On the Unity of Thought and Belief

On the Unity of Thought and Belief MILTON ROKEAGH Michigan State Umve

I N RECENT YEARS there have appeared a number of empincal studies on the relation between social belief and cognition In most of these studies the specific social belief under scrutiny was ethnic prejudice, or the authoritarianism conceived to underlie it (1) And the major findings of such studies are that ethnic prejudice and/or authoritarianism are significantly related to rigidity (4, 7, 16, 17,^ 31, 33, 48, 60, 61, 62), concreteness (49), and narrowness (50, 51) of thinking and problem solving, to premature closure of perception (3, 6, 16, 17, 59), and to distortion of memory (13, 16, 17, 52) Thus far the major attempt to account for such results is to be found in the work of Frenkel-Brunswik, who has written extensively on the problem of the interrelation among personality, belief, and cognition under the general heading of "personality-centered" approaches to perception (15-18) * Her mam thesis, derived from psychoanalysis, is that there exists a sort of lsomorphic correspondence between the cognitive spheres of behavior on the one hand and the emotional and social spheres on the other Stated very bnefiy, she starts with the psychoanalytic concept of emotional ambivalence and the role it plays in the development of the personality structure of the child, and develops the notion that as a result of early parentchild relationships there emerge individual differences in the abihty to tolerate emotional and social ambivalence which have im|rfications for differences in the cognitive spheres as well In this way a ' A revised version of paper prepared for the Conference on Thinking held in Cambridge, England, ui August, 1955 A major portion of the work reported herein was conducted dunng the author's tenure as a Faculty Research Fellow of the Social Science Research Council ' S e e also Levmson ( M ) , Weissk How can it be measured? As a result of the use of what Lewm has called the method of gradual approximation (34), there emerged, one might say, an increasingly differentiated conception of the organization of the behefdisbehef system A preliminary theoretical statement of how this organization may be conceived has already been presented elsewhere (54) The mam points of this analysis will be reiterated here, modified in several important respects This will be followed by a discussion of how the several properties of belief-disbelief systems were measured, and how such measures were employed m several researches designed to throw light on some of the structural mterrelations assumed to exist between belief and thought THE BELIEF-THOUGHT MODEL

Assume first that all cognitive systems are organized into two interdependent parts a belief system and a disbelief system Organization along a Belief-Disbelief Dimension A person's disbelief system is further conceived as being composed of several disbelief subsystems, each arranged along a continuum of similarity to the belief system Furthermore, the extent of a person's rejection of each disbelief subsystem, and the adherents thereof, is assumed to bear some systematic relation to this similarity-dissimilanty continuum This relation will be referred to as the disbelief gradient and a study will later be described in which the shape and height of this gradient is determined for several groups of subjects The belief-disbelief system is further conceived as having additional properties in terms of which it may vary 1 Isolation This refers to the degree of communication or mterconnectedness between and within belief and disbelief systems (29, 36) Cognitive manifestations of isolation would be the existence of contradictory beliefs withm the belief system—Orwell (46) has more picturesquely called this "doublethink"—the degree to which one emphasizes the differences rather than the similarities be-

THOUGHT AND BELIEF 231 tween belief and disbelief systems, the extent to which one perceives as "irrelevant" what may well be relevant, and the extent to which there is a cognitive denial of opposing or contradictory facts on such grotmds as "face absurdity," chance, or that it is the "exception which proves the rule " 2 Differentiation It is assumed further that there are varying degrees of discrepancy m the degree of differentiation of the belief system as compared with the disbelief system, and withm and among the several disbelief subsystems with respect to each other Cognitive manifestations would be the relative amount of knowledge possessed about one's belief system as compared with each of the disbelief subsystems, and the extent to which adjacent disbelief subsystems are perceived to be similar or different 3 Comprehensiveness or narrowness of the system This refers simply to the total number or range of disbelief subsystems represented m a given belief-disbelief system (9, 50, 64) Thus, to one Methodist, Mohammedanism and Taoism may actually be represented m the behef-disbelief system while to another they may be meaningless words Organization along a Central-Peripheral Dimension To handle a number of complex theoretical issues, it is necessary to distinguish further three sets of belief-regions organized along a central-peripheral dimension (1) A central region, representing what will be called one's "primitive" beliefs These have to do with all the beliefs a person has learned about the nature of the world one lives in, the nature of the "self" and the "generalized other" (43) ; (2) an intermediate region, representing the beliefs a person has in and about the nature of authority to be depended on to form a picture of the world one lives in, rational in so far as possible and rationalized in so far as necessary, to steer a homeostatic course through life; (3) a peripheral region, representing the beliefs perceived to emanate from authonty, whatever its nature, such beliefs filling m the details of this world-picture * ' T h e earher formulation of the organization of belief-disbelief systems (54) was in terms of a central region, corresponding to (2) above, and a peripheral region corresponding to (3) above Incidentally, "peripheral" beliefs are not necessarily ummportant to the person But such beliefs, as compared with central or intermediate beliefs, are more capable of being "jettisoned" if and when the person is confronted with dire threat

232

MILTON ROKEACH

1 The central region If one wishes to understand somethuig of the functicmal aspects of belief-disbelief systems we mtist first locrfc to the content of the central core of "pnmitive bdirfs" rather than to the content of the intermediate or peripheral beliefs. Note that content rather than structure is being emphasized here. What seems important to know about these primitive beliefs is their specific content about the social world one lives m, the content of one's sdfconcept and one's concept of others Such content, it is assumed, will have much to do with the formal organization of the rest of the belief-disbelief system 2 The intermediate region This r ^ o n is ctmceived to represent the beliefs different persons may have about authority, ranging from rational at one extreme to arbitrary at the other (19) By authority is meant not only positive authority but also a gradient of negative authorities, corresponding lsomorphically with the disbelief gradient In the intermediate region are also represented closely related beliefs about the existence—or nonexistence—of one true cause, one elite, one bible, etc Furthermore, there is represented within the intermediate region the beliefs one has about people who hold beliefs in agreement with one's own or in disagreement with one's own This formulation (of beliefs about people-who-holdbeliefs) IS an important one because it permits the binding together of questions often asked about the acceptance and rejection of ideas with questions about the acceptance and rejection of people. And here too it may be assumed that rejection of those who disagree with us IS not necessarily to the same extent. Corresponding to the disbelief gradient and to the negative authority gradient is assumed to be an intolerance or disbeliever gradient While our mam interest with respect to the primitive beliefs represented in the central region has to do with their specific content, this IS not the case with respect to the intermediate beliefs What seems of most interest here is its jormal content Consider, for example, two persons adhering to different ideological positions They may both believe that there is such a thing (or that there is not such a thing) as absolute authority, that there is such a thing (or not) as one true cause, one true bible, and so on It may be said that while the specific content of such beliefs ts different, the formal content is the same It is in this sense that the term "formal

THOUGHT AND BELIEF 233 ajntent" is used This conception of "formal content" is a helpful one because it enables us to look for similarities among persons with respect to their beliefs m authority and people m general, even though they are heterogeneous m their political, religious, or scientific orientations. 3 The peripheral region Represented within the peripheral region are all the beliefs and disbeliefs emanating from positive or negative authority, regardless of whether they are so perceived consaously by the person himself The speafic content of these beliefs and disbeliefs, like the specific content of beliefs m authority, will, of course, vary from one person to another, and it is such specific content which cues us m identifying a person as being of this or that ideological position But this is not what is of major concern here Rather what is of most concern is the structural interrelations among peripheral beliefs and, in turn, their structural interrelations with intermediate and central beliefs It is these various interrelations, rather than solely logical consistency, which give the total behef-disbelief system its holistic or systematic character. All information impinging upon the person from the outside must be processed or coded m such a way that the information is rejected or else fitted somehow in this system It is this processing-coding activity which we call thinking, and surely It must be within a context hke the behef-disbelief system that thinking must take place It IS far from clear how the processing-coding operation proceeds. But as a first approximation it may be assumed that this operation begins with the person screening the new information for compatibility with the primitive beliefs (central behef region) The initial screening may lead to the rejection or narrowing out of new information so that nothing further need be done with it Indeed, to one extent or another, we all narrow down our range of activities (eg, the people we choose and won't choose as friends, the books and journals we read and won't read, the churches we attend and won't attend, etc.) so that much new information will not even reach us But not all new information is handled this way Much new information does get through and need not be rejected if it can somehow become assimilated into the belief-disbehef system. This may

234

MILTON ROKEACH

necessitate the alteration-rationalization of the new information, to the extent necessary, by finding out what one's authority sources, positive and negative, have to say about it (intermediate belief region) And the final step is the fitting-filmg of this information, which now may or may not be new, into whatever world outlook one has come to call his own (peripheral belief region) Let us look at the fitting-filing process a bit more closely The new information is communicated from the central (primitive) region to the intermediate (authority) region and, in tum, to the peripheral region, and is represented m the latter in the psychological form of a new belief or disbelief Such a belief or disbelief may or may not communicate with other peripheral beliefs in the system, depending upon the degree of isolation among such beliefs The greater the isolation the less effect will a change m one part of the peripheral region have directly upon adjacent parts But there will still be communication between peripheral beliefs via the intermediate (authonty) region It is essentially m this way that one can conceptualize a "party-line" change and differentiate such a change from a more "genuine" change It shoxild be emphasized at this point that the processing-coding operation is not necessarily to be conceived as a coercing operation. The extent to which it is a coercing operation depends upon the degree to which the total belief-disbelief system is a closed or open system (more of this shortly) At the closed extreme, it is the new information which must be tampered with—by narrowing it out, by altering it, and by containing it withm isolated bounds, etc —while the belief-disbehef system is left intact At the open extreme it is the other way around new information is assimilated as is, and, m the process of fitting-filing, by communicating with other peripheral beliefs thereby changes the whole ideological system Organisation along a Time-Perspective Dimension A third major dimension of belief-disbelief system is tune perspective (14, 35), referring to the beliefs one has about the present in relation to past and future This dimension is conceived to vary from narrow to broad Cognitive manifestations of a narrowed time perspective would be the belief that the present is unimportant m its

THOUGHT AND BELIEF 235 own right, the belief that it is only the future or past that counts, and the belief that one knows what the future holds in store The Reduction of the Multt-dimenstonal Model to a Single Dimension Organisation along an Open-to-Closed Dimension The thought-belief model just presented is a three-dimensional one (belief-disbelief dimension, central-peripheral dimension, timeperspective dimension) and each of these dimensions, in turn, has additional properties But with an eye out for a way to measure quantitatively the whole behef-disbelief system along some single dimension it is convenient to assume further that these three dimensions and the several properties thereof are all lntercorrelated to such an extent that they are all reducible to a single dimension, namely, organization of the total cognitive system along a continuum from open to closed A belief-disbelief system will now be defined to be closed to the extent that A With respect to the organization along the belief-disbelief continuum, 1 The magnitude of rejection at each point along the disbelief gradient is relatively high, 2 There is isolation of parts withm and between belief and disbelief systems, 3 There is a discrepancy m degree of differentiation between belief and disbelief systems, 4 There is dedifferentiation withm the disbelief system B With respect to the organization along the central-peripheral dimension, 1 The specific content of primitive beliefs (central region) IS to the effect that the world one lives m, or the situation one IS m at a particular moment, is a threatening one, 2. The formal content of beliefs about authority and people who hold to systems about authority (intermediate region) IS to the effect that authority is absolute and that people are to be evaluated according to their agreement or disagreement with such authority, and according to their agreement or disagreement with the beliefs such authority represents, 3. The structure of beliefs and disbeliefs perceived to emanate

236

MILTON ROKBAGH

from absolute authority (peripheral region) is composed of substructures which are isolated from each other And finally C. With respect to the organization along the tune perspective dimension, there is a narrowing of time perspective To return now to the original point of departure which led to the model, namely, the problem of dogmatic thought and belief In terms of this model dogmatism can now be defined simply as a closed belief-disbelief system By belief-disbelief system is meant the thought-belief model, as set forth above, and by closed is meant one extreme form of it, as set forth above T H E MEASUREMENT OF BELIEF-DISBELIEF SYSTEMS

To be discussed now is how the model has been put to work A first step was to try to devise a reliable measure of individual differences m the extent to which belief-disbelief systems are open or dosed, derived from the analysis of the several dimensions and properties of the model A second step was to test the validity of such a measure, and along with this the validity of the model, by setting up several hypotheses, also generated by the model A series of belief statements is presented to the subject, who may agree or disagree with each on a -\-3 to —3 scale (with the zero point excluded in order to discouragi neutral responses) These statements address themselves specifically and systematically to various aspects of the cognitive model just presented, and taken together, comprise the Dogmatism Scale This IS not the place to present the full details regarding the content and statistical properties of this scale " However, to illustrate the derivational procedure followed, there are listed below in tabular form some of the properties of behef-disbelief systems and next to each some of the belief statements designed to measure them The belief statements shown in the table, as well as others constructed m the same way, have been found to hang together fairly well For various groups tested m the United States and in England the Dogmatism Scale has corrected reliabilities ranging from about 70 to about 90, depending upon the number of statements in the total scale ' To be presented elsewhere (M)

THOUGHT A N D BELIEF TABLE 1 REPKESENTAITVE ITEMS FXOM THE DOGMATISM SCALE DIMENSION

BeliefDisbelief Dimension

PROPERTY

In times like these it is often necessary to be mote on guard against ideas put out by people or groups m one's own camp than by tiiose in tne opposite camp

Isc^tion within belief system

The highest form of government is a democracy, and the highest fbrm of democracy is a government run by those who are most lnteUigent

Isolation between belief and disbelief system

The Umted States and Russia have just about nothing m common

Relative differentiation of belief and disbebef

It IS only natural that a person would have a much better acquaintance with ideas he bebeves in than with ideas he (^>p>oses

the disbelief system CentralPenpheral Duneiunon

B E U E T STATEMENT

The disbehef gradient

Speafic content of primitive beliefs

i here afe certain isms wiucii are really the same even thoi^h those who bebeve m these "isms" try to tell you they are different Man on his own is a helpless and miserable FundamentaUy, the world we hve in is a pretty lonesome place Most people just don't give a "damn" for

Formal content of intermediate bekef region

Of all the philosophies which exist in this world there is only one which is correct There are two kinds of people in this world those who are for the truth and those against the truth. ^4y bloou bolls whenever a person stuoDomlj renises to admit he's wrong

Isolation among penpheral behefs, communication between intermediate and penpheral behefs ("party-

In this comphcated world of ours the only way we can know what's going on is to rely upon leaders or experts who can be trusted

Narrowing

In the long run the best way to hve is to pick friends and assoaates whose tastes and behefs are the same as one's own

Narrowing fftiture-onented time perspective)

The present ts all too often full of imhappiness. It IS the ^ture than counts

L e ^ change)

TimePerspective Dimension

238

MILTON ROKEACH TESTING THE MODEL

Thus far there has been presented nothing more than a conceptual model of belief-disbehef systems ranging from open to closed, and a relatively reliable "global" measure purporting to measure it No claims can yet be made regarding empirical validity It is to this issue that attention is now directed Two sets of hypotheses have been tested thus far m our research program one set is concerned with attempts to achieve a broader conceptualization and measurement of general authoritarianism and general intolerance," the second set has to do more directly with the way we think and solve problems It is the second set of hypotheses which IS of most interest m the present context since it has to do with the interrelations between belief and thought One way we chose to explore the open-to-closed dimension of the thought-belief model was to invent a new system of reality, requiring a new belief system, and determining how readily a person can break away from his present system m order to adopt the new belief system as his own m order to solve problems withm this new system Imagine therefore a fictitious world wherein the rules of the game are m contradiction to the world we now live in Let the new beliefs be organized into a system wherein the beliefs must be m a high degree of intercommunication with each other if one IS to solve problems within this fictitious world Let the total number of such beliefs be relatively small, so that we can keep track of them Then, let us set up various problems withjn the framework of such a miniature world-system, the solutions of which hinge upon the subject's willingness or capacity to adopt this new system as "true," that is, to make of what is now not even represented in one's belief-disbelief system, part of one's belief system The more closed a person's present belief-disbehef system, the more difficulty he should encounter m solving problems withm this new system Or to put it the other way around, the more difficulty a person manifests in switching over to this new system, the more closed must be the organization of his present behef-disbelief system In one study by Rokeach, McCiovnQr, and Denny (55), and in another study recently completed by Vidulich (66), subjects are confronted with just such a hypothetical world. The main te-o of * "^ b

THOUGHT AND BELIEF

239

this world IS a little bug named Joe Doodlebug Joe, the subjects are told, is a strange sort of bug He can jump m only four directions north, south, east and west—^not diagonally, once he starts m any direction he must jump four times in that direction before he can switch, he cannot crawl, fly, or walk—he can only jump, he can jump large distances and small distances, and he cannot tum around Joe's master places some food, larger m diameter than he, three feet directly west of him Joe stops dead m his tracks, facing north After surveying the situation Joe concludes that he must jump four times to get to the food We impress upon the subject that Joe is quite right m his conclusion He must take four jumps, no more, no less We then ask the subject to describe the circumstances Joe must have been in which led him to reach this conclusion. He is given 30 minutes to solve the problem, to tell us why Joe reaches the conclusion he does There are three beliefs the subject must first overcome one by one, three new beliefs which he must then incorporate one by one mto his belief system, and then integrate into the problem solution, (a) The facing belief Joe does not have to face the food m order to eat It. He can land on top of it (&) The direction belief. Even though Joe is forever traiq)ed facing north, he can still change direction by jumping sideways and backwards (c) The movement belief When Joe stops to survey the situation he may not necessarily be a free bug His degrees of freedom may be limited by the fact that he could have stopped m the middle of a sequence of jumps rather than at the end of a sequence This problem is known as the Denny Doodlebug Problem and it IS a rather difficult one ^ Most subjects need help if they are to solve It in 30 minutes The reason why it is so difficult is that the subject must first overcome not one but three sets (beliefs) and replace them with new ones But this is not all. Even if he overcomes and replaces all three sets with new ones he must still put them together —integrate them—into a new system in order to solve the problem To keep track of the thought process the problem is deliberately made diffictilt, the subject is encouraged to think out loud, and he ' The sdtrtion • Joe had jumiwd once to the east when Ae food was p He has to jump thrice more to the east, and once back to the west, landing on top oAhe*f< -' ' e food

240 MILTON ROKEACH IS given hints along the way If the problem were too easy the subject would solve It too soon and we wouldn't be any the wiser as to how he arrived at his solution Other conditions are set m the experiment m order to find out what goes on by way of mental activity If the subject shows no signs of having overcome any of the sets (beliefs) by himself at the end of fifteen minutes he is given, m the form of a hint, the facing belief, at the end of 20 minutes, the directions belief, and at the end of 25 minutes, the movement belief In this way two discrete stages of the problem solving process are isolated and measured (a) how easily or quickly the subject is able to overcome the individual sets by himself, and (b) how easy or difficult it is for the subject to int^jate the new sets mto the problem solution after he has overcome the old sets. Let us look at these two stages a bit more closely Present-day theory and research are typically concerned with the first stage mentioned above, namely, the laws regarding the overcoming and learning of what are essentially single concepts, percepts, sets, Aufgabe, habits, expectancies, hypotheses, or attitudes (8, 12, 20, 25, 27, 30, 37, 38, 47, 67) There is little theory and less research about the development and learning of set-systems, hypothesis-systems, attitude-systems, etc Thus, Bruner (8) and Postman (47) are concerned with the nature of an hypothesis and the determinants of art hypothesis' strength * And as has been pointed out elsewhere "In previous research on thinking the obstacle to the problem solution IS usually a single set the overcoming of which is equivalent to the solution But real-life problems are generally more comphcated than this, typically involving the overcoming of multiple sets rather than smgle sets and involving further their integration mto new cognitive organizations" (55, p 87) It IS with considerations such as these m mind (namely, the separation of the two discrete problem-solving stages) that the writer • A major exception which comes to mmd is Tolman, who talks about "cognitive maps" (M) and "belief-value matrices" (IS) But no suggestions are made by Tolman on how such "maps" or "matrices," considered as a totality, may be measured. At first glance, Harlow's "learning sets" ( U ) seem to be another exceptiwi But closer scrutiny suggests that the "learmng sets" Harlow refers to are equivalent to what others have called concept formattcm ( e g , object quality discrumnations or posiUonal discnmmaticms or reversal dtscruntnatimis) These "learning sets" do not refer to combinations of such discntmnaticms operating simultaneously

THOUGHT AND BELIEF

241

has elsewhere (54) distinguished, at the personality level, the vanable of dogmatism from that of rigidity While the former has to do with systems of beliefs-disbeliefs resisting change, the latter has to do with smgle beliefs resisting change And the separation of the problem-solvmg process mto two discrete stages, as m the Doodlebug Problem, provides us with an opportunity to find out empirically not only whether the thought-belief model, and the Dogmatism Scale designed to measure it, have any validity but also with the opportunity to find out whether the distinction drawn between rigid and dogmatic thinking is empirically a supportable one If it is, subjects scoring low and high on dogmatism will have to differ from each other primarily with respect to the integration-of-new-beliefsafter-the-older-beliefs-have-been-overcome stage, while subjects scoring low and high on rigidity will have to differ from each other primarily with respect to the overcommg-of-mdividual-beliefs stage The results found m a study by Rokeach, McGovney, and Denny (55) are in accord with these expectations Subjects scoring high in rigidity, as measured by a personality scale (21), manifest significantly greater difficulty than subjects low in rigidity in overcommg the individual sets on the Doodlebug Problem But these two groups do not differ from each other, after the individual sets had been overcome, in their ability to integrate the new beliefs into the problem solution. Conversely, subjects scoring high and low in dogmatism do not differ from each other m ease of overcoming individual sets, but they do differ from each other in speed of integrating the new beliefs mto the problem solution Moreover, a record was kept of the various comments made by the subjects, some of which seemed to suggest that certain subjects, even though they overcome one or more mdividual sets by themselves, had no intention of "going all the way" to organize such beliefs into an integrated belief system "Let Joe starve"; "This experiment is ridiculous", "Only an Einstein could solve this one"—these are some comments to this effect In support of the mam findings, it was found that the low and high dogmatic groups differ significantly from each other m the frequency with which such rejecting remarks were made No such differences obtained between low and high rigid groups Moreover, some of the dogmatic subjects kept wanting to change the Doodlebug Problem. They want to solve the problem on their own terms rather than

242

MILTON ROKEACH

the experimenter's. Thus one subject remarks. " I don't believe that Joe has to take four jumps I think he can do it in o n e ' " W a n t i n g to change a problem and rejecting a problem seem hardly conducive to facilitate integration into a new system For if new beliefs are not really accepted they may not even become part of one's belief-disbehef system, and consequently there may be nothing to integrate Is it this that happens more frequently m the high dogmatic subjects' If these deductions are correct it should be expected first that highly dogmatic subjects will manifest a poorer incidental memory for the new beliefs than subjects less dogmatic Second, if a subject can somehow be made to accept new beliefs, whether he likes it or not, he will at least have some "bricks" to start with and this should mcrease the chances that the new beliefs will become integrated mto a new system Viduhch (66) has recently tested these deductions experimentally As before, the task is the Denny Doodlebug Problem but with important modifications The three new behefs (the facing, direction, and movement beliefs) are typed on separate cards As the subject discovers one or more of these by himself the appropriate card is placed before him Failing this, each of the three cards is placed before him at specified time intervals m the form of a "hint " The expenmental group is composed of a high and low dogmatic group They are allowed to keep the cards m front of them (the "Keep Card" condition) The control group is also composed of high and low dogmatic groups Here too the "belief cards" are shown as each specific set is overcome, with or without the experimenter's aid But each card is taken away immediately after its contents are read (the "Take Card Away" condition) As before, the subject is encouraged to think out loud, his comments are noted, hints are given at specified time mtervals, and the time taken to complete the two problem-solving stages is recorded Shortly after the end of the experiment the subject is asked to recall the three new beliefs and to answer several questions designed to find out if he really "accepted" the new world of Joe Doodlebug In this way the effects of both personality and situational variables are studied withm the confines of a smgle experiment. Both sets of variables address themselves fundamentally to the same issue.

THOUGHT AND BELIEF 243 namely, why the integration process fails and what must be done to and for the subject so that there is a better chance for it to succeed Theoretical expectations are confirmed in almost every respect As before (55), (1) the low and high dogmatic groups, under both experimental and control conditions, do not differ from each other m speed of overcoming individual beliefs (2) The high dogmatic groups, tmder both sets of conditions, clearly take longer than the low dogmatic groups to integrate the new beliefs into the problenysolution Furthermore, (3) the high dogmatic groups, under both sets of conditions, manifest at the end of the experiment a poorer incidental recall of the new beliefs than do the low dogmatic groups The greater difficulty m integrating manifested by the former, then, can be attributed to the fact that the new beliefs needed for integrating are less well remembered (4) The high dogmatic group which is allowed to keep the belief cards in front of them (Keep Card condition), as compared with the high dogmatic group which is not allowed to do so (Take Card Away condition) (a) integrate the new beliefs more quickly mto the problem solution, and (b) have a better incidental memory for the new beliefs (5) Results similar to 4 (a) and (fc) above also obtain for the two comparable low dogmatic groups but these differences are less striking (6) Other quantitative and qualitative results are to the effect that the high do^atic groups do not "go along" with the experimental situation as much as the low dogmatic groups These findings suggest that it is rejection of the experimental situation which leads to poorer memory for the new beliefs, which, in tum, leads to greater difficulty m integrating these new beliefs into a new belief system From these results it may be concluded that the more closed one's belief-disbelief system the more the reluctance to "empty out" old beliefs (10) in order to replace them with new beliefs Hence the greater inability to integrate seems to be a function of the fact that the new beliefs to be integrated into a new system are not all present in the phenomenal field But when the new beliefs are all together "pumped" into the phenomenal field by artificial-experi-

244

MILTON ROKEAGH

maital means, the integration of these beliefs into a new system is dearly facilitated The studies just ated have to do with the open-to-closed character of belief-disbelief systems Attention will now be directed to another kind of study, which has to do with the notion of the disbelief gradient W e asked ourselves Is it possible to demonstrate objectively the existence of the several disbelief subsystems along a continuum of similarity-dissimilarity to the belief system Is there a systematic empirical relation between sunilanty-dissimilarity and extent of rejection of the several disbelief subsystems' And is magnitude of rejection of each and

every disbelief-system along the similarity continuum a function of the degree of dosedness of the belief-disbdief system? Rokeach and Jensen have recently completdy mdependent studies with two groups one Catholic, the other Methodist • Each of these two groups was m turn subidvided into low and high dogmatic groups By means of a Likert-type scale, measures were obtamed of the extent to which Catholic and Methodist subjects rtj&A. Baptists, Catholics, Episcopalians, Lutherans, Methodists, and Presbyterians. Mean rejection scores for each of the five remaining groups, other than one's own, were obtained using the mean rejection score of one's own group as a base line These rejection scores were then plotted agamst extent of similarity of each of the five disbdief subsystems to one's own belief system, as determined objectively by an expert in the field of comj^rative religion That the expert's judgment IS valid is evidenced by the fact that it correlates 1 (X) with the mean similarity scores of the five groups ranked by all the Catholic subjects, and correlates 90 with the mean similarity scores of the five groups ranked by all the Methodist subjects The results are shown in Figure 1 It will be noted that even though the number of cases are quite small there is a certain regularity in the curves shown in Figure 1— the curves are either U-shaped or J-shaped. For both Catholic and Methodist groujs, there is relatively greater rejection of disbelief subsystems most similar and least similar to their own And in both Catholic and Methodist groups, there is least rejection of dis•A full report ol this research as well as further research subsequently conducted IS in preparatMm

THOUGHT AND BELIEF (Epis) (Luth) (Presb) (Meth) (Bopt 1 2 3 4 5

:;IIO^

6 * - • '"•^^ CATHOLICS High Dogmatic (N -13) _ ^ * ^ « ^ III _ ^ * ' " Total Group (N = 28) ""* * — • - -^Low Dogmatic (N=15)

5

t

4

High Dogmatic (N =13) Total Group (N=27) Low Dogmatic (NM4)

Ul I;

h UJ

2

MAGf

UIN

o 1

METHODISTS

J *

0 — — — — ^ - ^ -1 1 2 3 4 5 (Bopt) (Presb) (Luth) (Epis) (Coth) RANK ORDER OF SIMILARITY Fic. 1. The disbelief gradients for Catfadics and Metho(Ust8.

behef subsystems occupying intermediate positions along the sunilanty continuum. The curves shown m Figure 1, however, are probably truncated curves What if the Catholics were asked, for example, to indicate extent of rejection of ex-Catholics, Jews, Mohammedans and atheists, as well as for the five Protestant denominations' This is presently under investigation and the findings from this study should lead to further insights about the nature of disbelief gradients Another tantalizing problem for further research is to see whether similar results would obtain in the scientific realm Take, for example, a group of Freudians or Hullians, or what have you Would the relation between similarity and rejection of other theoretical positions by adherents to a given theoretical position also be U-shaped or J-shaped in function?

246

MILTON ROKEACH

There is another finding shown m Figure 1 to which attention should be drawn In accord with theoretical expectations, it is seen that for both Catholic and Methodist samples the magnitude of rejection of each and every disbelief subsystem is greater for the high dogmatic groups than for the low dogmatic groups, that is, all 10 sets of comparisons show that high dogmatists reject systems other than their own more than low dogmatists It must be confessed, however, that the thought-belief model is as yet insufficiently developed to permit us to specify m advance the exact shape of the disbelief gradients found But what is to be emphasized at this stage of the research is that the data do show, consistent with the model, that disbelief subsystems do indeed seem to be arranged along a continuum of similarity to one's belief system and that rejection of one particular disbelief subsystem (one outgroup) is not necessarily of the same magnitude as the rej'ection of another disbelief subsystem (another outgroup) Such rejections depend, at least in part, upon the degree of similarity to one's own behef system Hence we are reinforced in our notion that the organization of disbelief systems will have to be studied, in future research, m addition to and independent of the organization of belief systems CONCLUDING REMARKS

To return to the major theme of this paper, namely, the unity of though and behef The research program we are carrying forward regarding the relations between belief, as measured by the Dogmatism Scale, and thought, as measural by cognitive tasks, is being derived solely from the cognitive model which attempts to tie together the organization of thought with that of behef The reason why it has been possible to predict from scores on the Dogmabsm Scale to, for example, the way a person solves problems in the makebelieve world of Joe Doodlebug is that similar properties of beliefdisbelief systems are involved in responses to both sets of tasks The awicepts employed m describing such properties seem to be equally applicable to both hnes of inquiry and permit one to flit with rdative ease from behef to thought—^and back again to belief, so much so that it is becoming increasingly more difficult to delineate where believing ends and thinking begins. The outcome of the cognitive studies discussed in this paper sug-

THOUGHT AND BELIEF

247

gests that the model provides a fertile basis for further research relating belief to thought Thus, we intend to return to Doodleland to test other properties of the thought-belief model In collaboration with M Ray Denny and others research is presently under way m which subjects are first being "indoctrinated" to the point where they feel at home in Joe's world Then Joe's world is being tampered with systematically so that Joe has to figure out other ways to get nourishment, and, more extremely, Joe's world is so manipulated that it becomes intolerable to live in By such expenmental procedures we are studying such things as the thought processes and determinants of "party-line" changes, and the processes and determinants of "disillusionment" with and "defection" from a particular behef-disbelief system It IS further possible to address oneself within the framework of the model presented to a wide variety of other problems relevant to social psychology and personality theory In addition to the research reported here we are also curious about such things as the similarities and differences between the strong, passionately held beliefs of, say, an Einstein and those of a Hitler (41, 42, 57) , the differences and similarities in structure and content of fascist vs communist thought and belief, what effect external threat had historically on the dogmatization of thought and belief of such important ideological systems as Judaism, Catholicism, and communism, and what factors, rational and irrational, operate m the recruitment and entrenchment of scientific "gatekeepers" (36) having such-and-such belief-disbdief systems It has no doubt occurred to the reader that almost nothing has been said in this paper about the determmants of behef-disbelief systems. This is deliberate For the wnter believes that before one can ask questions regarding determinants one must first analyze and define carefully what it is one wishes to know the determinants of The analysis of the structural parameters of thought and belief, the extent to which such parameters are the same, and the measurement of such parameters seem to be logically prior to the analysis of the motivational variables underlying thought and belief. Once this prior task has been satisfactorily accomplished, attention can then be more fruitfully and more holistically directed to the issue of psychodynamic determinants For, the writer suspects, a given motiva-

248

MILTON ROKEAGH

tional variable may turn out to affect simultaneously and systemaUcally a whole class of structural properties, rather than just this and that property, of the cognitive system within which thinking and believing takes place The study of how motivation produces such simultaneous and multiple effects on belief and thought is a major task for future research REFERENCES 1 ADOTNO, T

W , FRENKEL-BRUNSWIK. ELSE, LEVINSON, D

J , & SANFORD, R

N

The authontartan personahty New York Harper, 1950 2 A S C H , S E Social psychology New York Prentice-Hall, 1952 3 BARRON, F Complexity-sunphaty as a personality dimension / abnorm soc Psychol, 1953, 41, 163-172 4 BECKER, W C Perceptual rigidity as measured by aniseikonic lenses J abnorm soc Psychol. 1954, 41, 419-422 5 BLANSHABO, P Communism, democracy and Catholtc power Boston Beacon Press, 1951 6 BLOCK, J , & BLOCK, JEANNE. An investigation of the relationship between intolerance of ambiguity and ethnocentrism / Pers. 1951, II, 303-311 7 Bw)WN, R W A determuiant of the relationship between rigidity and authontanamsm / abnorm soc Psychol. 1953, 41, 469-476 8 BauNEK, J S Personality dynamics and the process of percavmg In R R. Blake and G V Ramsey (Eds ) , Perception An approach to personality New York Ronald, 1951 9 BKtJNEH, J S , MATTER, J , & PAPANEK, MIRIAM L Breadth of learning as a fnnctum of dnve level and mechanization Psychol Rev. 1955, M. 1-10 10 BRUNEK, J S Going beyond the mfonnatton given Paper prepared for Cambridge Conference on Thinking, 1955 11 CuossMAN, R. The god that failed New York Harper, 1949. lZDuNCKEH, K On problem-solving Psychol Monogr. 1945, SI, N o 5 (Whole No. 270) 13 FISHER, J The memory process and certaui psychosocial atutudes, with sp«nal reference to the law of Pragnanz J Pers. 1951, II, 406-420 14 FRANK, L K Tune perspectives / soc Phil. 1939, 4, 293-312 15 FBENKEL-BSUNSWIK, ELSE Dynamic and cognitive categorization of qualitative matenal II Application to interviews with the ethnically prejudiced / Psychol. 1948, IS, 261-271 16 FKENKEL-BRUNSWIK, ELSE Intolerance of ambiguity as an emotional and perceptual personality vanable J Pers. 1949, II, 108-143 17 FRENKEL-BRUNSWIK, ELSE Personality theory and perception In R. R. Blake and G V Ramsey ( E d s ) , Perception An approach to personality N e w York Ronald, 1951 18 FsENKEL-BRtJNSWiK, ELSE Further explorations by a contributor to "the authontanan personahty" In R. Chnstie and M Jahoda ( E d s ) , Studies m the scope and method of "the authoritarian personality " Glencoe Free Press, 1954 19 FROMM, E. Escape from freedom N e w York Rinehart, 1941 a ) GOLDSTEIN, K . , & SHEERER, M Abstract and concrete behavior An expenmental study with special tests Psychol Monogr. 1941, M, N o 239 21 GouGH, H G , & SANFMOB, R N Rigidity as a psychological variable Unpubhshed manuscript Univer of Califonua, Institute of Personality Assessment and Research, 1952 22.HABLOW, H F The fOTmatKMi of l e a m u ^ sets Psychol Rev. 1949, U, S\-6S 23 HoFFra, E The true believer N e w York Harper, 1951 24 HvMAN, H H , & SHEATSLEV, P B "The authoritarian personality"—a methodological critique In R Chnstie and M Jahoda (Eds ) , Studies m the scope and method of "the authontartan personality " Glencoe Free Press, 1954

THOUGHT AND BELIEF

249

25 JOHNSON, D M The psychotegy of thought and judgment New York Harper, 1955 26 KATZ, D , & BRALY, K W Verbal stereotypes and racial prejudice. In G E Swanson, T M Newcomb, and E L Hartley (Eds ) , Readings in social psychology New York Holt, 1952 27 KENOUEB, H H , & ViNEBEKG, R. The acquisition of compound concepts as a function of previous trainmg / exp Psychol, 1954,