actor himself does not normally have any clear differentiation between
such different dimensions of locomotion.
A given goal object is "over there" in space and time,
mechanics, society, and so on, all at once.
The perceptions of the directions in all these dimensions occur
simultaneously as some sort of total Gestalt.
For the present, then, until further work has shown it necessary to
assume otherwise, I shall hold that locomotions, however complicated, can all be represented as occurring in a single
behavior-space plane. But let us consider now in more detail the
nature and result of such locomotions. The
Model 299 LOCOMOTION AND RESULTANT RESTRUCTURING OF THE BEHAVIOR SPACE Any diagram of the behavior
space (see C in Figs. 1 and 2) can obviously indicate a state only and not a process. The C's
in a diagram depict the assumed behavior space before locomotion (or other
restructuring) takes place. And the E's represent the new behavior
space after such locomotions or a succession of such locomotions (resulting in
learning or in one of the psychodynamic mechanisms) have taken place. What, now, is the nature of locomotion? First of all it must be emphasized
again that locomotion is an intervening
variable - a hypothetical construct - which is not identical with the
overtly measurable dependent variable of behavior to which it gives rise. Locomotion is a passage
of the behaving self from one region of the behavior space to another
(or through a succession of such regions).
It is such passages from region to region that express themselves in
overt behaviors; but such passages or locomotions are not identical with these
resultant behaviors. But this raises another
important question which we have slid over until now. What is meant by a region in the behavior space? A region in the behavior space is to be defined in the last
analysis by the sorts of behaviors which the actor perceives as possible for
the behaving self if the behaving self is in
that region (i.e., in the presence of such-and-such an object or
objects). Such an array of possible
behaviors is obviously dependent both upon the types of behavior which the
given actor is capable of, as a result of
innate endowment and previous sensory-motor learning, and upon the presence of
objects which will support such
behaviors. Behavior cannot take place in
a vacuum. When an actor perceives,
say, restaurant x as over there
(when, that is, be perceives the behaving self as now in one region and
restaurant x as in such-and-such
another region) he perceives his behaving self as now presented with
such-and-such an array of immediately possible behaviors. He also perceives that one of these
behaviors defining the present region will get his behaving self to the region
of restaurant x, which latter
will be defined by certain further possible behaviors (such as ordering steak,
eating, tasting, etc.). Locomotion is
thus a selection from perceived
immediately possible behaviors as the way to get to such-and-such another
region - other perceived immediately possible behaviors. And what we call an object in the behavior space is
essentially a part of a region and a collocation of perceived
"supports" for such-and-such particular behaviors. It must be emphasized further that among
such behaviors, which an object will support and which thus define any object
or region in the behavior space, are to be included the purely perceptual
discriminatory behaviors which such an object makes possible. These discriminatory behaviors constitute a
large part of the defining characteristics of any object or behavior-space region. That is, a behavior-space region or object
is defined both by qualities - i.e., the discriminatory behaviors - which it
will support, and by the types of other more manipulatory
behaviors which it will also support.14 300
A Psychological Model To summarize, a locomotion is a
selection from one or more perceived immediately possible behaviors (i.e., the
region in which the behaving self is initially located) as the way to get to
such-and-such other potentially possible behaviors (the region to which the
behaving self is locomoting). A
locomotion in the behavior space is thus not a
behavior itself but a selection or a series of selections which
result in a behavior or in behaviors. It is to be noted further,
however, that locomotion not only causes behavior but the continuance and success
of locomotion is, of course, contingent upon the fact that the behaviors to which it gives rise shall actually take
place and be successful.
Let me illustrate by an example.
Assume an actor who perceives his behaving self as in the region of
"being in the house" (i.e., as in
the presence of supports for such-and-such discriminatory and manipulatory
behaviors). Assume that he also
perceives another region, that of "being
outside in the garden," where such-and-such other discriminatory
and manipulatory behaviors would be supported.
The garden, let us say, has a positive valence, and we shall assume that
there is in the behaving self a need-push complementary to this positive
valence. The behaving self will,
therefore, be propelled
to locomote in the direction of the garden.
This locomotion will consist in a successive
selection from the perceived immediately possible behaviors of those
special behaviors which are perceived as appropriate for getting the actor into
the garden. If the original behavior-space
perception was veridical, then these successive selections - this locomotion -
will give rise to a series of actual sensory-motor acts which will in fact get
the actor into the real garden and thus lead to a new behavior space in which the behaving
self will be perceived as in the garden. If the
original behavior-space perception had, however, been nonveridical, then the
locomotion would have consisted in the selection of
inappropriate behaviors - ones which would not have got the actor into the real garden or the perceived behaving
self into the perceived region of a garden. The
Model 301 14 These two types of
support are what I have previously distinguished as "discriminanda"
and "manipulanda" [22]. The E's in the diagrams (see Fig. 1) represent, then,
the new layout or restructuring of the behavior space which results when
the behaving self has locomoted, let us say successfully, to a new region - a
region which was perceived as in a certain direction and at a certain distance
in the original behavior space before such locomotion took place. It is to be noted that such restructurings
as a result of mere successful locomotion are quite simple. They consist merely
in the fact that the behaving self is in a different position relative to
objects and regions than it was before.
If the actor were to be put back into the same initial objective stimulus situation that he was in
before the locomotion and consequent behavior took place, his behavior space
might well be, to all intents and purposes, the same as it was on the previous
occasion. Restructurings
of another sort, however, do occur.
In such cases, if the actor is put back into the original objective
stimulus situation, he will perceive a definitely different behavior space from
that which he perceived on the original occasion. How do these more permanent restructurings come about? They also
originate out of locomotions. But in
these cases the locomotions lead not simply to a new position of the behaving
self in the behavior space but rather to enlargements or other fundamental restructurings of the behavior
space. Further, these enlargements or restructurings may be roughly
separated into two main classes: (1) those
resulting from learning and (2) those
resulting from the psychodynamic
mechanisms. Moreover, it will also appear that learning and the
psychodynamic mechanisms involve not only such enlargements or restructurings
of the behavior space but also correlated changes
in the belief-value matrix system.
Before
passing on, however, to these questions of learning and the psychodynamisms we
must briefly note the other main item among the intervening variables of the
model: capacity and temperamental traits. CAPACITY AND TEMPERAMENTAL TRAITS (T) At the top of the middle portion
of Fig. 1 there is an entity labeled T.
This symbol represents a whole collection of individual
difference variables or traits.
These traits are shown to produce an array of radiating causal arrows
which are conceived to impinge (although these impingements are not indicated)
upon all the other intervening variables and upon the various interconnecting
functions. This feature of the diagram
is, of course, no more than a mere formal acknowledgement that there are trait
variables. The fact that no terminations for the arrows have been
indicated is a confession that no clear hypotheses have been worked out as to
the relations between trait variables and the other or "content"
variables. The truth
seems to be that the methods hitherto used to arrive at trait variables -
ratings, test scores, inter- correlations,
and factor analyses - have as yet for the most part never been closely
integrated with the methods and variables used in the determination of the
content variables. Traits are presumably constants or parameters in
the equations determining the magnitudes of the content variables, but they
have practically never been studied as such. 302
A Psychological Model For example, we do not yet know whether such a trait as
the I.Q. (obviously one of the most
studied) consists primarily in
individual differences in the formation of appropriate belief-value matrices, in the
ready perception of adequate behavior spaces, in the
rapid learning of specific behavior-space expectations, in the
readiness to translate behavior-space locomotions into appropriate muscle
responses, in the
presence of useful and the absence of handicapping psychodynamic mechanisms, or
in some
combination of all of these. We must
conclude, in short, that the necessary empirical and theoretical work which
must be done in order to integrate traits with contents is still largely
untouched. With this very brief and purely
formal acknowledgement of the problem of traits, let us return now to a
discussion of learning and the psychodynamic mechanisms. 303 3 Learning and the
Psychodynamic Mechanisms There
seem to be two distinctive types of empirical problems with one or the other of
which practically all studies of learning have been concerned. These two problems or setups I shall call
the pure association setup and the reward setup. Let us consider them
successively. a. The
pure association setup.
Here the stimulus situation presented to the actor consists in two (not
at the time specifically valued or valenced) objects which are presented in a
given temporal and/or manipulative order.
For example, a buzzer is sounded and followed after a mere time interval
by a light. or the
animal goes from one place in, say, a maze by a certain manipulative running
activity of his own to another place in the maze. If we assume that the actor's behavior-space perceptions
correspond more or less veridically to the objective stimulus objects, this
means that the rat's or the human being's behaving self is first in the region
of the one object and then - either through the locomotion
of "selecting to wait through time" or else through
that of "selecting out a special set of spatial, mechanical, or other
manipulatory be- haviors"
- it comes into the region of the secondd object. When this succession of the two objects (together with the
interconnecting locomotion) has been repeated a number of times, it will be
found (if, that is, learning takes place) that on a subsequent occasion the stimulus
situation of only the first object will evoke an enlarged behavior space which
will contain not only the perception of this first object but also a "perception" of the second object
and of the direction and distance of the locomotion which led from the first to
the second. It is my contention that such
learning by "seemingly pure association" can and often does take
place without either of the two objects being,
to any appreciable degree, an immediate goal object
for any specific utilitarian need such as hunger, thirst, sex, or the like. Thus, for example, rats in the so-called “latent
learning" experiments can apparently often learn "what leads to
what" in a maze, even though the final "what" may not be a goal
object for any special need which may be operative at the moment. However, it
also appears that, if the animal is highly
apathetic or has in his behavior some other object or objects
with very high valences for specific needs, which are then and there active,
the animal is less likely to learn (i.e., acquire this type of associative
enlargement of its behavior space). 304 A Psychological Model We have
to explain two contrasting facts: first, the fact that such learning tends to
take place without either of the two objects being specifically a goal object;
and second, the fact that such learning does not tend to take place if the
behaving self is either totally lacking in need-pushes or contains a need-push
for which some other irrelevant, but strongly valenced, object is present. To explain these contrasting facts I shall
have recourse to the notion of a general exploratory
or curiosity or 'placing" need. The need compartment corresponding to such an exploratory or
placing need is assumed to be in close communication with the compartments for
all the other needs so that the arousal of any of these other needs - hunger,
thirst, fear, sex, dominance, etc. - will also arouse the general exploratory
need. However, the laws by which other
needs are to be assumed to arouse this general need are not simple. Thus it would appear that, whereas an
increase of a specific need will tend to cause an increase in the general
exploratory need up to a certain magnitude
of the specific need, beyond that point further increases in the magnitude of the specific need will tend to decrease the magnitude of
the exploratory need. In any event, given that the exploratory need is aroused
in some degree (whatever the cause), this, by definition, will mean that all
types of new objects will have positive value in a corresponding exploratory or
placing belief-value matrix and that perceived instances of such objects in the
behavior space will have positive valences.
To reach these valences the behaving self must locomote in the direction
prescribed by the character of each object.1 I conclude that in this way "seemingly pure
associations" are acquired. This
does not mean, however, that such associations are acquired without motivation,
but only that the motivation is the relatively
"disinterested" one provided by a general exploratory need. Specific acts are not stamped in but new field relationships are learned. Fig. 5 will
indicate more clearly what this means in terms of the model. The diagram shows the assumed state of the
need system, of an attached belief-value matrix, and of the behavior space upon
the first presentation to the actor (let us assume a rat) of a simple T-maze. The behaving self of the rat is shown in the region of the
bifurcation between the left-hand alley and the right-hand alley. A general exploratory need activates a belief-value
exploratory matrix so that all alleys and other types of objects such as goal boxes are believed to have positive value for gratifying the
exploration need and hence all instances of such objects will be perceived and
valenced in the behavior space. Given,
then, that the rat is in the presence of the two alleys, he will perceive and
be ready to explore them both. Further,
after he has locomoted by releasing exploration down each alley he will also
perceive the characters and contents of the goal boxes reached. 1 See below
the discussion of the Bruner and Goodman experiment. Learning and Psychodynamic Mechanisms 305 Turn
now to Fig. 6.
This shows two behavior spaces: one when the behaving self of the rat
has locomoted to the left-hand goal box; and one when it has locomoted to the
right-hand goal box. The causal arrows
- the broken lines drawn from both behavvior spaces back to the belief-value
matrix - indicate that, as a result of perceiving each alley, locomoting down
it, and then perceiving the resultant goal box and its contents, new beliefs tend to be produced in the
belief-value matrix. These new beliefs
are lassos issuing from each type of alley, involving a certain type of
behavior, and arriving at a given type of goal box. This means that when, on a subsequent occasion the rat is
presented with a left-hand turn or a right-hand turn, his belief-value matrix
will contain beliefs about the further types
of object he would arrive at by making such left- or right-hand turns; and his
behavior space, at the moment of presentation of the two types of alley, will
be enlarged so that he will then
"perceive" not only these immediate turns but also their expected
consequences of one kind of goal box on one side and another kind of goal box
on the other side. Such learning is cognitive in nature, and there seems to be
no differential
"reinforcement" involved in it. 306 A Psychological Model Let us
turn now to the other type of empirical setup in which differential
reinforcement may (without further analysis) seem to play a part. b. The reward setup. The type of experimental setup designated by
this term is the more conventional one. An actor, animal or human, is
confronted with a number of possible alternative behaviors and is motivated by
some specific need, such as hunger, thirst, desire for praise, or the like; one
of these alternative behaviors leads to an appropriate goal for the specific
need while the other alternative behaviors do not. The usually accepted theory to explain learning in this kind of
situation is the so-called reinforcement Fig. 6
BACK ACTION OF LGCOMOTION ON CATEGORIZATIONS AND BELIEFS theory.
Reinforcement theorists argue that learning is produced not by the gratification
of a general exploratory or curiosity or placing need, for which all objects
and relations are equally rewarding, but rather by the gratification of the
specific need through the reaching of the specific goal object for that need. Thus, it is said, because the actor has behaved in a
certain way in the presence of a given stimulus situation and has been led
thereby to a goal object appropriate to the aroused specific need, the
consequent reduction of this need (even though slight and relatively temporary)
increases the tendency to perform this Learning and Psychodynamic Mechanisms 307 same response to the same stimulus on a subsequent
occasion. As an empirical fact this, of
course, is usually true. But in terms
of our present analysis the important point is that the crucial behavior-space objects and directions of locomotion shall have been "noticed," i.e., shall have
brought general exploratory-need gratification whichever response
occurred. For I would contend that even
with rats, to say nothing of men, there is already much evidence to show that
the learning of what would appear to be merely a single response to a single
stimulus may actually result in the ability to make
wholly new but appropriate responses to a set of field relationships. See, for example, the studies of spatial
learning in rats by Ritchie [19] and by Tolman, Ritchie, and Kalish [2628], and
by others [11] which indicate that rats, having learned one path on a maze to
get to food, may be able to short-cut over a new path or to approach the food
correctly from a totally different starting point. According to orthodox reinforcement theory, a stimulus-response
connection is "stamped in" if any need reduction (relevant or irrelevant)
takes place in close temporal continguity after the response. According to the argument presented here, learning consists rather in the acquisition of
perceptions of objects, directions, locations, and valences in the behavior
space and eventually in the resultant acquisition
of generalized categorizations, beliefs, and values in a superordinate
belief-value matrix. And the latter
kinds of learning take place as a result of the gratification of a merely
"cognitive" exploratory need.
However, this exploratory need will itself have in many instances been
heightened by communication from some special specific need. But, in any event, it
is the gratification of the cognitive exploratory need and not the
gratification of the special need which determines the learning. So much
for learning; let us turn now to the second type of process through which
behavior spaces and belief-value matrices may be enlarged or otherwise
restructured: the so-called psychodynamic mechanisms. THE PSYCHODYNAMIC
MECHANISMS I shall
limit myself here to a consideration of only four of these mechanisms. I should hope, however, that the treatment
of these four will indicate the general pattern of approach which could be used
successfully for the consideration of all the other mechanisms. The four I have chosen are: (a)
identification, (b)
the self-ideal, (c)
repression, and (d)
symbols and symbolic substitution. Before considering these
individually, it should again be emphasized that any such mechanism will be
conceived to involve, first, a restructuring of the behavior space relative
to an initial, particular stimulus situa- tion, and second, a resultant and relatively persistent
change in one or more superordinate belief-value matrices. Learning and Psychodynamic Mechanisms 308 a. Identiflcation. Identification seems to arise out of
initial, concrete behavior-space situations in which the behaving self contains
a need-push for love and approval that impels it to locomote toward a region of
behavior exemplified and approved by parent or other loved individual. As a result of such locomotions and the
discovery of how to behave to obtain the love and approval of the parent, the
given actor may develop a general belief that to
behave "similarly" to the parent or other authority figure is
a good way to get such love and approval. Fig. 7 BEHAVIOR SPACE IN IDENTIFICATION Freud,
in first introducing the concept of identification, assumed, as is well known,
that the small boy identifies with the father (i.e., comes to behave as the
father requires and in a manner similar to the father) to retain or regain not
so much the love of the father as the love of the mother. That is, Freud believed that there was a
really sexual sort of competition between the small boy and the father for the
mother's love and that the boy came to believe that, if he behaved like the
father, then the mother would love him as she did the father. Whether or not we accept all the
implications of this Freudian analysis, it seems to be pretty much agreed that
identification does involve at least two factors: (1)
locomotion toward a
region of love and approval from some alter or alters by ego's behaving in a
prescribed way similar to that of the alter or alters; and (2)
locomotion away from some
other region of valenced activity because of the stronger need-push to get to
the region of love and approval. Learning and Psychodynamic Mechanisms 309 In Fig. 7 I have diagrammed the behavior space for the
case of the small boy whose father is very strict about
"punctuality." Look first at
the symbols for the self at the left and right of the diagram and drawn in
dashed lines. They are selves toward which the behaving self may
locomote. They are shown in regions of
the behavior space which ego perceives as potential goal regions. This introduction of self symbols into goal
regions is to take care of the fact that, at least for human actors, goal
regions and subgoal regions may contain not only other objects but also the actor's own self
as potentially present and as acting upon, or being acted upon by, the other
objects.2 The valence
charges deposited on these goal regions will, that is, often be attached
primarily not to the other objects per se but rather to the self, as acting on
the other objects or being acted upon by them.
Thus, in this particular example, it is the self in the region of
punctuality and therefore receiving love and approval from the mother (and
presumably also from the father) which has the positive valence. The self in the region of being late,
because continuing to play and therefore receiving disapproval and having love
withheld, has the negative valence. It is
important to emphasize that the region of approval from the father and/or
mother, the region of "punctuality," is also a region in which the
father himself is said to be located.
This highlights what is probably a crucial empirical problem in
connection with the establishment of successful identification. If a father and mother preach punctuality but the father is
himself unpunctual (i.e., is himself not
in the region of punctuality), the boy's identification with the father is
going to be relatively difficult. The boy's behaving self is pushed toward the region of the
mother's love. He is told and he
believes that to gain love he must be like
the father. And he is also told to be punctual. But he discovers that the father is not
himself punctual (i.e., is not in the region of punctuality). How, then, is the small boy to gain the
mother's love: by obeying the mother's and father's prescription and locomoting
toward the region of punctuality, which lies in one direction; or by becoming
like the father and locomoting in an opposite direction toward the region of unpunctuality where the father actually is? Furthermore,
it is to be observed that, in locomoting toward the
region of punctuality (assuming that it is not only prescribed
but that the father himself is actually in it) the behaving self is at the same
time locomoting away from a region which has
a positive valence - the region of unpunctuality or
"continuing to play."
The behaving self of the small boy in locomoting toward the region of
punctuality and identification is locomoting against a behavior in the opposite
direction. That is, as was pointed out
above, identification always involves the not-going to some other region or
regions which in themselves have positive valence. 2
“Objects" includes, of course, "other persons. 310 A Psychological Model We have stated the essential features of identification, as
described in terms of the behavior space, to be (1)
the locomoting toward a region of likeness to another or
others in order to gain love or approval, & (2)
the simultaneous locomoting away from some opposite region
or regions which have their own positive valences.
Whether or not good identification will be achieved will thus depend on
the relative magnitudes of a number of factors, such as the clearness and
unambiguity with which a single clear direction of locomotion to reach love and
approval can be perceived, the strength of the need-push for such love and
approval, and finally, the strengths of the opposing need-push or need-pushes.3 Finally,
when complete identification has been achieved either with another individual
or with a group,4 ego perceives not only his "self" as a goal but
also his behaving self as practically always in the region of the approved
behaviors of
the individual (or the group) with which he has identified. 8/23/01 jjd:
note this is life in the personality (should I say the flesh)
and not according to any adult social participation in the group. Behaviors which are approved and exhibited by the alter or
alters have, in cases of strong identification, such powerful positive valences
that they tend to win out over all others. An actor with strong identification
may, in fact, come to sacrifice every other need, even
life itself, because his behaving self is so strongly attracted
by the positive valences of the "identification" region belonging
both to him and the alter or alters with whom he has identified. In Fig. 7 we depicted identification in terms of what happens in the behavior space. It is obvious, however, that identification
involves not only changes in the behavior space but also resulting changes in
the belief-value matrices.
When ego accedes to alter's prescriptions on specific occasions in order
to achieve love and approval, he not only perceives that love and approval as
lying in the region of a given set of behaviors, also exemplified by alter, but
he acquires an accompanying belief that these kinds
of behaviors are in general the way to get to love and approval. Many of the problems involved in successful
or unsuccessful identification
undoubtedly arise in connection with the nature of the generalization gradients
which ego develops with respect to such behaviors in his belief-value matrices. Fig. 8 shows a belief-value matrix in identification. The generalization dimension represented at
the left has arrayed along it types of activity which are to be
"punctually" dropped. The
generalization dimension depicted at
the right has arrayed along it mother, father, and other individuals from whom
acts of love and approval will, it is believed, result if such activities are
dropped. Finally, there is the further
belief that love and approval from mother, from father, and from others will
lead in decreasing degrees to grati- fication of the need love and approval. The exact character of the identification
achieved and represented in such a diagram will be indicated by what gets
placed where along each of these generalization dimensions and by the shape and
spread of the forked tail of each of the belief lassos. 3 It will be seen below that identification
may sometimes be aided through the simultaneous operation of the mechanism of
symbolic substitution which permits a "surreptitious" expression of a
seemingly abandoned need-push. 4 Space will not be taken to elaborate upon the
problems of identification with a group.
But such group identification seems to involve practically the same
principles as identification with a single individual. The actor wants love and approval from the
group, and to get them he has to behave in ways similar to those of the
group. In so doing he has to locomote away
from behaving in other ways which, because of other needs, also have for him
positive valences. Learning and Psychodynamic Mechanisms 311 Fig. 8
BELIEF-VALUE MATRIX IN IDENTIFICATION b. The
sell-ideal. What I am
here calling the self-ideal Freud
discussed under the two separate
heads of the superego and the ego-ideal.
The superego according to his
analysis consists, in my terms, of acquired negative values and valences for
those types of behavior in which one should not
engage; and the ego-ideal
consists of acquired positive values and valences for those types of behavior
in which one should engage. The two concepts are, however, I believe,
best conceived as but obverse sides of one and the
same process. The formation of a
self-ideal is the final establishment of
categorizations of the self (in belief-value matrix terms) and of perceptions of the self (in immediate
behavior.space terms) as itself an alter which responds with love and approval
to certain acts of the behaving self and also responds by withdrawing love and
exhibiting disapproval to certain other acts of the behaving self. Next,
it is important to note that the establishment of such a self-ideal seems to
grow out of identification. (This also
was assumed by Freud.) But in a
self-ideal mere identification as such has been gone beyond, in the sense that
ego now perceives the positive goal region no longer as one in which the self merely behaves in ways similar to and approved by a
"judging alter or alters" but as one in which the self behaves in
ways similar to and approved by a "judging self." The proper diagram for the behavior space in
the case of a self-ideal is shown by Fig. 9. 312 It
would appear, however, that in order for such a
judging, approving, or disapproving self to develop, there must also be, or
have developed, self-love. For,
only insofar as self-love, narcissism if you will, is present, will there be
gratification as a result of the self's act of loving and approving
itself. Only if the need from others
has developed into a tertiary need for "love from
self" will positive values and valences be attached to
self-approval. Fig. 9
BEHAVIOR SPACE-- SELF-IDEAL A
complete diagram for the self-ideal is, therefore, represented in Fig. 10. This
diagram indicates that not only does a judging self become a particular perceived entity in particular
behavior-spaces but also that the self as a type of judge will be located along a
generalization dimension of judges (such as mother and father) in the belief-value matrix. Furthermore, the belief-value matrix is here
shown to correspond to a sort of transition stage between mere identification
and a true self-ideal. For it is
indicated that the self-love of ego is still affected by the approval of these
other judges. In the final stage of a self-ideal the self alone would be the sole and final judge. 313 If the
above analysis is correct, then the important causal problems will consist in
attempting to discover (1) what empirical conditions in early childhood favor
the development of strong identifications
and (2) what conditions in early childhood favor the development of a requisite degree of self-love so that the
individual will come to have as a goal not merely those types of behavior in
which the self will be approved and loved by others but also those types of
behavior in which the self will approve and love itself. Fig 10 NEED SYSTEM, MATRIX, AND BEHAVIOR SPACE FOR SELF-IDEAL 314 Further,
it would appear that, in extreme cases, a self-ideal may eventually accord
positive values to behavior which only the
self itself approves. Whereas
the behaviors for which the self will love and approve itself will be at first
primarily those which alters, with whom ego has identified, will love and
approve; the self, eventually, if self-love is strong enough, may come to love and approve behaviors which are totally
idiosyncratic. Hence, from a
practical, social-welfare point of view the fact that ego has acquired a strong
self-ideal may not necessarily mean that this self-ideal is, from the point of
view of the society, good. The self-ideal may be that of a criminal
or of an ego-maniac.
But, obviously, all sorts of detailed empirical investigations will be
necessary to discover just how socially nonacceptable, rather than socially
acceptable ego-ideals, are acquired. The problem may well involve
such factors as the strength of early identifications; the characters of those
identified with; the magnitude of self-love; and the particular stage in
development at which self-love developed. c. Repression. This mechanism may go off simultaneously with
instances of the others: thus, for example, in the above examples of
identification and of a self-ideal repression may also tend to be
involved. As was indicated before, the
behaving self in locomoting to a region of identification locomotes away from a
region of continuing to play and this latter region has a positive value of its
own But continuing
to play also has a negative valence corresponding to disapproval and
withholding of love. And it is
this phenomenon of a region which is simultaneously both positively valenced
and negatively valenced which, I shall assert, gives rise to repression.
Repression is a blotting out from conscious awareness of a given region or
regions of the behavior-space and a simultaneous blotting out of objects with
plus and minus values from belief-value matrices. As an
example of repression I shall choose a case which bulks large in the literature
- ego's repression of aggression againstt in-group members in order to retain
the love and approval of such members. Fig. 11 represents the diagram for this kind of
case. Look first at the behavior
space. The behaving self is represented as between two
regions. It may locomote toward a
region, at the right, of identification with in-group members - a region in
which the self and others in the in-group behave similarly and also approve the
self and each other (the acts of approving the others have not been indicated);
or it may locomote to a region, indicated at the left,
labeled "play." The term "play" here represents, for
purposes of illustration, any type of activity disapproved by the in-group. Further, it
has been indicated that disapproval by in-group members is to be represented as a barrier in the behavior space on the way to
play. Again, when any such barrier appears in the behavior space, aggression against that barrier will have positive
valence. The total region, at the left, contains, then, both a positive valence
on play and a positive valence on aggression against the barrier which is set
up by the in-group members. It also
contains two negative valences: one attached to disapproval for indulging in
play and one attached to disapproval for indulging in aggression. But it is my hypothesis that a region which has both positive and negative valences
will tend to be blotted out from the conscious awareness in the behavior
space. This blotting out I have represented by stippling the region. 315 The
next point to note is that such a perceptual blotting out, I believe, acts back
(see dashed arrows) upon the corresponding belief-value matrices activated by
the play and the aggression needs so that there is also a blotting out of the
beliefs and resultant values for play and aggression in these gov- erning matrices. The dashed arrows are to be conceived as
transmitting the blotting out (represented by stippling) from the behavior
space back to the belief-value matrices. Fiq. 11
A CASE OF REPRESSION In other words, the actor in
this situation neither consciously perceives nor
consciously believes that he wants to play or that he wants to aggress
against in-group members for their blocking of such play. Nor does he
consciously perceive or believe that he would be disapproved or punished
for these behaviors. His behaving self is, however, still left
with the need-pushes to locomote toward play and toward aggression as well as
with the need-push to locomote toward approved in-group behavior. This leaves the behaving self in a restless
condition. The two unreduced need-pushes will keep the behaving self restlessly locomoting. In fact, it is such repression - that is,
such unresolved need-pushes - which, I believe, are the basis of much neurotic and maladaptive
behavior. An apparent solution for such
a neurotic state is, it appears, often brought about through the cooperation of
still another mechanism - one which I would label symbolic substitution. To this we may now turn. d. Symbolism and symbolic substitution. In order to understand symbolic
substitution, we must first consider simple psychoanalytical symbolism (in
which substitution in the sense employed here need not as such be
involved). For such a case of simple
symbolism let us return to the example of restaurants and food. It will be
recalled that the value depicted on a type of restaurant in the belief-value
matrix was shown to depend, not only on the presence of a belief as to what
types of food the given type of restaurant will yield, but also upon the belief as to how much money will have to
be spent in such a type of restaurant. The effective positive value of the type of restaurant may
be much reduced if the given actor has a high value for controlled spending and
the given kind of restaurant costs much money.
fig 12:
Controlled spending: a
psychoanalytical symbol for controlled elimination A common psychoanalytical
suggestion as to why controlled spending should have a high positive
value for a given individual is that money is for
this individual a symbol for feces. Such an assumption (represented in Fig.
12) would be that the actor in question retains in his matrix the belief
that controlled elimination is the way to get to mother's approval, which latter is cathected to gratification of his
need for approval.5 The
crucial feature in this figure is the fact that it places money, the supposed
symbol for feces, along the same generalization dimension as the feces
themselves. That is, if controlled elimination of feces is believed to be
good, then this belief would be assumed to generalize
to the controlled spending of money.
On such an assumption, a type of object and accompanying behavior are
included along the generalization gradient of a belief, simply because they are
similar to another type of object
and its accompanying behavior which are already believed to lead to the given
type of goal. It is a case of
"overgeneralization." The
symbolic object and behavior have come to be "believed in" as ways to
reach the goal in question, not because they are really actually appropriate
for so doing, but merely because they are similar
to an original object and behavior which are, or were, appropriate. But we
shall see now that such psychoanalytical symbolism may be used in a further way
to permit symbolic substitution in cases
of repression. In the case of
controlled spending as a symbol of controlled elimination, the controlled elimination itself is presumably not
repressed. It also continues to
occur. Hence the controlled spending does not take the place of -
is not a substitute for - controlled elimination; the case, as we have said, is
merely one of an inept overgeneralization. In our
previous case of the repression of aggression
against in-group members, it will
appear now that aggression against the self and/or against out-group members may enter the picture as
"symbolic substitutes" for aggression against the in-group. This
possibility is represented in Fig. 13. Look first at the belief-value matrix on the
right. It will be seen that out-group
members and the self have both become placed along the same generalization
dimension as in-group members, because they are in many ways similar to, and
hence may act as symbols of, in-group members.
In other words, the original belief that by aggressing against in-group
members ego will satisfy both his need play and his need aggression has been
generalized to out-group members and to self so that ego now believes that
aggression against out-group members, or against self, will also gratify need
aggression and need play. The first
of these resultant beliefs is veridical in the sense that such aggression
against out-groups or self will actually gratify
aggression. The other belief,
that by aggression ego will gratify the play need, is
nonveridical. Moreover, because
it corresponds to a region in the behavior space which is blotted out from
conscious awareness, it will also be blotted out in the belief-value matrix. ======================================================== 6
Whether or not this is a likely assumption must he left to the
psychoanalysts. But, if this seems to
the reader not too probable an example of symbolism, there are other examples
that could he cited which the reader undoubtedly would accept and the same
principles would hold for them. 318 A Psychological Model Fiq. 13 SYMBOLIC SUBSTITUTION Look
next at the behavior space. The regions
for aggression against out-group members or against the self, since they have
no negative valences resulting from disapproval, will not be blotted out and
hence the behaving self will tend to locomote toward one or the other. In fact, the in-group may approve aggressions against the out-group or even, in some cultures,
against the self. Hence these regions
may have even stronger positive valences added to them. As a result, the behaving self will locomote
toward aggression against one or the other, and the need-push aggression will
be reduced. But the behaving self will
not locomote to the original play region and the need-push for play will
remain unreduced. Hence such a substitute symbolic aggression,
while adaptive in allowing release of aggression, is not successful in getting
ego to the original positive goal. Furthermore, given the conditions of the
actual world, it may lead ego to very bad subsequent situations such as retaliation from
out-group members, or extreme sickness or injury in the cases in which
he aggresses against himself. 319 4 Further Problems
Connected with the Model We
may turn now to several matters, briefly touched upon in preceding sections,
but which need further elaboration.
Specifically, let us consider in more detail: (1) the need system -
especially the problem of tertiary needs and of functional autonomy; (2)
further problems concerning matrices and their effects on the behavior space;
(3) the problem of the discourse use of symbols; and (4) the problem of
operational definitions. THE NEED
SYSTEM As has
already been indicated, the need system is
to be thought of as a set of interconnecting compartments, each
compartment corresponding to a differentiated need. The energy or tension in such compartments is conceived to be
made up of positive and negative charges analogous to electromagnetic charges. These
charges are supposed to be capable of spreading from one compartment to another
through the dividing walls, which walls may be thought of as in the nature of
semipermeable membranes. In considering the process of arousal of any specific need
the following principles should now also he included: (1)
Any independent physiological drive condition or stimulus
which arouses a need will be assumed to do so by first
increasing the total amount of charges in the libido compartment. (This compartment, it will be remembered, is
conceived to be in contact with each of the specific need compartments.) (2)
The arousal of a specific need will also be assumed to
involve an increased permeability of the membrane dividing such a specific need
compartment from the libido compartment so that there will be an increased flow of charges from the libido into the
specific need. (3)
It will be assumed, further, that the arousal of a specific
need may also cause an increase in the permeability
of the membranes dividing this specific-need compartment from certain other
specific-need compartments.1
(4)
Still further, such increased permeabilities of the
membranes between pairs of specific compartments must be assumed to
be either bidirectional or unidirectional. Thus, (a) if the permeability is bidirectional,
an increase of charges, resulting from arousal in either one of a pair of two
specific-need compartments will tend to flow into the other. That is, an increase in the magnitude of
either need will tend to enhance the magnitude of the other. (b) If the increased permeability of the membrane between a
pair of needs is unidirectional, it may mean
either (i)
that, when need A is aroused, need B will also tend to be
aroused because of an increase in permeability in
the direction from A to B, or (ii)
that if need A is aroused, need B will tend to be lessened
because of an increase in permeability in the direction B to A. 1
It 'was suggested previously that in order to indicate diagrammatically such
interactions between needs a hyperspace diagram would he necessary, so that
each specific-need compartment could he depicted not only as in contact with
the libido compartment but also in Contact with each and every other need
compartment. 320 A Psychological Model On the
basis of the above assumptions I would attempt to integrate and to explain such
already fairly well established empirical findings as the following: First,
the arousal of any specific need tends to be
correlated with an observable increase in general energy. Example: a hungry rat
also tends to be a lively and an active rat.
(Explained by principles 1 and 2 above to the effect that, in the
arousal of any specific need, charges first flow
into the general libido compartment and then into the
specific-need compartment.) Second,
if either one of certain pairs of needs is
aroused, this will tend to increase the magnitude of the other need of the
pair. Example: an increase in need
dominance tends to give rise to an increase in need aggression; conversely, an
increase in need aggression tends to give rise to an increase in need dominance.
(Explained by principle 4a of a bidirectional increase in permeability
of the membrane between the need compartments A and B.) Third,
when one of a given pair of needs is
aroused, it may tend to increase the magnitude of
the other need of the pair, but not vice versa. Example: if a basic viscerogenic need such
as hunger is aroused, it will tend to increase the magnitude of, say, the
aggression need, but not vice versa.
(Explained by principle 4b (i) of a unidirectional increase in
permeability from the hunger compartment into the aggression compartment, but
not vice versa.) Fourth,
when one of a given pair of needs is aroused
it may tend to decrease the magnitude of the
other need of the pair, but not vice versa.
Example: if a basic viscerogenic need, such as hunger, is aroused it
will tend to decrease the magnitude of, say, the aesthetic need, but not vice
versa. (Explained by principle 4b (ii)
above of a unidirectional increase in permeability into the hunger compartment
from the aesthetic compartment, but not vice versa.) It may
further be noted that with the aid of the above principles Maslow's striking
and important hypothesis of a hierarchy of needs [16] could be explained. Maslow assumes that needs are to be arranged
along some sort of continuum so that only when the basic "lower"
needs are in a state reasonable gratification can the "higher" ones
develop any appreciable magnitudes. The
viscerogenic hungers would be the outstanding basic needs lying at the bottom
of the continuum whereas the intellectual, aesthetic, and the more "disinterested" types of social needs
would lie at the top of the continuum.
This assumption, translated into the above terms, would mean that when
any need higher in the scale is aroused at the same time a "lower"
need is also aroused, the permeabilities between the two compartments would be
unidirectional in the sense that the arousal of the lower of the two needs
would always tend to drain away the charges from the higher need. Only when the lower need is gratified would
the higher need have a chance. Further Problems 321 Another
question about the need system which we have thus far avoided concerns a basic list of needs. I shall not attempt any final and precise
answer to this question. I shall,
however, assume, first, that there is
a list of primary needs - i.e., a set
of basic hungers and avoidances which man shares with his nearest kin the
anthropoid apes - such needs as, say, hunger, thirst, sex, pain avoidance,
aggression against obstacles, and a general exploratory, curiosity, or placing
need. The final and precise statement
of this list is yet to be agreed upon. I shall assume, secondly,
that in addition there is a list of not as yet clearly differentiated secondary or socio-relational needs, such as
affiliation (need for love and approval), dominance, dependence, submission,
and the like.2 I shall also assume that there is
probably in addition some set of tertiary
needs (which must definitely be assumed to be the product of learning) which
are fairly universal in any given population.
The tertiary needs will consist in wants to get to and from, to
manipulate (as ways of getting to or from) certain relatively universal types
of culturally provided goals; for example, in
our culture: the want to get to wealth and away from poverty;
the want to get to professional and business success and away from business
failure; the want to get to a college or university degree and away from
flunking out; the want to get to strolls in the park and away from the house or
office; the want to get to a vacation in Miami and away from work; the want to
play the violin; etc. It is, of
course, such an assumption of acquired "tertiary" needs for
culturally provided goals which is contained in Allport's doctrine of "functional autonomy" [1].
My point of view here, however, is perhaps somewhat different from
Aliport's. For I would grant, as
Allport perhaps would not, that a complete analysis of both the conscious and
the unconscious features of the given personalities might indicate that in
cases of seemingly tertiary needs, the goals of these tertiary needs
are not really final goals but mere
means or subgoals connected by “beliefs"
to more basic goals. However, these
connecting beliefs may have become so stable and unvarying in the given
individual that, for the purpose of a given empirical study, an independent
tertiary need can, pragmatically speaking, be assumed. The empirical criterion in such a case would
be that, whenever the individual (or individuals) in question can be said for
all practical purposes to persist
to or from a certain type of culture object (wealth, academic success, etc.) irrespective of any further consequences or lack of
consequences, then it is pragmatically useful and legitimate to
assume an independent tertiary need. 2
The question as to whether these secondary needs are themselves to he
considered innate, or as acquired early in life, must for the present he left
open I tend to believe that they are largely
innate and can, for example, be studied reliably in chimpanzees, who are
very like men but free from the teachings of any culture. 322 A Psychological Model But it
is to be remembered that this assumption (like the assumption of all the other
entities in the model) is no more than a pragmatic construct to be retained so
long as, and only so long as, it proves helpful. Thus, we would be led into no dire consequences if we first
assumed a separate, functionally autonomous, tertiary need which later, as a result of
further empirical investigation, turned out to be better conceived as a mere
means activity connected by the belief system to the goals of some more basic
needs. Suppose, for example, that from
studies of college sophomores we were first led to
postulate a tertiary need (stronger in some individuals than in others)
for "approval from college authorities." Suppose, that is, that one finds in the
general area under investigation that such an assumed tertiary need has
explanatory value. By this I would mean that for the students in question,
"approval from college authorities" appears to be sought irrespective
of any further consequences and enjoyed in
consummatory fashion per se, and that ratings of the
different intensities of this need in the different individual students was
found to correlate sensibly with some other
variable or variables. These initial
conclusions would not be
materially upset - although they might be further refined and amplified (i.e.,
some previous unexplained residual variance might be taken care of) - if, upon
further investigation, it turned out that the goal "approval from
authority" was more usefully conceived, not as an independent goal with
its own final consummatory response, but as a subgoal
connected by beliefs to some more ultimate goal such as that of being loved. If this proved to be the case, no
basic feature of the previous findings would be upset. It would simply appear that by now examining
variations in the more basic love need and in the
connecting beliefs, some of the preyiously unexplained variance could be
explained. The final conclusive test,
however, in terms of our model, as to whether in the last analysis a tertiary
need is or is not to be assumed will lie in the degree to which useful
deductions can or cannot be made from the assumption of such a special need compartment connected to other needs by membranes of
certain types of permeability. Further Problems 323 Finally,
if true tertiary needs really become formed, as they would according to the
doctrine of functional autonomy, we would still be left with the empirical
problem of establishing the laws and conditions under which the initial development of these functionally
autonomous new needs takes place.
This is a question which to date seems to have been but incompletely
experimented upon. Is it, for example, the frequency and consistency with which
a given type of means or subgoal has led to a given primary goal which causes the former to assume goal character? Or is it perhaps a lack of consistency of
this sequence which (as has, in fact, been suggested by some experimenters)
causes the means or subgoal to take over a functionally autonomous goal
character and to be accompanied by the development of a new tertiary need of
its own? These are questions which
require further investigation. FURTHER PROBLEMS
CONCERNING MATRICES AND THEIR EFFECTS
UPON THE BEHAVIOR SPACE A first
point to be emphasized under this heading is a fact already suggested, but not
previously stressed, that the typed images for one
and the same kind of objects may
appear simultaneously in two or more matrices and the fact that
the values given these images in the different matrices will be different. That is, the images may be
ordered, or located, quite differently along the respective different
generalization dimensions of the different matrices. Thus, for example, to return to our perhaps overworked
example of foods and restaurants, it would appear that a given set of food
images may be classified by the actor not only along the functionally defined
generalization dimension of leading-to-hunger
gratification, but also in a second matrix, along a functionally defined
dimension of leading-to-palatability gratification.5 This second matrix and the palatability need to which it is
connected plays, no doubt, a dominant part in most middle-class, high-nutrition
groups, whereas in a low-nutrition group the taste or palatability need may be
conceived as drained through the increased permeability of the membrane into a
strongly aroused hunger need. Or,
again, restaurants may be arranged by the actor not only along the functional
dimension of food-providing but also along
such different functional dimensions as costing-money
and prestige-providing. Fig. 14 indicates a possible diagrammatic way in
which such further complications of two or more matrices may be represented; it
represents the orderings of three types of restaurants, X, Y, and Z, in three
different matrices. The three
matrices are conceived to be connected respectively
to the three needs: (1)
hunger and/or palatability need, here grouped together as
simply a "good food" need; (2)
a not-wanting-to-spend need; and (3)
a prestige need. According, then, to this hypothetical
example, X, Y, and Z are believed by the actor to be in the order named relative to providing
"good food." As regards
spending, X is rated highly expensive, Y and Z as less
expensive. Finally, in regard to
prestige-producing values, X is believed by the actor to be very high in leading to prestige
gratification; whereas Y is believed to have low prestige~producing value, and
Z, although cheap, is believed to have more prestige~producing value (for
example, an artist's hangout) than Y, but still not as much prestige value as
X. 3
Sensations of palatability can he objectively - behavioristically - defined
without resort to phenomenology by means of discrimination experiments. And so defined I have labeled them "discriminanda" [22]. 324 Psychological
Model Fiq.14
THE SAME TYPES OF OBJECTS (TYPES OF RESTAURANTS) ORDERED DIFFERENTLY
IN DIFFERENT MATRICES That
one of the three types of restaurant will be gone to in a particular case (if
concrete instances of all three types are present in the environment and can be
got at with equal ease) will be determined by some
interactive effect on the behavior space of the values arising out of
the three sets of positive and negative values. In the present simple example, it is probably accurate enough to
assume that the interactive effects of these values will
be additive and this will make a restaurant of the type X the most
valenced and most likely to be locomoted to by the behaving self. 325 Fig. 15
ACADEMIC OR VISCEROGENIC
GRATIFICATION? A second
point in connection with matrices which deserves further elucidation is whether
a given type of object is to be conceived as the final goal object in a special
matrix attached directly to a special tertiary need of its own, or whether it
is to be conceived as merely a means object on the way to some more distant
goal of a more basic need. The two
alternatives are represented in Fig. 15, based
on the example of working for academic degrees. In Fig. 15, A represents the assumption of a small
belief-value matrix in which academic degrees are the finally cathected goals,
and it is indicated that such a matrix is attached directly to a new tertiary,
functionally autonomous, "need for academic
success." On this assumption the absolute
strength of the final gratification and deprivation values of such a matrix
will be determined directly by the strength of the directly attached academic
success need. That is, in diagram A it is assumed that there is an acquired
need for academic success which, insofar as it is aroused, will determine
directly the absolute magnitude of the gratification charge of this matrix,
which in turn (given the generalizations and beliefs in the matrix) will
determine the respective values for possessing professional degrees of Ph.D. or
M.A. and for studying books as the way of obtaining such degrees. In diagram B
of Fig. 15, on the other hand, I have presented
the alternative possibility: namely, that possession of the Ph.D. and the M.A.
does not in itself gratify a final need for academic success but is merely a
means on the route to the goals of some more basic needs such as, say, the
viscerogenic hungers. 325
Psychological Model Fig. 16
STUDYING TO OBTAIN LOVE AND APPROVAL But
perhaps a still more reasonable assumption would be that the obtaining of
academic degrees is a means not so much to the gratification of the basic
viscerogenic needs as to the gratification of the need for love and approval
from an alter or alters. Fig. 16 suggests this possibility. Here the resulting values on the Ph.D. and
M.A. degrees will be determined not by the viscerogenic needs but by the need
for social love and approval. The
significant new feature in Fig. 16 is that the
final valued goal objects are the perform- ances - overt or implicit acts of approval - represented by the corkscrew arrow issuing
from the image of an alter or alters.
It is these performances of the imaged alter, the "enjoyment" of
which is believed by ego to lead to the gratification of the need for social
love and approval.4 4
These assumptions of a social love and approval need and of the goal as being
the "approval response" of the alter or alters or even of a judging
self have, of course, already been made above in our discussion of the
development of identification and of the self-ideal. Further Problems 327 Fiq
17 BEHAVIOR
SPACE -- ESTIMATION OF SIZE A third
example of a more complicated problem in connection with matrices is to be
found in the well known Pruner and Goodman experiment [2]. In this experiment a group of children (who
were separated into two subgroups of "poor children" and "rich
children") were presented with coins: a penny, a nickel, a dime, a quarter,
and a half-dollar. The children were
asked to match the sizes of these coins by adjusting a diaphragm with a knob so
as to make a patch of light shining through a ground glass screen appear the
same size as the given coin. The
children were also given the task of matching small circles of cardboard - one
corresponding in size to each coin - in similar fashion. The results were briefly as follows: (1)
In general, all children tended to overestimate the sizes of
the coins as compared with the sizes of the corresponding equal-sized pieces of
cardboard. (2)
This over-estimation of the coins was greater, the greater
the money value of the coins (the dime was for some reason an exception). (3)
The poor children tended to overestimate the coins more than
did the rich children. 327 A Psychological Model Let me first
indicate how such an act of estimating size
is, as I see it, to be depicted in a behavior-space diagram (Fig. 17). In
this figure, x is the perceived
cardboard or coin to be matched. The
act of adjusting the diaphragm until it appears equal to this object x is conceived as the "locomotion"
of "placing" this object in
the correct "size direction."
The angular directions around the "nose" of the behaving self
here represent thus not spatial directions but "size
directions." The average size of
the given set of cardboards, or coins, is represented as straight ahead of the behaving
self's nose. Sizes larger than the
average are represented as radiating above and sizes smaller than the average
as radiating below the behaving self's nose.
The extreme limits of the given range of actual sizes used in the given
set of objects would fall somewhere near the right angle upwards and somewhere
near the right angle downwards, respectively.
The locomotion consists of "pushing" x in the correct direction and reaching the
region of "having matched." Fiq. 18
BELIEF-VALUE MATRIX-- ESTIMATION OF SIZE The next
point to be noted is that in this size-matching task the positive valence to be
reached by the act of matching will be the same irrespective of whether the
given size is large or small. In other
words, the positive values in the controlling
matrix are to be conceived as equal for all sizes. To indicate just what
is meant by this let me present now the diagram for the accompanying matrix (Fig. 18). It
will be observed that in this matrix the different sizes are arranged along one
generalization dimension. The belief as
to the result of an "act" of "placing" (indicated by the
corkscrew arrow) is assumed to have a flat generalization gradient in the sense
that in such a matrix the resultant instrumental value from placing any one of
these sizes will be the same. In other words, the same degree of gratification
will, it is believed by ego, be received whatever the size of the particular
object that is placed. Further Problems 329 Finally,
one further query about such a matrix. Is such a matrix to be conceived as attached directly to the independent curiosity,
exploratory, or placing need? Or is it
to be conceived, rather, as but a submatrix within
a larger matrix, of which the ultimate goal is, to take the present
example, to get "love" and "praise" from the
experimenter? Personally, I am inclined
to think that both types of causation were probably involved in this particular
experiment, which used young children.
However, I should certainly hold that with adults, and often also even with children or rats,
a general exploratory, curiosity, or placing need
has also to be postulated (see above).
Further, this need, as already indicated, is assumed to be in contact
with all the other needs, so that whenever a special need is aroused some of
the tension from this special need will tend to flow into and arouse in some
degree the pure exploratory or ''cognitive placing'' need. Hence even in this case the need of the children to get love and praise from the
experimenter would pass over into and activate in some degree the
pure exploratory or cognitive placing need.
But let
us turn now to the purchasing power of the
coins. For it would appear that a coin
may also be locomoted toward for its purchasing power and that the valence for
each coin will be greater, the greater its purchasing power. Fig. 19 presents
a diagram which indicates this. There
are a number of new assumptions introduced here. (1)
It is assumed that the purchasing-power matrix and the
cognitive matrix both produce an array of
directions in the behavior space. (2)
It is assumed further that these two sets of directions are
"orthogonal" to one another. This has been indicated by centering the
array of purchasing-power directions around a northerly axis and the array of
size directions around an easterly axis.
(3)
It is assumed that the attempted locomotion of
"placing" of the object according to size
is distorted by a simultaneous locomotion of placing according
to purchasing power. (4)
It is assumed that this distortion in the direction of
purchasing power is greater the greater the valence of the given coin, and also
the poorer the children (i.e., the stronger the average valences for all the
coins). In a word, the behaving self of
the child locomotes in such a direction that the path which it takes, though
supposedly a size-placing path, is in reality a compromise between size placing
and going toward the valence of the coin because of its money value. 330 A Psychological Model Although
the argument is complicated, the essential point is, I believe, clear: namely,
that many "distortions" of
perception may be conceived as distortions of discriminated
directions of locomotion in the behavior space; and further that such
distortions of direction, when they occur, may be a result of the interference in the behavior space of a
categorization of directions coming from one matrix upon the categorization of
directions coming from another matrix. Fig. 19
ESTIMATION OF SIZE AS AFFECTED BY PURCHASING POWER Let us
turn now to a consideration of still another problem concerning matrices - the
use of language, or discourse, symbols. Further Problems 331 THE DISCOURSE USE
OF SYMBOLS AND SOCIAL RELATIONSHIP UNITS We have
considered the nature of symbols in the psychoanalytical sense. There the symbol object was conceived to get
placed along the same generalization dimension as the symbolized object and
therefore to be behaved toward in the same way in which the symbolized object
is behaved toward. But now
we must turn to symbols in the more ordinary, discourse or language, sense of
the word. It is obvious that the
"discourse" use of symbols is first of all a type of communicatory
behavior. Furthermore, communicatory
behavior obviously occurs only in the presence of an alter or alters and in the presence (but usually the "non-pointable to"
presence) of either some concrete object or of some conceptual object. Ego then behaves so as to indicate to an
alter or alters (or to himself in the temporary role of an alter) what the
character of this concrete or conceptual object is. Fig. 20 BELIEF-VALUE
MATRIX-- COMMUNICATORY BEHAVIOR How is this use of symbols taken
care of by the model? Fig. 20 suggests a way in which it may be done. Note first that the entities placed along
any generalization dimension are no longer, as in previous examples, single objects (nonhuman or human) relative to
which ego has beliefs and values but are larger social-relationship objects containing the "typed" self,5 and a typed
alter, and other possible objects, labeled O. Further, these
larger social-relationship units are defined not merely by self and alter and
other objects but also by ego's differentiation of
the communicatory and other interrelationships between these
latter. Thus, in the social-relationship
unit depicted on generalization dimension I,
I have indicated the self as present to an alter and vice versa (double-headed
arrow) and an object O is
represented as present to self but not to alter
(single-headed arrow). The curved
corkscrew leading from O to alter
represents the type of symbolic response (e.g., "naming") by which
ego "believes" that the type of social-relationship unit depicted on
generalization dimension II will be achieved.
This relationship is depicted as one in which O
is now present to alter as well as to self.
The zigzag arrow indicates the type of dominance response by which ego
believes that (now that O is
present to alter as well as to self) self can dominate alter and the
social-relationship unit of the type depicted on generalization dimension III is to be achieved. And, I have depicted on generalization dimension III a social-relationship unit in which alter is submitting to self. And, finally, this is believed by ego to lead directly to the gratification of his, ego's, need for dominance. 5
The "self" as here depicted Ie ego's "universalized" image
of himself. 332 A Psychological Model The above use of symbols is a type
of behavior in which ego presents language symbols
to an alter in default of being able to present the actual symbolized
object. Often the object, which is in
ego's behavior space and belief-value matrix and which ego wishes to bring into
alter's behavior space and belief-value matrix, may be merely an ordinary
environmental, but distant, object as in the example just presented. But language may also perform a quite different
function. It may be used by ego to
designate either (a)
a conceptual object, a "universal," which cannot
actually be pointed to, or (b)
a "private" internal object which exists only in
ego's own belief-value matrix and behavior space. How shall we conceive these two further uses of language? (a) Our
explanation of the use of language symbols by ego to indicate a ''universal''
to alter (or to himself as alter) apparently
depends first on the circumstance that universals actually
constitute, according to the model, directions in the behavior space. When an actor matches a given gray in a
series of grays or a given sized object in a series of objects of varying
sizes, he may do so by having an array of concrete gray objects or an array of
concrete objects of different sizes as samples actually before him, and he may
merely place the object to be matched on top of the most nearly
similar sample in such an array. This is a simple response to the likeness of the
two concrete objects. The actor may go a step further and simply arrange a
series of test objects according to "grayness" or according to
"size" with no chart or exemplar
before him. He is then matching each
object not to a concrete sample but to the
"conceptuahzed" direction in the behavior space. All this can be done by an ego without the
use of speech. When, however, the ego
has learned the use of speech, he has learned a symbol which he
can use to indicate a given variety of gray (or a given size), i.e., a given
direction. In making use of this
symbol, emitting this word, he is performing a new sort of
locomotion in his behavior field.
Instead of the behaving self merely placing the given object in the
proper direction, the behaving self is also "conveying"
this direction to an alter (or to self as alter). Further Problems 333 It appears, however, that there is
a still more recondite use of Ianguage by human beings. They may use language not merely to convey objects and
directions in their behavior space and the categorizations on their
generalization dimensions, but they can also use language to convey their "con- veyings." They can say not only: "This is yellow," "This is
large," and so on; but they can also say:
"I perceive this as yellow," or, "I believe
this to be yellow." In other words, they can introspect or use verbal
reports. They can talk about their
behavior spaces or about their belief-value matrices. And this is a very crucial and important phenomenon. It means that the verbal reports
of the actor can often be used in lieu of many complicated observations of
actual behavior to discover both what an actor perceives and valences and what
he conceives and values. Another
crucial feature which, however, must be kept in mind in connection with such
use of verbal reports is that, as has already been noted in our discussion of
repression, significant parts of the behavior field or of the belief-value
matrix may not be accompanied by conscious
awareness may not, that is,
be available for such verbal report.6 An
actual working out in terms of our model of the exact ways in which the causal
determination of such verbal reports is to be conceived will not be attempted
here. Suffice it to say that we do
assume that by virtue of such "verbal reports" the constitution of an
actor's belief-value matrix or of his behavior space can be to some degree
correctly conveyed to another. Only,
however, after the details of our assumptions concerning this process have
actually been worked out in model terms, shall we be in a really good position to understand, better than we do now, the conditions under
which such verbal reports (introspection) are or
are not reliable indicators of the actual content of a matrix or
of a behavior space. By an operational definition of an intervening variable
I shall mean, first, a statement about a standard
defining experiment in which a certain measurable variation in some
feature of the observed behavior will, by definition, be assumed to be a direct
measure of corresponding variation in the magnitudes of a given intervening variable. Second, such a definition will involve an assumption about the linear or nonlinear nature of this
mathematical function connecting the measured feature of the dependent behavior
to the intervening variable. And, third, the
specific constants in this form of mathematical function must also be
known, or assumed, before such definitions will be final. And it must be admitted that at present our
notions concerning such defining experiments and the precise forms of the
mathematical functions connecting the intervening variables to the dependent
behavior fall far short of the required precision. Our attempts, now, can be only
partial and tentative. This does not,
however, I believe, prevent our model from having considerable theoretical and
empirical usefulness, even at this time.
For the model and the variables within it set a program for further
empirical investigation, and prominent among such suggested new investigations
would be the proposed defining experiments themselves. Furthermore, as the data from such
experiments come in, not only will our store of empirical knowledge be increase
but the model itself will thereby suffer corrections and revisions, or perhaps
even final abandonment, and a new and better one will be arrived at. In the
meantime, the model will have served its purpose in stimulating research an in
leading to the discovery of ever more empirical facts. 6
Conscious awareness means, I
believe, in the lost analysis, availability for verbal report. 334 A Psychological Model Let us
turn now to an indication of at least the sorts of defining experiments which
can at this time be suggested for most of the intervening variables. To keep the task clearly before us, it will
help to group the variables in the following outline.7 VARIABLES TO BE OPERATIONALLY DEFINED a. List of
needs. b. Magnitude
of a given need at a given moment. c. The
magnitude of the gratification and deprivation "values" of the given
matrix at a given time. d. The
shape of the cathexis belief attaching the various types of goal object
(arrayed along a given generalization dimension) to the gratification end of a
given matrix. e. The
shape of the means-end belief attaching types of means object (arrayed along a
given generalization dimension) to a given type of goal object or subgoal
object. f. Perceived
qualities, distances, and directions in the behavior space. g. The
strength of a need-push. h. The
strength of a positive or negative valence. i. The
strength and direction of a field force. j. The
identification of a locomotion. 7
It will he noted that I have omitted from the outline the individual difference
variables, because, as indicated above, the main technique by which such
variables have been arrived at to date - that of factor analysis - fails to
integrate them in any way with such a "content" scheme as that of our
model. Eventually, the manner in which individual difference
variables will have to be handled to fit into our type of scheme will be in
terms of individual differences in the magnitudes of the constants or
parameters in the mathematical functions which are assumed to connect the
intervening variables with each other and with the independent variables on the
one hand, and with the dependent behavior on the other. But the task of developing the final forms
of these equations and of measuring the variations in their constants and
parameters for different individual lies. as has been suggested, a long way
ahead. Further Problems 335 This
is a formidable array of items. All that can be attempted here will be very
brief, schematic, and tentative suggestions as to the types of experiments or
other empirical set-ups to be used for defining certain of these items. a. List of needs. As has already been indicated, a need is to
be defined as a readiness or tendency to
persist toward and to perform a consummatory respouse relative to a
certain more or less arbitrarily chosen "standard" goal object or
situation and to avoid or go away from certain other objects or situations. Once the consummatory response to the standard goal
object has been achieved, the given tendency to persist toward, and away
from, ceases. Using this
definition we would postulate, then, as many needs in a given actor as we can
find standard objects or situations for which such tendencies to and from can
be demonstrated. At
the present time psychologists seem to be fairly well agreed, in considering
adult human beings, upon a list of common viscerogenic hungers, such as food
hunger, palatability hunger, thirst, sex, temperature control, oxygen intake,
rest and sleep, etc. - each of these to be defined and measured in terms of the strength of the propensities to go toward a
"standard" food, a "standard" taste, a "standard"
liquid, a "standard" sex object, a "standard" temperature,
a "standard" degree of oxygen intake, a "standard" object
to rest and sleep on. In such experiments, however, it would be necessary not
only to undertake the measure of the strength of the given need by employing a well-chosen "standard" goal object but
also by using a controlled, "standardized" test situation -
one very familiar to the given actor so that the measurement of the strength of
the given need will not be corrupted or diluted by any modifying simultaneous
influence from some other need. Thus,
for example, we would have to be sure in measuring the food-hunger need that
our measures were not distorted by some hidden and unrealized simultaneous activity of, say, the thirst need or the fear need. Second,
it would appear in the cases of human beings (and, in fact, in the cases of all
mammals) that there are also to be listed two other basic primary needs
in addition to the viscerogenic hungers.
These are fear and aggression. It is to be noted, however, that for both of
these needs the standardized defining situation or object chosen for measuring
will not be a standard variety of the to-be-got-to goal (i.e., safety in the
case of fear and destroyed opposition in the case of aggression) but rather a standard variety of the situation to-be-got-from, (i.e., a
standard type of pain or injury in the case of fear, and a standard type of blocking or obstruction
in the case of aggression) .8
8
We are using the term "aggression" somewhat differently from the way
in which it was used in the earlier sections of this book. 336 A Psychological Model Third,
we have suggested in the preceding discussion that in human beings, and even in
rats, there must also be assumed a general exploratory
and curiosity, or "cognitive placing," need. The standard defining goal object for this
need would be some sort of standard “new"
object to he explored, examined, or matched. It will
obviously require considerable experimentation and analysis to decide what to
use as a standard "new" object for defining and measuring the basic
strength of such a curiosity or exploratory drive. However, I believe that, in
time, a concrete and satisfactory answer to this problem can be found. Fourth,
I would argue that there is probably also in man and in the higher apes a list
of secondary innate social needs such
as need gregariousness, need love, need approval, need dominance, need
submission, and the like. The exact
list of these is still to he determined.
Again, whether or not we assume any one of these needs will be decided by whether we can
agree upon a defining social goal situation, which can he made standard and
used for measuring, for each of these needs.
Fifth,
it has already been suggested that acquired,
functionally autonomous, tertiary needs for culturally provided
goals may also become established in human beings. Insofar as this assumption stands up after further empirical
investigations, then we must also be able to set up defining standard goal
objects or situations for each of these tertiary needs. b. The magnitude
of a given need at a given moment. It has just been indicated that for the measurement of the
strength of a need at any time an agreed upon standard
goal object for that need and an agreed upon standardized
testing situation must be decided on.
And it appears, further, that a measure of some quantitative aspect of
the vigor of the behavior (e.g., frequency, intensity, force, consistency,
latency, or the like) toward or away from the standard goal object in the
standardized testing situation must also be agreed upon as constituting the
measure of the strength of the given need.
The
discovery of, and agreement upon, acceptable standard goal objects, acceptable
standardized testing situations, and acceptable to-be-measured features of the
behavior constitute the main task. The reaching of final agreement upon all these matters in
connection with, say, the social needs seems to be a long way in the
future. However, this has not prevented
personality and social psychologists from making beginnings and from trying out
quantitative measures (often ratings by an observer) of verbal or other
responses of given subjects in either real or projective situations as indications
of the strength of needs (see, for example, the original studies by Murray [17]
and the later ones by Frenkel-Brunswik [4]).
Further Problems 337 An
example of a relatively straightforward way of measuring the hunger and other
viscerogenic needs in rats is the Columbia
Obstruction Box [29]. By
this method the number of crossings of an electric grill in a twenty-minute
period to a standard food or other goal are counted and taken as a measure of the strength of the given need. As so measured, the strength of the hunger need, for example, is
found to increase progressively over some forty-eight hours of food
deprivation. After longer periods of
starvation, it declines. This I would
take as indicating that at first the hunger need increases more or less
linearly with the physiological hunger-drive condition, but with greater
increases in the hunger-drive condition, other physiological factors intervene
which lower the general libido need and because of this lowered libido the hunger need as such (not the hunger drive) declines. The electromagnetic charges in the
hunger-need compartment tend, after greater periods of starvation, to be
drained off into the libido compartment.
c. The magnitude
of the gratification and deprivation "values" of a matrix. My assumption will be that, whereas the
gratification values and deprivation values in any matrix vary proportionately
(linearly) with the magnitudes of the need to which they are attached, their
absolute magnitudes may be less than those of the total need. That is, any given need compartment may have
attached to it numerous different matrices and the proportion of the total need
strength which goes into a given matrix may be less than 1. For example, if the matrix aroused
by virtue of the presented stimulus situation be that of restaurants and foods
obtained in restaurants, the strength of the gratification and deprivation
values for this matrix may be less (given one and the same actual hunger need)
than the gratification and deprivation values of a matrix relative to
home-cooked foods. This means, of
course, that whereas the standardized testing situation for the pure hunger
need must be one in which neither the peculiarities of restaurants nor of homes
come into play (I am not sure what such a standardized situation might be), the
defining situation for measuring the specific gratification and deprivation
values of a given hunger matrix must be one which contains the specific types
of means objects as well as types of goal objects constitutive of the given
matrix. d. The shape of the cathexis belief attaching
the various types of goal object (arrayed along a given generalization
dimension) to the gratification end of a matrix. It must be pointed out first that, by definition, the most believed
in type of goal object placed on the final generalization dimension of a matrix
will be said to have a belief strength of 1.
Other goal objects arrayed along this same dimension will, according to
the shape of the initial, generalization end of the belief lasso be less
strongly "believed in." It is, then, the shape of this
generalization fork which determines the degree to which the given type of goal
object is believed to be good for reaching the given gratification end of a
matrix. The empirical question which
arises is how do we determine the shape of this forked tail. I suggest two empirical methods of attacking this problem.
The first
method is to arouse the given need in our experimental subject and to discover,
under the given conditions of need arousal, what will be the relative order of
actual going to and performing the consummatory response Upon the given selected array of goal
objects in question. We may, for example, make the actor hungry and measure
quantitatively how he actually selects from an array of different types of food
in terms of going to and eating. 338 A Psychological Model (This is the method of P. T. Young in his important
studies ~3O] on hunger and appetite in rats.)
Further, it is to he noted that it may turn out, when the test is made
separately for different strengths of the hunger need, that the order of
selected food types will he different for different strengths of the hunger
need (i.e., for different magnitudes of positive value charge in the
gratification end of the given matrix).
In other words, the shape of the generalization end of the belief lasso
may be different for different strengths of the hunger-gratification
value. This is an interesting and
important possibility which will need much further empirical study. Finally, it is to be noted that this method
of testing the shape of the generalization end of a belief lasso is the only method which we can use with the lower
animals. We have to make them
actually hungry, thirsty, sexually aroused, or whatever, and then measure some
quantitative feature of the relative order in which they will select from
presented alternative types of goal object.
There
is, however, a second method which can be
used with human beings. We can
use their verbal reports and we can obtain these stated preferences under
conditions in which the corresponding need is not aroused, or not appreciably
so. We can ask Mr. X what are his
preferences relative to a list of types of food. And it may well turn out that we may get much the same answer
whether or not his hunger need is aroused at the moment. That we can thus determine by verbal
responses the shape of a food-gratification belief lasso indicates again the
point made above; namely, that verbal reports of human beings must be assumed
to be capable, with some degree of accuracy, of betraying directly the features
of a belief-value matrix (however imperfect our conceptualization of the causal
determination of verbal reports still remains). e. The shape of
the means-end belief attaching types of means objects (arrayed along a given
generalization dimension) to a given type of goal object or subgoal object. Here, again, it would appear, in working
with rats or other subhuman animals, that the only way we can arrive at the
shape of a means-end belief - that is, the relative preferences with which
certain types of objects will be selected as means to certain other types of
objects as goals - will be through a long experimental procedure in which we
first establish a type of goal object and then test out preferences for types
of means objects over a long series of experiments in which many concrete
instances of the goal objects and the means objects are used. In the case of men, however, the approach may differ. We can ask them about their beliefs when
they are under conditions of no actual needs other than a
social-approval need which leads to the readiness to "convey"
information. For
example, we ask them, when they are not hungry, to rate restaurants of types X,
Y, and Z in terms of the degree to which they believe that these types of
restaurants will lead to a good steak.
And we shall probably find that their verbal ordering of these
restaurant types when they are not hungry is very much the same as the ordering
which would occur as a result of actual selection of restaurant types when they
were hungry. And this means that verbal
reports (with some validity) can indicate the characters of means-end beliefs
in a belief-value matrix, even though the utilitarian need in question is not
actually activated at the moment. Further Problems 339 Before leaving this matter of the operational
definition and measurement of cathexes and means-end beliefs, one more point,
which has not perhaps been sufficiently emphasized, should be brought out. In
the measuring of a belief relative to types of objects (if the measurement is
carried out in non-verbal fashion) the measuring
has to be done over a whole range of particular instances of each of the given types of object. If, for example, we are to conclude that a
given actor believes strongly that Child's restaurants are better than ABC restaurants for leading to good
flapjacks, we must try him out (or if possible a whole series of identical
"him’s" out) to see the consistency with which in Child's restaurants
he will order flapjacks and in ABC restaurants he will order not flapjacks but,
say, tea and buns. Similarly, in the case of rats, we
will be able to say that a given population of rats (after a given training)
has come to believe more strongly that right-hand alleys lead to food than that
left-hand alleys lead to food, only if we try these rats on many pairs of
right-hand and left-hand alleys. Only
in this way will we know that it is right-handedness versus left-handedness as
such, and not some peculiarity of a particular pair of alleys, that determines
the selection of one turn or the other.
But let us turn now to the problem of the
operational definition of features of the behavior space. f. Perceived qualities, distances,
and directions in the behavior space. In working with human subjects, the method usually resorted to
for attempting to discover and define an actor's perceived qualities,
distances, and directions is again verbal (introspective) reports. The subject says that he perceives
such-and-such objects, qualities, or performances, and that these objects are
placed in such-and-such ways relative to one another. As we have already seen, however, those features of the behavior
space of which the actor is “consciously aware" (i.e., which he can
verbally report) are not necessarily coextensive with the total behavior space
which actually governs his locomotions and hence his actual behavior. It appears, therefore, that sometimes we must resort to
other methods similar to those which we have to use with the lower
animals. The
methods required for the lower animals all reduce essentially to experiments in
which, by presenting the actor with successive pairs of stimuli (colors,
shapes, tones, spatial directions, spatial distances, etc.), we eventually
develop tables from which we can infer that if the given actor is presented with such-and-such actual physical qualities, distances,
or directions in the stimulus situation, he will (other things being equal) be
likely to perceive (contain in his behavior space of the moment) such-and-such
objects in such-and-such directions and distances. 340 A Psychological Model Thus from discrimination-box experiments with rats, we
can now say, for example, that a rat will perceive only shades of gray and not
chromatic colors; that he will perceive spatial length in the distances which
he runs over, only if one length is at least 10 per cent longer than the
other. Finally, as regards simple
spatial direction, we are merely beginning to be able to say something (see
experiments by Tolman, Ritchie, and Kalish [26~28]) about what sorts of
orienting extra maze cues are necessary to enable rats to recognize a spatial short cut to food or to be able to
choose the correct approach to a well-located food from some wholly new
starting point on the other side of the maze.
As for the problem of determining the probably perceived mechanical,
social, aesthetic, etc., directions and distances of human beings without
resort to verbal reports, hardly a start has been made. g. The strength
of a given need-push. I
shall assume, very dogmatically, that a need-push in a given behavior space is
measured directly by the strength of the deprivation value (or its equal the
strength of the gratification value) in the controlling belief-value
matrix. Thus, to return to our original
example of restaurants and foods, I shall assume that the strength
of the hunger need-push resulting from the activation of the
restaurant-plus-food matrix is to be measured by the average magnitude of the
eating or getting-to behaviors in consuming the most preferred restaurant food
(under the given strength of the hunger need).
h. The strength
of a positive or negative valence. The strength of the valence on the percept of, say, a given food
will be a function of the general need-push for food activated by the given
controlling matrix plus the degree of cathexis to the particular variety of
food as determined by the shape of the generalization fork. In other words, whereas the need-push and
the gratification value will be the same, and measured by putting the actor
before an instance of the most preferred food, the valence for any given food
may be less and can only be tested by putting the actor before an instance of
this particular variety of food.
Similarly, the strength of a negative valence will be measured by
putting the actor before an instance of the special variety of - in the case of
fear or aggression - a frightening or blocking object and measuring in some way
the vigor of the actor's avoidance behavior in the presence of such
object. i. The strength
and direction of a field force.
The strength of a field
force 9 is, I shall assume, directly proportional to the product of
the need-push and the determining valence in question and inversely
proportional to the square of the behavior-space distance between the region of
the behaving self at the moment and the region of the corresponding
valence. If, then, we know the
magnitude of the need-push and the magnitude of the valence and the magnitude
of the perceived distance (not the actual physical distance) between the
behaving self at the moment and the perceived location of the valence, we could
compute (if we had the proper constant or constants) the strength of the
resulting field force. Obviously,
however, any such computations will require more precise measurements than we
now have. 9 The terms
"field force" and "behavior force" are used
interchangeably. Further Problems 341 The direction of a behavior force is the
direction as perceived by the actor between the region of the behaving self and
the perceived region of a positively or negatively valenced distant object or
region. The problem of direction in the
behavior space is a complicated one. We
are at the barest beginfling, for example, in learning something about the
nature and precision of the behavior-space directions of rats when compared
with actual directions in physical space.
Furthermore, as has already been indicated, the concept of direction in behavior space has been broadened by the
present writer to cover every sort of discriminable differences - qualitative,
mechanical, aesthetic, and other differences as well as spatial ones. That is, insofar as two objects are close
together qualitatively, mechanically, aesthetically, or whatever, the more
nearly they will be in the same "direction." Discriminable differences become, in terms
of the behavior space, angular differences around the nose of the behaving
self. This whole
concept of direction is, in short, as yet only sketchily developed. And certainly a great deal more analysis
will have to be done to make it really workable and to allow for the
complications which arise when both directions and distances are combined. Only when all this has been done, will we be
able to handle such phenomena as short cuts and roundabouts not only with
regard to space but also with regard to quality relations, aesthetic relations,
mechanical relations, and the like. j. The identification
of a locomotion. In
considering the operational identification of a locomotion it must be
constantly emphasized that what is actually observed and measured is a
behavior, or some quantitative aspect of a behavior, and not a locomotion. A locomotion is a purely hypothetical
construct (an intervening variable). It is correlated with a behavior
but is not the behavior itself. To take
a simple example, assume that we observe a rat approaching a choice point in a
maze, turning left and right a number of times when he reaches this point, and
then finally proceeding down the one alley or the other. These would be behaviors. What, however, would be the correspondmg
locomotions of the rat's behaving self?
The answer must be that only after we have been able to identify the probable
behavior-space regions corresponding to the different objective features of the
total stimulus situation would we really be able to
identify the locomotions. A
region in behavior space, it will be remembered, is to he defined as a set of
potential behaviors, which the actor is ready to perform when in the presence
of a given stimulus situation. And a
locomotion was defined as a selection
of one or more such behaviors out of this set of possible ones as the way of passing to a next region. 342 A Psychological Model Only
when we finally know enough to be able to say that an actor at this point had
such-and-such behavior possibilities which he could release and that, out of
all these, this one was selected and got him to such-and-such another set of
possible behaviors, will we finally be able to identify his actual
locomotions. In the
case of a human actor we would, of course, again resort to verbal reports,
though with some uncertainty as to their ultimate reliability. We would ask him, "Where were you
first; that is, what objects (behavior possibilities) were then before
you? What behavior did you select and
what new objects (behavior possibilities) were you then brought into the
presence of?" And so on. After we had asked him these questions, we
would also want him to report on such further complicated interrelationships as
those concerning the roundabout routes and short cuts which he realized that he
could have taken to arrive at one and the same final region. Obviously, it will take many experiments to
pin all this down. Hence any final and
complete definition of a rat's "locomotion" or a man’s
'locomotion" still lies far in the future. To sum
up, let us look once again at this matter of operational definitions as a
whole. All such definitions reduce in
the last analysis to assertions that in the attempt to think of precise ways of
identifying and measuring each of the assumed intervening variables we shall be
led, should our theoretical model have value, not only to clearer ideas about
these hypothetical variables themselves but also to important empirical
experiments. These experiments,
whatever the truth of the model, will uncover new relationships between controlled,
and measured, variations in independent variables, on the one hand, and
resultant measured changes in final behavior, on the other. Having
now reviewed, as best we can, the parts and interrelations of our model, let us
finally, in the next chapter, make an attempt to show its usefulness and
applicability to some of the problems and concepts of the sociologist and the
anthropologist as presented in other parts of this book. 343 5 Value
Standards; Pattern Variables; Social Roles; Personality The four
concepts of value standards, pattern variables, social roles,
and personality have been selected for final consideration
because they have figured prominently elsewhere in the book, and perhaps, more
particularly, because they happen to be ones in which the present writer is
personally interested and which he believes can be helpfully analyzed in terms
of the model. Cultures
have value standards - cognitive, appreciative, moral. These standards tend to be acquired by the
actors living in these cultures. That
is, an actor tends to differentiate between the true and the false, the
beautiful and the ugly, the good and the bad, in ways prescribed by the
culture. He adopts his standards from
the culture. What now in terms of the
model are such standards and how are they acquired? It has
already been argued that among the list of relatively basic needs present in
both man and the lower animals is an exploratory or "cognitive
placing" need.1 This
need leads to the development of belief-value matrices and of behavior spaces
in which gratification is obtained from the mere piecing of objects according
to their qualities, directions, and distances.
It will now be suggested that cognitive - or "understanding" -
"value stand-ards" are no more than such "placing matrices"
as regards qualities and locations,2
which the given culture tends to inculcate in its individual members. 1
See the discnssion above of latent learning and of the Pruner and Goodman
expriment. . 2
The term "location" is here used to cover all the spatial,
mechanical, mathematical, logical, social, etc., directions and distances
between objects or between objects and the behaving self. 344 A Psychological Model However,
before we analyze further this matter of the inculcation of culturally approved cognitive placing matrices,
let us note that a similar situation seems to hold in relation to the other two
types of value standard - the appreciative and the
moral. I shall argue, in fact,
that in addition to a purely cognitive placing need and resultant matrices,
there are, or are developed, an appreciative (or aesthetic) placing need and
matrices and a moral placing need and matrices. The appreciative placing
need leads to matrices which categorize objects with respect to final resultant beliefs
concerning their leading to immediate enjoyability - such beliefs as "this
is pleasanter than that"; "this is nicer than that"; "this
is more beautiful than that"; and so forth. The moral placing need leads to matrices which
categorize objects (in this case primarily performances of self and others)
according to resultant beliefs concerning their likelihood of receiving moral
approval - such beliefs as: "this is good"; "this is bad";
"this is virtuous"; "this is sinful"; and so forth. Further, it will also be maintained that such appreciative
and moral "placings," like the cognitive placings, are, as far as the
resultant gratifications are concerned, equally
rewarding whether the judgments (beliefs) in question assert the beauty or
the ugliness, the virtuousness or the sinfulness, the truth or the falsity, of
the given objects. At this
point, it may well be argued that all three of these placing needs are
really subsumed under the first
- that is, that all such placings are However, this original, placing
need soon becomes differentiated into the three subtypes:
appreciative placing, moral placing, and cognitive placing. And it appears that these three subneeds may
acquire quite different average magnitudes
in different cultures, or in different individuals within one and the same
culture. Thus, some cultures (or
individuals) may be primarily cognitively oriented, others may be primarily
aesthetically oriented, and still others morally oriented.3 Granted
that cognitive, appreciative, and moral evaluations express three different
placing needs, we must still consider the nature and meaning of
"standards." For a culture
not only tends to encourage cognitive, appreciative, and moral placings, it
also tends to impose its rules or standards about just what is "so"
or true, what is beautiful, and what is good.
These standards, I would assert, consist merely in the fact that the basic ways of discriminating, generalizing,
and believing which become incorporated into the placing matrices are imposed
by the culture - learned from the culture. The precise empirical conditions which favor such learnings are
still unknown. I shall, however, hazard the assumption that two main
factors operate. 3
It must be pointed out, as has been indicated several times before, that to
date we have practically no good empirical techniques for the discovery of how
such placing needs are acquired, or if they should have innate beginnings, how
these beginnings are variously strengthened (or weakened) through
learning. Value Standards 345 First,
a culture has "names." It
has, that is, symbolic ways of focusing the
attention of its participants upon the particular discrimination and
generalization units and beliefs that it favors. The result of such names is to point out to the actor
repeatedly "what" he is to perceive and conceive, "how" he is to differentiate it from other "what's," and
what further "what's" it will lead to. Hence, an actor growing up in a culture learns to
"place" according to the standards of that culture, largely because
language brings about a focusing upon and a frequency of presentation of
such-and-such objects and relationships and no others. Second,
culture prescribes positive sanctions for
discriminating and generalizing and believing according to its rules and
negative sanctions for discriminating and generalizing and believing
otherwise. As a result, any nonapproved discriminations,
generalizations, or beliefs will tend to be repressed according to the
principles of repression suggested earlier.
It follows that the degree of acquisition of the value standards of the
community will he affected strongly by the magnitude
of the actor's need for social approval and hence his sensitivity
to such sanctions. The stronger this
need happens to be, the more likely the individual is to develop the
"accepted" placing matrices. Still
another point is to be emphasized. It
appears that value standards - i.e., cognitive, aesthetic, or moral
discriminations, generalizations, and beliefs - do
not necessarily bring about corresponding action. One may know the true, the beautiful, and
the good without seeking them. Other
practical, utilitarian needs such as the viscerogenic hungers may be too
strong. Ordinarily, placing matrices
indicate the layout and are then incorporated in other utilitarian matrices in
which cognitive gratification, appreciative gratification, or moral gratification
are not the main goals.4 In
summary, cognitive, appreciative, or moral value standards are, according to
this analysis, merely imposed ways of discriminating, generalizing, and
believing, which become established in the placing matrices of given cultures
or of given individuals. The culture or
the subculture determines by "namings" and by social sanctions the
precise units and the precise limits of such discriminations, generalizations,
and beliefs which will tend to be accepted.
However, it is also to be noted that even though an individual has been highly trained to "know" the true, the
beautiful, and the good, this does not necessarily mean that he will also want
them. Actually
to behave, an actor must not only have value standards, he must also have then-and-there
activated values. His utilitarian as well as his placing needs
must be aroused. But we will
understand this better after we have considered the pattern variables. 4
The fact that an understanding matrix may later be
incorporated in a utilitarian matrix was assumed in our discussion of
the latent learning of rats. It was
there implied that the discriminations, categorizations, and beliefs concerning
maze alleys and the locations of food or water which the rats achieved as a
result of a mere general exploratory (understanding) need, and which were laid
down in an understanding placing matrix, could then be incorporated into a
practical food-getting-to or water-getting-to matrix. 346 A Psychological Model The
five pairs of pattern variables, presented in Chapter I of Part II, were listed
as (a) affectivity vs. affective neutrality; (b) self-orientation vs.
collectivity-orientation; (c) universalism vs. particularism; (d) ascription
vs. achievement; and (e) diffuseness vs. specificity. Each pair will here be conceived to consist, not in a pair of alternate
moral value standards, but in two opposing
"moral placing" matrices, one or the other of which will tend to
determine the behavior of the given individual, or the given culture, relative to a certain area of object relationships. a. Affectivity vs. affective neutrality. When an object or situation presents an
actor with an opportunity either for a relatively quick and easy gratification
of a "lower" need or for a gratification of a relatively postponed,
difficulty-to-be-arrived-at, and supposedly greater and "higher" 5
need, some individuals (or cultures) will "place" the postponed
gratification as leading to moral gratification and the immediate gratification
as leading to moral deprivation, whereas other cultures or individuals will
"place" the two types of gratification in the reverse order. This means that in the moral placing matrix
for the one culture or for the one individual, immediate gratifications are
categorized as leading to final social disapproval; whereas in the moral
placing matrix for the other culture, or individual, they will be categorized
as leading to final social approval. b. Self-orientation
vs. collectivity-orientation.
This alternative arises when an
individual situation may be classified either as an opportunity for the
gratification of ego's own relatively private utilitarian needs or as an
opportunity for the gratification of the needs of a collectivity of which ego is
a part. In the one case, ego's moral placing matrix will categorize final moral
gratification (i.e., final social or self-approval) as resulting from ego's own
private gratification; whereas in the other case, ego's moral placing matrix
will categorize final social or self-approval as resulting from the
gratification of the needs of the collectivity. 5
See the brief remark on an earlier page concerning "higher" and
"lower" needs in accordance with Maslow's theory [16]. 347 c. Universalism vs. particularism. In the case of "universalism"
certain types of environmental objects and standards, for example, the
relationships between patients and doctors, will be discriminated and
categorized as situations in which the patient will choose (and behave to) a
doctor (and vice versa) according to the generally accepted, universal
standards of medicine. In the case of
"particularism," on the other hand, the relationships between
patients and doctors will be discriminated and categorized as ones of unique person-to-person
relations such as those of friendship or kinship. This means, further, that in the moral placing matrix of the one
culture the general relationship will be believed to lead to final moral
gratification of ego, and the particular relationship to final moral
deprivation of ego; whereas in the other culture the reverse will be true as far as such types of
relationship are concerned. d. Ascription vs. achievement. This contrast corresponds to two different categorizations by ego
of both self and alters in a moral placing matrix: a categorization primarily
in terms of their social statuses; or a categorization primarily in terms of
their performances and achievements.
Further, for one culture or individual the moral placing matrix says
that to categorize according to status is proper, right, and virtuous (i.e.,
will meet with final social and self-approval); whereas the moral placing
matrix of the other culture or individual says that to categorize human beings according
to their behaviors or achievements is proper, right, and virtuous (i.e., will
meet with final approval). e. Diffuseness vs. specificity. In this pair of pattern variables, "diffuseness"
corresponds to a moral placing matrix which makes no differentiation between
the different "segmental" relations of an alter to ego and vice
versa. If ego is involved in one sort
of relationship with alter, he is categorized as also being involved in every
other sort of relationship with alter.
In the case of "specificity," differentiations between the
different "segmental" relations of alter to ego, and vice versa, are
made in the moral placing matrix. If alter is categorized as in one of these relationships to
ego, he need not be categorized as also in the others. For example, in the case of diffuseness, if
alter is classifled as a sex object, he tends to be classified also and
forthwith as guide, counselor, and friend. In the case of specificity, on the
other hand, differentia- tions are made between the various potential roles of alter
to ego; the fact that alter is classified in one of them does not as such lead
to his classification in the others.
Finally, the moral placing matrix of the one culture will thus contain
the final belief that the diffuse categorization wfll lead to ego's moral
gratification (i.e., to social- and self-approval) ; whereas the moral placing
of the other culture will contain the belief that it is the specificity
categorization which will lead to ego's moral gratification (i.e., to social-
and self-approval). In the
above discussion of the pattern variables and in the preceding discussion of
value standards certain assertions and implications have been made about
placing matrices and their relationships to other matrices which perhaps can
only be rendered completely intelligible with the help of a diagram. Let me present, then, by way of
clarification a figure to represent one of the pattern-variable
alternatives. I will choose the first
pair - affectivity vs. affective neutrality.
A diagram for any one of the other pairs would involve the same general
principles. 348 A Psychological Model Fig. 21
AFFECTIVE
NEUTRALITY Fig. 21 presents a concrete illustration of a moral
placing matrix in the case of an individual, or a culture, for the case of affective
neutrality. It also indicates two
interacting utilitarian matrices. The
placing matrix is shown as the horizontal one.
The actor is conceived as being faced with money and the moral placing
matrix shows two alternative sets of behavior and accompanying beliefs. One of these (represented as radiating
slightly upward) envisages that money may be spent directly and lead to ego's
own viscerogenic gratifications, which, however (given the general culture or
the individual's private culture of "renunciation"), leads to the
further belief that this will result in moral deprivation. The other set of behaviors and accompanying
beliefs (represented as radiating slightly downward) envisages that money may
be put in a bank to collect interest and then may be willed to offspring. This belief, given the culture of
renunciation, leads to the further belief that this will lead to ego's own
moral gratification. As far as the
placing need is concerned, either of the belief chains will lead to the same
final placing gratification. Value Standards 349 We must
assume in addition, however, in order to explain what will actually happen, the
two "utilitarian" needs of (1) the viscerogenic hungers and (2) the
moral (social- and self-approval) need. These two needs lead to further matrices, shown as
cross-cutting the moral placing matrix, which indicate that if the social- and
self-approval need is very strong
(so that it heavily outweighs the viscerogenic hungers), then affective
neutrality will have the greatest terminal positive value and ego will tend to
locomote in his behavior space in the direction of saving his money and willing
it to his offspring. If, however, the
viscerogenic hungers are exceedingly strong, then, even though the moral
placing matrix accepted by the culture and by the individual in question
indicates that immediate viscerogenic satisfaction will lead to moral
deprivation, the viscerogenic gratifications will have such strong positive
values that ego will locomote toward them rather than toward saving. (Note: the
double-headed arrows indicate that the gratification and deprivation values of the cross-cutting matrices are, for
convenience of drawing, merely repeated inside the moral placing matrix.) So much
for a diagram of the matrix system for the pattern variable of affective
neutrality; the matrix system for its antithesis of affectivity would be
exactly the same, save that it would be the spending and immediate consummatory
alternative which, in the moral placing matrix, would lead to final moral
gratification and the other alternatives which would lead to moral
deprivation. It is obvious that a pure
gratification culture would produce practically no conflict. The viscerogenic hunger needs and the
social- and self-approval needs would both lead to positive values. Any
social system is made up of mutually overlapping social structures or
institutions. Each structure or institution (for example, the family) is
constituted by a number of complementary social positions or, as Linton [14]
calls them, social statuses. Thus, the
institution of the family, as we know it in America today, is made up of the
social status of a husband and father, the social status of a wife and mother,
and the social status of each child (which will he different according to age,
sex, and birth order). Further, the set
of behaviors which ego is expected to perform by virtue of his position as
father, mother, or child, is called a social
role. A man in his role
as father is expected to behave in certain ways; a woman in her role as mother
in certain other ways; and each child according to age, sex, and birth order in
still other ways. A role is thus a
series of appropriate and expected
ways of behaving relative to certain objects, by virtue of a given individual's
status in a given social structure or institution. 350 A Psychological Model Further,
these expectations that individuals in given statuses will behave in
such-and-such ways are called role expectations. This term has a double meaning. It applies
not only to the expectations of the alters (in this example, the other members
of the family and the public at large) that ego will behave in certain ways but
it applies as well to the expectations of ego that if he behaves in these expected
ways, the alters will meet his behavior with approval (or at any rate with lack
of disapproval) and with other
appropriate, complementary, meshing behaviors of their own. If ego does not so behave, the alters will
meet his behavior with disapproval and with other behaviors of their own which
do not mesh. Ego thus comes to expect
these resultant approvals or disapprovals and these meshing or non-meshing
complementary behaviors from the alters; and the alters come to expect the
complementary behaviors from ego. It is
obvious that these concepts of social roles and role expectations are very
close to the concept of the self-ideal presented in the section above on
psychodynamic mechanisms. The only
difference between the concept of a social role and the concept of a self-ideal
seems to be that the concept of the social role is limited to a relatively prescribed set of
stimulus situations, whereas the concept of the self-ideal concerns a
relatively wide, more abstract, set of behaviors, such as courage, honesty,
neatness, punctuality, and so on, which are demanded of ego in practically all
types of social situation. Indeed, it
would appear that the self-ideal probably develops out of the more limited
social roles. Ego learns the expected
behaviors in narrow role contexts and then generalizes them to all
contexts. He thus builds up his general
self-ideal. But there is a problem
involved; the behavior which is expected, by society in one role may turn out
to be quite different from that which is expected and required in another
role. This
brings us to the concept of role conflicts. The term “role conflicts" has been used
to refer to this very fact that any ego is usually involved at different times,
or even at the same time, in several different social structures or
institutions and that the sorts of behaviors expected of him in these different
social structures or institutions may be
incompatible. For example, an
adult male in our society will have to play the roles of father, lover,
businessman, community leader, club member, helpful domestic, and tidy gardener (especially in California).
And it appears that the types of behavior-space regions which be must
locomote to and from in one such social role may be quite different from those
which he must locomote to and from in others of these roles. The categorizations, beliefs, and values
which he acquires in connection with one role may conflict seriously with the
categorizations, beliefs, and values which he acquires in connection with other
roles. The more complicated the society
or the more fluid the institutions, the more likely such role conflicts
become. Value Standards 351 The
actor may meet these conflicts by such mechanisms as repression and symbolic
substitution, by a thoroughgoing neurosis or a psychosis, by discovering more
general categorizations and beliefs, by establishing new needs which allow for
the synthesis of such conflicting categorizations and beliefs and values,
and/or by easy "compartmentalization," which prevents generalizations
appropriate to the matrices required in one role from spreading over to those
required in other roles. The
concept of role acceptance also has
to be considered here. It concerns the conditions underlying an actor's original
acquisition or acceptance of a role.
Whereas many roles in our society (for example, that of a member of a
family) cannot, generally speaking, be escaped, choices regarding the acceptance or adoption of certain types of role are
encouraged. This is especially true in
a society such as ours. For example,
young people in our society are, in most cases, encouraged to select their own
professional or work roles. What are
the determining factors of such selections?
The answer seems to be that the factors consist mainly in the histories and
strengths of the successive positive and negative identifications which the
child has made with father, mother, teachers, gang mates, or external prestige
groups. The contents of these
identifications and the order in which they occur seem to be the important
factors. The youth aspires to a
particular job, or other career, because he wants to be like his father or
because he doesn't want to be like him; because he is trying to live up to some
social ambition implanted by his mother; because he and his gang are protecting
themselves against the ideals of teachers or parents; because of special school
encouragements or school discouragements; and so forth. To put this in other terms, the
factors that seem to determine ego's acceptance of any role, professional or
other, would he all the previous multidinous relationships with alters which
have developed ego's beliefs concerning the ways to obtain social approval
and/or self-approval. And these beliefs
are, in essence, that the way to obtain approval is by becoming like, or
unlike, such-and-such other individuals.
The tracing out of all the determining social experiences behind these
beliefs in any concrete individual case, or in any characteristic group of
cases, becomes, of course, a tremendously difficult empirical problem. An exciting first attack upon this problem,
with special reference to the adoption of work or job aspirations by junior
high school boys, has, however, already been made by Parsons, Stouffer, and
Kluckhohn [18]. The problems of role
acceptance, however complicated, are actually capable of empirical attack. Let us
turn now to our last concept, personality.
352 A Psychological Model This
topic, though of such tremendous importance for social science, can be touched
upon only briefly here. In an earlier
chapter, the significance of a personality as an on-going system was
emphasized. What, in terms of the concepts of the present chapter, would be the
necessary conditions for the non- disintegration, the
on-goingness, of a personality system? The answer would be that all the "needs" and all the
attached "belief-value matrices" which are relatively prominent in
the make-up of a given personality must somehow coincide and fit together so
that the actor possessing this personality will not be led into frequent
conflict; will not, that is, be presented in his behavior space with regions
that contain at one and the same time both a strong positive valence and a
strong negative valence. This requires
special types of coordination among the different belief-value matrices and
their attached needs. In an earlier
chapter the concepts of "integration" and "allocation" were
advanced as useful analytical tools for the study of such types of
coordination. Let us see how these
concepts may be helpful. First,
let us consider integration. An integrated personality would be one whose belief-value matrices (i.e., whose view of the
world) would add up to a set of relatively consistent ways of valuing
any given group of objects.
Probably, most important in achieving such an integrated personality
would be the consistency of the actor's beliefs about parents and
later about other authority figures. He
should not be subject to too much strain of both love and hate for the same
figures. And this would seem to require
further that his moral placing matrix - his beliefs about how he is to achieve
social- and self-approval - should not be too
inconsistent with his beliefs about how he is to obtain his primary
gratifications. The early authorities
who helped, loved, and approved him must also have encouraged
and not blocked too many of his simple primary needs. This sort of an integrated personality would
be integrated because the parents and the culture would give approval for the sorts of activity also
believed to be necessary and good for attaining gratification of the basic
biological and social needs. The
individual personality would tend to be integrated because the culture was
integrated. Further, it must also be pointed
out that a really integrated culture – producing really integrated
personalities - would have to be integrated not only internally, as just suggested (that is, no
conflicting moral and utilitarian values), but also externally (that is, no values, resulting
from moral or utilitarian matrices, which were incompatible with the actually
available, environmental conditions).
One of the obvious examples of a lack of such "external"
integration in our culture and a resulting difficulty in achieving personality
integration is the emphasis put by us upon standards
of material welfare which only a few can attain. This means the development of a type of
self-ideal the achievement of which is constantly met by negative barriers in
the behavior space. Conflict,
repression, symbolic substitutions, and other tendencies toward disintegrations
seem almost inevitable. Value Standards 353 Let us
turn now to allocation. Through allocation, as we use the concept
here, a considerable measure of integration can be achieved in a culture, or in
a personality, even though there are, in fact, a number of conflicting
values. This is achieved by allocating
the gratifications of the values to separate times and places. The
culture and the personalities nurtured in the culture have learned to draw
clear distinctions between acceptable and nonacceptable places and occasions
for the gratification of specific needs.
Thus, we have the traditional story of the New England
dairyman and church warden who worshipped God and gave to the poor on Sundays
and watered his milk on weekdays.
Integration (i.e., lack of conflict) was achieved through allocated, non-overgeneralized, beliefs as to when and
where this, that, or the other need is to be gratified. An on-going society and on-going
personalities would, in a successfully allocative society, be brought about, on
the one hand, by diminishing the original chances of overgeneralization and, on
the other, by psychotherapy which would consist in the successful correction of
such overgeneralizations when these overgeneralizations have occurred earlier
in a given personality. The
above discussion of personality has been far too brief and far too
abstract. It has not sufficiently included the specific concepts and
problems of the personality psychologist.
Further, it has not, as it should have, amplified the
remarks, made in the earlier portions of this paper, about learning and the psychodynamic mechanisms. It should also have analyzed the ways in
which the conscious and the unconscious portions of the belief-value matrices
interact with one another and what happens when one or the other of these
portions dominates. A further
elaboration concerning the influence of sequences of roles should have been
attempted; for it appears that any actor's final personality develops in large
part through his being forced into such-and-such a series of successive
roles. He acquires matrices (beliefs
and values) appropriate to each successive role. His final personality depends,
therefore, not only upon his innate, or early acquired, capacity and
temperamental traits, which make some roles easier for him than others, but
also upon the concrete nature of the roles, upon the order and the intensity
with which they have been accepted by him, and upon the ways and the degrees to
which they have or have not conflicted with one another. 354 A Psychological Model A few final remarks it must be
reemphasized that the study of personality, stated in terms of the present
model, is a study of belief-value matrices - their integration or lack of
integration - plus a study of need systems, the list of needs, the ways in
which the individual needs do or do not enhance or depress one another, and
their attachments to specific matrices.
Further, it may be pointed out that although the "adjusted
personality" in any given culture is perhaps usually the individual whose
own belief-value matrix system is most consonant with the "modal
matrix" 6 of the culture in general, this need not always be
the case. For would it appear that
there are types of unique individuals who deviate relatively widely from their
culture and yet achieve an integration of their own - some sort of an inner
consistency in their various need strengths and belief-value matrices which
holds up even though these needs and matrices run counter to those of their
fellows. Finally, it may be noted that
there are so-called creative individuals who set new pat- terns, new value standards, for a society. Their need
systems and belief-value matrix systems also differ from those of their
society. But these creative individuals may or may not be internally
well-integrated, well-allocated personalities. Many seem to be just the
opposite. They are neither stable because
they conform nor stable because they have an internal strength of their
own. They are often highly unstable;
and yet, in terms of the social system of which they are a part, they may make
tremendous social innovations - sometimes good, sometimes bad. To
summarize, personality is to be defined in terms of innate and acquired need
systems and in terms of belief-value matrices.
Personalities may be integrated and likely to survive or not integrated
and likely to break down. An integrated
personality is one which does not have conflicting values or whose conflicting
values are taken care of through allocation of their respective gratifications
to socially acceptable times and places.
The worth of a personality type, in terms of its contributions to the
social system of which it is a part, cannot be evaluated directly in terms of its
own degree of inner integration. Some
nonintegrated, very unstable, but creative personality types may make great
social contributions, some of which may add notably to the permanence and
on-goingness of the given social system.
6
See above for the meaning of this term. 355 6 Summary
and Conclusions An
attempt will be made here to summarize more specifically the distinctive
contributions of the model, as the present writer sees them, to psychology and
to social science as a whole. FOR PSYCHOLOGY The
average psychologist, immersed as he is in the details of empirical problems,
perhaps centered around one narrow area or perhaps scattered over a wide range
of areas, will ask what meaning such a model can have for him. My answer would be to point out, first,
that the model indicates the loci within a single theoretical framework of all
the different empirical problems with which psychologists concern themselves;
and second, that the model also presents specific hypotheses concerning
the functional, causal relationships involved in each of these empirical
problems. The classical chapter
headings of any general textbook in psychology run somewhat as follows: Sensation
and Perception; Memory; Learning, Thinking, Motivation, Attitudes, Skills,
Personality; Individual Differences.
The point I should like to make is that the
empirical problems designated by each of these chapter headings finds its
own natural and, I believe, illuminating locus or loci in the model. Let us consider each of these usual chapters
one at a time. A
chapter on sensation and perception
would contain, first, statements of the detailed functional relations between
types and intensities of delimited physical stimuli and resultant contents
(discriminanda) in the behavior space.
These statements are usually gathered together under the subheading sensation.
But such a chapter would also
contain considerations of the ways in which such details of
discrimination will also be affected by the larger spatial and temporal
stimulus contexts within which the given restricted stimuli are presented. Studies of this second
type are those which were largely initiated by Gestalt psychology;
such studies as those on figure and ground, completion of incomplete figures, pragnanz,
distortion in simple forgetting, and the like. 356 Both of
the above types of study would be located in the model by the arrow running
directly from the independent variable of the stimulus situation to the
intervening variable of the behavior space.
In addition, however, a chapter on sensation and perception
would include a study of all the further variations in the discrimination contents of the behavior
space, which are brought about by variations in the beliefs and values in the
superordinate belief-value matrices. It
would include such studies as that by Bruner and Goodman on the effect of the
money values of coins on the perception of their sizes, together with a host of
other recently reported "new look" experiments - all of which
indicate the tremendous effects on sensation and perception of the specific
values in the belief-value matrices. All these "new look" studies would be located on our model by the causal arrows running
from matrices to behavior spaces. A
chapter on memory (as I am using the
term here) would consider primarily the independent effects of single past
stimulus situations on present behavior spaces. The laws of memory as thus restricted would seem to be very much
the same as those for immediate sensation.
However, the term "memory" is often used to cover the problem
of the extrapolation of a behavior space owing to the fact that the now
presented stimuli have in the past been presented in temporal, spatial, and
other conjunctions with such- and-such other stimuli.
But in this sense the chapter on memory passes over into the chapter on
learning, to which latter we may now turn.
A
chapter on learning, as will be
recalled from our previous discussions, will from the point of view of our model,
have to deal with three main topics: (a) the establishment of new
differentiations in the behavior space; (b) the establishment of new
categorizations, new beliefs, and new values in belief- value matrices; and (c) the establishment of new acquired (tertiary)
needs in the need system. Let us
consider each of these further. (a)
The study of the establishment of new
differentiations in the behavior space would be a study of the
functional determination represented, in the model, by the arrow from the stimulus
situation to the behavior space. These
would be determinations over and above those considered in the chapter on
sensation and perception. It would be a
study also of an interaction effect between locomotion and the behavior
space. (b)
The study of learning as
a change in more enduring and generalized differentiations, beliefs, and values
in the matrix system would be a study of the laws represented in the model by
the arrow running back from the behavior space to the belief-value matrix. This study would consist in further
investigations of the facts already known - though in very inadequate detail -
concerning permanency and generalizability of learned relations; facts that in
former days were considered mainly under the heading of "transfer of training." (c)
Finally, the study of the acquisition
of new needs would be located on the model by the arrows leading back
from a belief-value matrix to the need system.
The laws and conditions of such learning are, however, as has already
been indicated, as yet practically unknown.
Summary and Conclusions 357 A
chapter on thinking would concern
itself with modifications of belief-value matrices, not as a result of concrete
behaviors but as a result of some internal type of process. This points to a feature in connection with
the model which we have up to now largely side-stepped. It arises out of the fact that human actors can apparently talk to
themselves, or in some other not as yet clearly formulated way, modify
the categorizations, generalizations, and beliefs in their belief-value
matrices without overtly behaving - without,
that is, actually coming in contact with any new stimulus features. Furthermore, the processes by which they do
this cannot be observed by an outsider.
An observer can tap the successive stages of the process by presenting
simple objective tests every so often and discovering how the contents of the
matrix have successively changed.
But he cannot study the process itself which produces the changes except
by asking for introspections. And the
unsatisfactory and unreliable character of introspection, when it comes to
reporting on thought processes, is notorious. The
process of thinking, then, seems to consist in some type of internal activity
which enables an actor to bring into play the consequences of given potential
types of behavior without, however, actually carrying out such behaviors. And as a result of these brought-into-play
consequences, he modifies or reformulates or expands his behavior space and his
belief-value matrix. Further, it is to
be emphasized that these reformulations are not the ordinary ones which have,
for the most part, been considered above as the result of actual behavior, but ones which result from purely sitting and thinking. Satisfactory psychological studies of this
process are as yet, I would submit, almost nil. A real chapter on thinking has yet to be written. A
chapter on motivation would
concern itself first with the relations between drives as independent
physiological variables and the need system as a set of consequent behavior
readinesses. It would, that is, be
concerned with the arrow (see Fig. 2) running
from the independent variable of drives to the intervening variables of the
need system. The chapter would, secondly,
also concern itself with the arrow running from the stimulus situation to the
need system; that is, it would have to consider not only the arousal of needs
as a result of physiological states but also as a result of presented stimuli. Third, this chapter would be
concerned with the causal functions represented both by the arrows from the
need system to the belief-value matrices and the arrows from the belief-value
matrices back to the need system. That
is, both the problem of the cathexis of a given matrix to a given
need and the problem of how a change in a matrix may lead to a new need would
have to be included. Still further,
the chapter would be concerned with the problem of how values in a matrix act
to produce valences in a behavior space as well as with the problem of how the
deprivation value in a matrix produces a need-push in the behavior space. Finally, a chapter on motivation
would concern the problem of the strengths of the resultant field forces in the
behavior space as a result of the interaction of need-pushes, valences, and
distances in the behavior space. In a word, it appears that the chapter would
recapitulate from its own angle practically everything said in all the other
chapters. 358 A Psychological Model A
chapter on attitudes would, from
my point of view, be nothing more or less than a descriptive study of beliefs
and resultant positive and/or negative values in belief-value matrices in
specified populations of individuals.
Indeed, attitude questionnaires, as was suggested earlier, would seem to
be one of the main techniques for investigating belief-value matrices. Heretofore, of course, most attitude
studies, resorting to questionnaires or to interviews, have been concerned with
discovering merely the presence or absence of specific attitudes (belief-value matrices) in given
populations for practical social or political purposes. In the sort of chapter on this topic which I would envisage,
the emphasis would be rather on a study of the dynamic interrelations between
different attitudes - different belief-value matrices. The ways in which one matrix can or cannot
include another, the necessary relations of superordination between matrices,
and the like - these would be studies which, from the point of view of our
model, would be the most exciting. A
chapter on sensory-motor skills
is for the most part yet to be written.
In terms of the model a discussion of sensory-motor skills would be a
discussion of how locomotions in
the behavior space get translated into overt behaviors. It would also be a discussion of how the
sensory-motor activities of the organism - the actual behaviors which he learns, or
develops through mere maturation - come to react upon and become constitutive
of regions of his behavior space. But,
as I have said, such a chapter has as yet never really been attempted. A
chapter on personality, as would
appear from the preceding section, would be concerned with the individual's
need system and his belief-value matrices.
But it would also be concerned, by tradition, with the nature of the
psychodynamic mechanisms. Still
further, it would have to consider individual differences. Thus, to a considerable degree, a chapter on
personality, like a chapter on motivation, would be a chapter recapitulating,
from its own point of view, practically the whole of psychology. This chapter would, that is, be concerned with psychology from the special angle of the
resultant integration or non-integration, smoothness or non-smoothness, of the
resultant functioning of the individual actor.
Summary and Conclusions 359 Finally,
a chapter on individual differences
would be concerned, as suggested above, with determining the values of the
constants for different individuals in all the causal functions (represented in
the model by arrows) considered in the other chapters. But this is a topic upon which I am not
qualified to speak. The
individual-difference psychologists, with their tests, their rating procedures,
and their statistical techniques, have acquired an array of tremendously
important facts, but these facts have still to be fitted into the structural concepts of such a model as ours. This remains a major task which, it is
hoped, may be undertaken at some future time.
Let us
turn now to the question of the significance of our model for social science as a whole. FOR SOCIAL SCIENCE The
argument here will be brief. It has been indicated in the other sections of
this book that the basic categories for the structural analysis of social
systems, of culture systems, and of personality systems must be consonant with
one another. For individual actors are
but role-takers in collectivities; and both individuals and collectivities
internalize culture patterns and respond to cultural accumulations; and,
finally, individual actors respond to collectivities and collectivities respond
to individual actors. The psychological
model presented here is then, I would assert, no more than a restatement of the
same conceptual categories developed in the other chapters of the hook for social
systems and culture systems, but colored somewhat by the special needs and
interests of the psychologist.
Psychology is in large part a study of the internalization of society
and of culture within the individual human actor. Therefore, our psychological model has sought to ferret out and
to classify the internal workings of the individual human actor by adopting the
same kind of taxonomic principles used in the sociological and anthropological
chapters of this book. And, although I
have not always used the same words or retained all the same emphases, I have,
I hope, remained true to the sense and to the spirit of these other social and
cultural analyses. Bibliography 360 Bibliography 1. AlIport,
G. W. Personality, A Psycfiological Interpretation. New York: Holt, 1937. 2. Bruner,
J- S., and C. C. Goodman. "Value and Need as Organizing Factors in
Perception," burn. Abnorm. and Soc. Psychol., XLII (1947), 33~4. 3. Brunswik,
E. "Organismic Achievement and Environmental Probability," Psychol.
Rev., L (1943), 255-272. 4. Frenkel-~runswik,
E. "Motivation and Behavior," Genet. Psychol. Mono g., XXVI (1942),
121-264. 5. Guthrie,
E. R. The Psychology of Learning. New York: Harper's, 1933. 6. Guthrie,
E. P., and G. P. Horton. Cats in a Puzzle Box. New York: Rinehart, 1946. 7. Hebb, D.
0. Organization of Behavior. New York: Wiley, 1949. 8. Hull, C.
L. Principles of Behavior. New York: Appleton-Century, 1943. 9. Krech,
D. "Notes toward a Psychological Theory," fourn. PersonaL, XVIII
(1949), 6~7. 10. .
"Dynamic Systems, Psychological Fields, and Hypothetical Constructs,"
PsychoL Rev., LVII (1950), 283-290. 11. Kendler,
H. H., and W. P. Gasser. "Variables in Spatial Learning: I. Number of
Reinforcements during Training," Journ. Corn p. and Physiol. Psychol., XLI
(1948), 178-187. 12. Ko~hler,
W. Dynamics in Psychology. New York: Liveright, 1940. 13. Lewin,
Kurt. "The Conceptual Representation and the Measurement of Psychological
Forces," Contrib. Psychol. Theory, I (1938), no.4. 14. Linton, R.
The Cultural Background of Personality. New York: Appleton-Century, 1945. 15. MacCorquodale,
K., and P. E. Meehl. "On a Distinction between Hypothetical Constructs and
Intervening Variables," Psychol. Rev., LV (1948), 95-107. 16. Maslow, A.
H. ' 'Higher' and 'Lower' Needs," burn. Psychol., XV (1948), 433~36. 17. Murray, H.
A. Explorations in Personality. New York: Oxford University Press, 1938. 18. Parsons,
Talcott, S. A. Stouffer, and Florence Kluckhohn. Unpublished manuscript
concerning Research Seminar on Social Structure, Harvard University. 19.
Ritchie, B. F. "Studies in Spatial Learning: VI. Place
Orientation and Direction Orientation," Journ. Exp. Psychol., XXXVIII
(1948), 659-669. 361 20. Sears, R.
R. "Effects of Frustration and Anxiety on Fantasy Aggression." Address delivered at Sixtieth Anniversary
Celebration, Clark University, April 1950.
21. Skinner,
B. F. The Behavior of Organisms: An Experimental Analysis. New York: Appleton-Century, 1938. 22. Tolman, E.
C. Purposive Behavior in Animals and Men. New York: Appleton-Century, 1932. 23. -
"The Nature and Functioning of Wants," Psychol. Rev., LVI (1949),
357-369. 24. ~
"The Psychology of Social Learning," Journ. Soc. Issues, V (1949),
Supplement no.3. 25. .
"Interrelations between Perception and Personality: A Symposium, Part I,
Discussion." Journ. Personal., XVIII (1949), 4850. 26. Tolman, E.
C., B. F. Ritchie, and D. Kalish. "Studies in Spatial Learning:
1.Orientation and the Short-Cut," burn. Exp. Psychol., XXXVI (1946),
13-24. 27. .
"Studies in Spatial Learning: II. Place Learning versus Response
Learning," Journ. Exp. Psychol., XXXVI (1946), 221-229. 28. -
"Studies in Spatial Learning: IV. The Transfer of Place Learning to Other
Starting Paths," Journ. Exp. Psychol., XXXVII (1947), 39-47. 29. Warden, C.
J., et al. Animal Motivation: Experimental Studies on the Albino Rat. New York:
Columbia University Press, 1931. 30. Young, P.
T. "Food-seeking Drive, Affective Process, and Learning," PsychoL
Rev., L (1949), 98-121.
ideological furnishings for the homeless
mind
daurril
library: talcott parsons
LEARNING
Fig. 5 RAT AT CHOICE- POINT; EXPLORATORY NEED
OPERATIONAL
DEFINITIONS
Need Systems
Belief-Value Matrices
Behavior Spaces
VALUE
STANDARDS
PATTERN
VARIABLES
SOCIAL
ROLES
PERSONALITY