Reprinted from Acta Sociologica vol 2 (1957), no. 4, pp 179-201.

[p 179]


by Hans L Zetterberg

We might assume that in social intercourse some persons will describe various events, some will evaluate events, and some Ė perhaps the same persons Ė will prescribe what should be done about them.2 Our task in this paper is to search for a few hypotheses that indicate in what ways descriptions, evaluations, and prescriptions, existing in our social milieu, might influence us. In short, we shall study the acceptance of pronouncements such as "The President of the United States has had a heart attack" (a description), "Imperialists are evil" (an evaluation), and "Participate in the pension plan!" (a prescription). We consider this three-fold delineation essential to the study of compliance.3 We shall assume the existence [p 180] of some descriptions, evaluations, and prescriptions, and omit problems concerning their origin or modification (problems of social change) and problems concerning their allocation to and differentiation into social roles (problems of social structure). This theory, therefore, differs in its focus from Tardeís classical theory in this field4 and from other summaries or theories of studies of social influence that have appeared in recent years.5 Our summary, furthermore, will be organized as an axiomatic theory.6


The field we shall survey lacks a high level of conceptual clarity. For example, the fact that members of a group evaluate each other favorably is in one research tradition called "sociometric score", in another "cohesiveness", in a third "sentiment" or "rank", and in a fourth school of thought it is called "morale". Not only do terms differ among various schools and research traditions, but also within any given tradition there is considerable conceptual haziness. This theory shall choose as building blocks or primitive terms the notions of "actor" and "actions", and among the latter we shall include the above mentioned sub-classes of "descriptions", "evaluations", and "prescriptions". The choice of these sub-classes is not coincidental. It can be demonstrated that the lionís share of sociological concepts can be defined in these terms. Knowledge of the laws of the social use of descriptions, evaluations, and prescriptions can, therefore, be used to derive laws concerning a large number of sociological phenomena.

Considered in terms of building blocks such as descriptions, evaluations, and prescriptions, much current terminology appears less precise. Professor Segerstedt brought to our attention the failures in differentiating prescriptions from prescribed actions.7 For example, we have the term "behavior pattern" which sometimes means patterns of behavior (action) and sometimes means patterns for behavior [p 181] (prescriptions). We read that a "social role" includes such and such a performance (action) and at other times we read that the social role expects that and that action (prescription). In one report we find that the "norm" or "group standard" in one group is a large proportion of college enrolment (action), while in another report we find that there are strong norms or group standards encouraging college attendance (prescription). In one lecture we hear that "folkways" and "mores" are simply behaviors A, B, and C (actions), and in another lecture folkways and mores refer to the fact that A, B, and C ought to take place (prescriptions). Confusion in sociological language between descriptions and evaluations should also be brought to our attention. As a case in point we may take the concept "definition o£ a situation". This term may stand for a description of a situation ("This paper deals with social psychology") and an evaluation of a situation ("This paper is not difficult to understand"). In other contexts we have eagerly utilized the similar distinction between facts and values but not in the formation of this concept. Furthermore, confusions between evaluation and prescription are often found. The term "value", for example, sometimes stands for an evaluation, sometimes for a prescription, and sometimes for both. Behind the label of "opinion", finally, we may find descriptions, evaluations, and prescriptions. A personís opinion about subversive infiltration into government, for example, is likely to consist of his description of the number of subversives, his evaluation of them as more or less dangerous, and his prescription that something should be done about them.

We believe that the fusion of these aspects of action is blocking our progress in formulating and testing hypotheses. There is no reason to believe that activities and prescriptions follow the same scientific laws as our conceptualizations of folkways, mores, etc, suggest. Rather, it may be assumed that prescriptions are among the causes of certain activities. Also, evaluations may not follow the laws of descriptions, nor of prescriptions, but each action-type may have laws of its own.


In view of such considerations let us introduce a series of derived terms that maintain the above distinctions. First we may simply sort out all actions that are alike. These may be called uniform actions. A uniform action is any action that is equal to or synonymous with another action, for example, Mr Xís reading of an editorial and Mr Yís reading of the same editorial, or Mr Xís reading of an editorial one morning and his reading of another editorial the following morning. We might further distinguish between intra-individual and inter-individual uniformities. A dispositional action is a uniform action by one and the same actor, for example, Mr Yorkís habit of voting for the Democrats in every election. An aggregate action is a uniform action by several actors, for example, the Democratic vote of New York City. This delineation between dispositional actions and aggregate [p 182] actions states, in more general terms, Deweyís well-known distinction between habit and customs.8

Aggregate actions are the main topic of sociology.9 Sociological concepts such as status, role, group, institution, class, culture, and so forth represent complexes of aggregate actions. Dispositional actions are the main topic of psychology. Psychological concepts such as trait, ability, drive, personality, and so forth represent complexes of dispositional actions. The border field between these two sciences is social psychology. The topic of this paper falls in this border area.

We might now consider each of our primitive terms denoting behavior as either dispositional or aggregate actions. Thus we arrive at precise definitions of a series of well known concepts. A cognition is a dispositional description, for example, the readerís knowledge that this journal is devoted to sociology. An attitude is a dispositional evaluation, for example, the readerís emerging feeling that this might be a difficult paper. An expectation is a dispositional prescription, for example, a teacher assigning papers on social influence to his students. When we say that

"Acta Sociologica is devoted to sociology", (a cognition) or "This paper is difficult", (an attitude) or "Read Festingerís experiments on social influence for the next class", (an expectation) we are not giving completely unstable responses. We repeat them as the occasion arises. If these statements are random in the sense that one minute we say something and the next minute we say the opposite we cannot call them dispositional actions. Thus the measurements of the reliability and validity of responses by means of checks of homogeneity, unidimensionality, test-retest correlation, split-half correlations, et cetera, upon which our methodology texts insist, become relevant here. If these tests reveal low consistency of response we cannot legitimately say that they measure cognitions, attitudes, expectations, or any other dispositional action.10

The parallel aggregate actions can also be formalized. A social belief (sometimes social knowledge) is an aggregate description, for example, the general knowledge that the Soviet Union has hydrogen bombs. A social value (sometimes social valuation) is an aggregate evaluation, for example, the general feeling that those who would begin an atomic war are evil. A social norm is an aggregate prescription, for example, the generally accepted law prohibiting private individuals from the manufacture of atomic bombs. The above terms differ from the previous one in the same manner that the sentence "I, as chairman, suggest that Mr X be excluded from our club" (an expectation) differs from the sentence "We, as [p 183] members of the central committee, have ruled that Mr X shall be excluded" (a social norm). Perhaps the easiest way to visualize the latter aggregate terms is to think of a table summarizing poll data from a large number of people. When a given percentage of them say, for example, that "all children ought to have a high school education" we have a social norm. When one father repeatedly says to his son that he ought to finish high school we have an expectation.

Usually, aggregate and dispositional actions overlap. The fatherís expectation about education enters as one element in the social norm about school attendance. However, there are also completely idiosyncratic expectations, and panicky aggregate outbursts that have no dispositional counterparts. Moreover, even if we assume that overlap is normal, points of strain will remain. Modern society is far from homogeneous in terms of social beliefs, values, and norms. There will be frequent instances when established cognitions clash with new social beliefs, and when accustomed attitudes encounter new social values, et cetera. Therefore, we must not take the overlap between the aggregate and the dispositional for granted, but rather view it as problematical. This is done when we study compliance, that is, the degree of such overlap.

The sine qua non of compliance is, of course, that the persons who affect each other have (or have had) some kind of contact with one another. As a term denoting one kind of contact we shall choose social visibility. Social visibility is the probability that one actor has cognitions of action by (an)other actor(s). The cognitions we have of each othersí actions in conditions of social visibility need not be scientifically correct. W I Thomasí dictum applies here: if people define a situation as real it becomes real in its consequences.

When conditions of high social visibility exist we might speak of an action system. A set of actions is an action system, if and only if, all actors have high social visibility of these actions When the word "system" is used here an assumption of interdependency is made; we do not call a set of units a system unless changes in some units affect other units. The use of the term action system, therefore, implies that we plan to write specific hypotheses about the relationships of its various parts.

An informal luncheon discussion is an action system: each participant knows what views the others advance. What is important to us here is to note that most laboratory experiments on small groups are experiments with action systems. They are usually not experiments with "social groups" in the complex sense in which sociologists use this term. The notion of action system will make it possible for us to absorb the great value of the experiments for sociological problems without changing the traditional meaning of the concept of social group. It is fruitful to distinguish between one-way visibility and two-way visibility in action systems. [p 184] Some sociologists have talked about a system with one-way visibility as a "mass" and a system with mutual visibility as a "public."11 (Words such as mass-media and "public opinion" refer perhaps to this distinction). In this paper, however, are not going to include any of the variables of social structure, but we shall rather attempt to write a theory for any type of action system.


To be able to write hypotheses about events in an action system we need to make some assumptions about the motivation of the participants. Human motivation is, of course, immensely complex and it is impossible to do justice to all its facets and layers. Experience in other fields shows that simplified assumptions will suffice in macroscopic analyses of human behavior. Thus, economics has been quite successful by assuming only one motive, that is, the profit motive. We shall make a similarly drastic simplification by assuming, not manís desire to maximize profit, but a desire to maintain and to maximize favorable self-evaluations. As a formal postulate this might be stated:

An actorís actions have a tendency to become dispositions that are related to the occurrence of favorable self-evaluations.

This will be called the Postulate of Ego-Needs. It is in one form or another common in modern writing in psychodynamics.12 It is the only assumption of individual psychology we need to make in this paper.

A difficult aspect of this postulate concerns the interpretation of the words "related to". The learning theorists have developed several conceptions as to the occurrence of behavior is "related to" the occurrence of gratifications. Some of these conceptions are simple notions that equate "related to" with temporal proximity. Others are complex ideas that equate "related to" with the functional significance an event may have for need reductions.13 We shall not enter upon a discussion of all these possibilities here. To give one illustration, if we equate "related to" with "compatible with" we obtain the well-known principle that we change our actions in a defensive way to make them compatible with favored self-attitudes.14 In our culture the homosexual, for example, tends to avoid recognition of this homosexuality in order to preserve his self-respect (repression). In addition, he may deny that the homosexual impulse comes from himself, saying by no means, "I am homosexual", but rather "You are homosexual", (projection). Furthermore, he may convert his feeling of affection into its opposite, saying not [p 185] "I love him" but "I hate him" (reaction-formation). Also, he may transfer the latter into "He hates me" (new projection) and thus develop paranoid ideas.15

What the Postulate of Ego-Needs suggests about any actions must, of course, also include the dispositions which we call cognitions. We might now recall that the likelihood that a person has cognitions of the actions of others is labeled the social visibility of these actions. Hence, the Postulate of Ego-Needs involves a special case of social visibility: actions have a tendency to have high social visibility to an actor to the extent that they arc accompanied by his favored self-attitudes. Since actions with high visibility are labeled action systems, we might state this special case in this way:

Actions have a tendency to be part of an action system to the extent they are related to favored self-attitudes among the actors of the system.

This hypothesis is an aspect of a well-known theory of perception. This theory holds that those stimuli that are relevant to need-satisfactions are more likely to be noticed than others.16 It is relevant because it indicates that an action system is not a random selection of the actions that the participants might physically describe. Actions which affect our self-attitudes favorably are over-represented in action systems. Conversely, the more an action lowers our self-attitudes, the less likely it is to be included in our action systems. Thus we have an interpretation of the frequently observed phenomenon of "psychological evasion of propaganda". Prejudiced persons evade the normal interpretations of a series of cartoons that depict them as ridiculous.17 We avoid exposure to voting norms from which we deviate; we read the campaign material from our own party to a larger extent than that of opposing parties, although both find their way to our mailboxes; we look at "our" candidateís speech in the paper, although the other candidateís speech is also in the paper. As Professor Lazarsfeld and his co-workers pointed out after having discovered this in a presidential election:

In recent years, there has been a good deal of talk by men of good will about the desirability and necessity of guaranteeing the free exchange of ideas in the market-place of public opinion ... Now we find that the consumers of ideas, if they have made a decision on the issue, themselves erect high tariff walls against alien notions.18

[p 186]

This tendency toward evasion, that is, the exclusion from our action systems of anything that threatens our self-evaluation, is one reason why attempts at influence may fail.19


We might now begin to formulate some hypotheses about compliance in action systems. These hypotheses summarize past findings and predict future ones. It should be readily acknowledged that each hypothesis if taken by itself is trivial. The findings in this field have not been particularly startling. However, when the implications of all of these hypotheses have been spelled out, many conclusions appear more interesting. The first hypothesis relates uniform evaluations in an action system to a participantís attitudes:

In an action system any actor has a tendency to develop attitudes that are synonymous with uniform evaluations (attitudes and/or social values) in the system.

This shall be called the Postulate of Evaluative Compliance. Among the many studies documenting this tendency is one by Sims and Patrick.20 They ascertained some attitudes toward Negroes among students in the North (Ohio) and among students in the South (Alabama) and found, as expected, that the Northern students had more favorable social valuations of Negroes than the Southern ones. They also located 115 students from the North who had enrolled in psychology classes at the University of Alabama. The students came South as freshmen with attitudes practically the same as those of the Ohio students, but as time passed their attitudes approached those of the typical Southern students, until in the Junior and Senior year, their attitudes did not differ significantly from those of their Southern classmates. These findings are in agreement with the Postulate of Evaluative Compliance. Among other studies which also render support to the Postulate we might mention only an experiment on attitudes toward Russia,21 and a survey analysis on soldiersí evaluations of their readiness for combat.22 Students of social structure also implicitly accord significance to this postulate when they deal with problems related to the acceptance of systems of social stratification. Rank and honor are social evaluations of statuses and roles, and new members in a group normally accept the already existing system of stratification of rank and honor.

[p 187] A particularly interesting special case of the above postulate concerns an individualís attitude toward himself. Let us call any evaluation that has to do with self-respect, self approval, et cetera, by a special term, self-attitude. A self-attitude is any attitude of an actor toward himself and/or toward his actions. We might now state the Postulate of Evaluative Compliance in such a way that it explicitly includes self-attitudes.

In an action system, any actor has a tendency to develop self-attitudes that are synonymous with the uniform evaluations of him that are in the system.

In general systems, the hypothesis suggests that our self-evaluation is dependent on evaluations from others. We recognize this hypothesis (and a following one concerning self-cognitions) as the theory of "the looking-glass self" formalized by Charles H Cooley.23 Let us call it the first Cooley hypothesis. In spite of the fact that the theory of the looking-glass self was formulated over 50 years ago, not much research has been centered on it. Some qualitative support might be found in case histories of religious conversions. The typical convert has been brought into contact with a religious group which evaluates as "sinful" some of the actions his previous groups thought of as normal or acceptable. As he becomes more entrenched in the religious group he comes to evaluate such activities as his "sins". Other qualitative support has been found in observations of the self-image of members of low-ranking ethnic groups. It has been observed that some members of these groups develop "self-hatred".24 They feel hated by society and this valuation is reflected in their self-evaluations. A more direct test is provided by Miyamoto and Dornbusch who asked members of ten student groups to evaluate themselves and all other members of their group in terms of intelligence, self-confidence, physical attractiveness, and likeability.25 A memberís self-rating conformed to the rating of others in 35 out of 40 tests, and it conformed to what he felt to be the rating by others in 40 out of 40 tests. Similar results have been found by Israel, who experimented on students in Stockholm. He finds an agreement between his subjectís estimates of their leadership ability and their estimates of their ability by the other group members, particularly among males.26 This study is also interesting due to its distinction between two types of self-evaluations that are conceptually different but empirically highly correlated.

The First Cooley Hypothesis has also been tested in research on marital success. To appreciate marriage research in this context we have to realize that the hypothe-[p 188]sis implies that the more a personís actions of behavior receiving a favorable attitude or valuation in his action system, the more favorable are the self-attitudes of this person. In a marriage study Ort ascertained, independently for husband and wife, which actions each spouse particularly valued and appreciated.27 He then recorded whether the actions of each spouse were of the kind that the other spouse appreciated. The degree of correspondence was counted, and labeled "marital happiness". It correlated positively with conventional conceptions of this phenomenon. If we make the reasonable assumption that "marital happiness" is the extent to which we derive favorable self-attitudes from the action system called marriage, the findings of Ortís study are in agreement with the First Cooley Hypothesis. Karlsson has developed and tested a theory of marital satisfaction that makes explicit the assumptions in Ortís procedure.28 He also provided a test of the general assumption that these findings hold only to the extent the spouses form an action system with respect to the relevant evaluations. He checked whether each spouse knew what actions the other spouse appreciated. He found that increases in this knowledge Ė that is, increases in the social visibility of these evaluations Ė correlates positively with marital satisfaction.

The Postulate of Ego-Needs suggests that an individualís dispositions develop to the extent that they are related to his favored self-attitudes. We might recall, that the First Cooley Hypothesis maintained that a personís self-attitudes tend to be consequences of the attitudes and social valuations in his milieu. Hence the Postulate of Ego-Needs and the First Cooley Hypothesis render the deduction:

Theorem 1. An actorís actions have a tendency to become dispositions that are related to the occurrence of favored uniform evaluations of the actor and/or his actions in his action system.

This deduction is of paramount importance to sociology since it indicates the ways in which our social environment can manipulate our actions through rewards of favorable evaluations.29 It is, of course, always the case that any theorem has to earn its way and prove its predictive worth. Theorem 1, however, seems to be a strong contender for the position as the major Motivational Theorem in sociology. The maximization of favorable attitudes from others would thus be the counterpart in sociological theory to the maximization of profit in economic theory.

We tested Theorem 1 in a study of participation in student politics. Students who attended the meetings of six religious organizations at the University of Minnesota were asked to fill out a questionnaire that, among other things, contained the questions, "Please mention three things which a member of your organization could do for which the organization would praise him", and "Please mention three things which a member of your organization might do for which the organization would criticize him." The items which mentioned that participation in campus politics were positively valued and that apathy in campus politics were negatively valued were counted. It was found that the Congregational group had a high number of such responses while a Lutheran group had no such response at all. In short, the former group placed a more favorable social valuation on participation in campus politics than the latter group. We tested the consequence predicted by Theorem l through a later questionnaire study of a sample of all Congregationalists and Lutherans on the campus (return: 66 per cent). The questionnaire contained a Guttman-type scale (reproducibility: .91) measuring participation in campus politics. Members of the organizations previously interviewed were identified and compared with regard to political participation. The mean participation score for the Congregational group (N=41), that valued campus politics, was 8.9, while for the Lutheran group (N=72) the score was 6.7 (critical ratio: 5.17). The data thus conform to Theorem 1. Indirect evidence for the same hypothesis can be obtained by juxtaposing different studies. We know that the higher classes value education more than do the lower classes in the United States, and from another study we learn that the higher classes are more likely to send their children to colleges.30 The implications of these simple findings are obvious. It is not enough [p 190] to provide polling places and other aspects of democratic machinery if the social values in the sub-groups of society do not cherish political participation, nor enough to provide university facilities and financial aid to working class students if the social values in the working class are indifferent to higher learning.

The Postulate of Ego-Needs and Theorem l describe rewards in the form of self-attitudes and other peopleís attitudes. Some such rewards seem essential for effective operation of the processes of social influence described in the Postulate of Evaluative Compliance. We know, for example, that some efforts to change prejudiced opinion without any manipulations of rewards have met with failure.31 This might be due to the action systems in which the attempts were made, that is, lectures or films in class-rooms, visits to Negro hospitals. Such action systems are relatively unimportant compared to all other influences in the milieu of the subjects. In such situations the changes might be undetectable unless they are motivated by rewards. Let us, therefore, rewrite the Postulate of Evaluative Compliance take into account the operation of rewards according to the Postulate of Ego-Needs and Theorem 1. The Compliance Postulate suggested that a personís attitudes tend to conform to the repeated evaluations in his environment. Qualify this the Postulate of Ego-Needs, and we have:

Theorem 2 An actorís tendency to have attitudes that are synonymous with uniform evaluations in his action system increases to the extent these attitudes are related to the occurrence of favored self-attitudes.

We have not been able to locate any study that tests this hypothesis. There are, however, many reasons to believe that ego-involvement Ė that is, involvement our self-evaluation Ė in an opinion increases the likelihood that we accept the opinion. For example, one study shows that persons who are manipulated to defend a view tend to accept it more than others.32

If, however, the Compliance Postulate is combined with Theorem l instead of the Postulate of Ego-Needs, we obtain a more testable hypothesis:

Theorem 3. An actorís tendency to have attitudes that are synonymous with uniform evaluations in his action system increases to the extent that favored uniform evaluations of him occur in the system.

Among the evidence supporting this hypothesis is Professor Newcombís study of attitude changes at Bennington College.33 It depicts a campus community in which attitudes and social values are liberal while most students arrive there with [p 191] conservative views. As predicted already by the Postulate of Evaluative Compliance the new students exposed to the liberal climate of the campus tended to change their attitudes toward more liberal ones. While this over-all trend in the data is obvious, it is equally plain that not all students accepted liberalism to the same degree. Additional information allows us to test whether or not Theorem 3 can account for any of the variations in the studentsí acceptance of liberalism. We know the evaluations the students gave each other when they were asked to give the names of those most worthy to represent the College at an intercollegiate gathering. Those who received five or more nominations scored a mean of 60 points on a scale of conservatism, those receiving one to four nominations scored on the average 65 points, and those who where not favored at all in this respect show a mean of 67 points. We see Ė exactly as predicted by Theorem 3, that the more favorably evaluated students were those who accepted more of the values and attitudes of their student community.

A very common way of giving a person a more favorable evaluation is to allow him to associate with persons of higher rank or prestige. In general, we are more influenced in role relationships with high ranking persons than in those with low ranking persons. Evidence for this is found in the large number of studies conducted on so called "prestige suggestions".34 Some studies of this type have also hinted that we might forget the message at a different rate from that with which the reward is forgotten. We obtain then the "sleeper effects" of propaganda.35


Thus far we have dealt with some simple problems of the effects of social influence on our attitudes. Before we look at some additional illustrations to these hypotheses we shall deal with the corresponding effects of influence on our cognitions. Our reasoning here will be parallel to the discussion just completed. Analogous to our previous postulate we have:

In an action system any actor has a tendency to develop cognitions that are synonymous with uniform description (cognitions and/or social beliefs) in the system.

This hypothesis will be called the Postulate of Descriptive Compliance. There is considerable evidence for this hypothesis, particularly if we stress the word tendency in the formulation. Students of social structure are apt to note that peopleís acceptance of status designations follow this rule. Statuses are a special kind of social [p 192] belief, and normally we learn them through the process described in the Postulate. For example, when we call someone "father" it is not because we have first-hand knowledge of our conception but because others talk about a certain person as our father.

Less impressionistic confirmations of the Postulate of Descriptive Compliance have been furnished by experiments on the social psychology of perception. If rephrased in our terminology, the classic experiment by Sherif, testing the effect of social influence on the autokinetic effect, might serve as a first illustration.36 Subjects, separately viewing a light, described its "movements" in a rather consistent manner; that is, they developed cognitions about the length of the "movement". These cognitions differed from person to person. Later, two persons with differing cognitions were asked to describe to length of the "movement" for each other, that is, they joined an action system. Each subject then compromised his earlier cognition in the direction of his experimental mate. Thus each subjects own cognition changed in the direction of the other cognition in his action system as stated in the Postulate. An equally well-known experiment by Asch demonstrates the same point for social beliefs.37 Seven persons describing the length of a line to each other, by prior agreement chose an obviously absurd description in the presence of the subject. In short, they established a social belief contrary to the sensory experience of the subject. Of 50 subjects, 84 per cent yielded to the majority at one time or another during the twelve trials, and reported cognitions in conformity with the absurd social belief. Similar results have been found by Berenda who used children as subjects.38 The tendency to accept social beliefs, indicated in the Postulate, has also been documented by a large number of other investigators, for example by Bovard using descriptions of triangles,39 by Cantril, Gaudet, and Herzog studying a radio broadcast describing a landing by Martians in New Jersey,40 by Jennes who utilized descriptions of the number of beans in a jar,41 and by Luchins, who experimented with descriptions of complex drawings.42 The topics of these studies vary but the relationship persists.

[p 193] A special case of the Postulate of Descriptive Compliance that has captured the imagination of sociologists concerns a personís descriptions of himself. We know that such descriptions often occur: we can readily tell our age, our occupation, our height and weight, our dress, et cetera. Let us call such descriptions self-cognitions. A self-cognition is any cognition that an actor has about himself and/or his actions. Whatever is held to be true about all cognitions must, of course, also be true of self-cognitions. Let us, therefore, rewrite the Postulate of Descriptive Compliance to deal with this special case of self-cognitions. We obtain, then, the Second Cooley Hypothesis:

In an action system any actor has a tendency to develop self-cognitions that are synonymous with the uniform descriptions of him that are in the system.

This theory of the social origin of the self-conception has been tested on medical students.43 Apparently it makes some difference to the self-image of a medical student, if on his visit to the ward, he is greeted by nurses and patients with "Good morning, young man" or "Good morning, doctor". In one school of medicine where first year students have considerable contact with patients it was found that 39 per cent of the students who felt their "patients" thought of them as doctors regarded themselves primarily as doctors, in contrast to 6 per cent of those who felt their patients thought of them as students.

While it is probably possible that cognitions can occur without the presence of any reward ("latent learning") there are many reasons to believe that rewards do speed up the process. If we introduce in the Postulate of Descriptive Compliance the rewards indicated in the Postulates of Ego-Needs and Theorem 1 we obtain the following theorems:

Theorem 4. An actorís tendency to have cognitions that are synonymous with uniform descriptions in his action system increases to the extent these cognitions are related to the occurrence of favored self-attitudes.

Theorem 5. An actorís tendency to have cognitions that are synonymous with uniform descriptions in his action system increases to the extent that favored uniform evaluations of him occur in the system.

An experiment by Kurt Back confirms the latter hypothesis.44 The Back experiment is similar to the Sherif experiment we discussed previously. Subjects were given three rather ambiguous photos and told to write a story about them. To make sure that the stories would differ, some details on the photos varied. When two [p 194] subjects had written their stories, they met in order to tell their versions of three photos to one another, and to discuss them. This resulted in a modification of their descriptions. When they were separated again, and asked to rewrite the stories, the new versions had a greater degree of similarity. So far the experiment reveals only what we have learned from the Sherif experiment and have stated in the Postulate of Descriptive Compliance. However, Back introduced an important variation. Some of the students were given more favorable evaluations than others. For example, some were told that psychological tests about the kind of persons they were and the kinds of persons they liked had been successfully used in the selection of their experimental partners. These students were made to feel that they liked each other very much. Others got their self-attitudes built up through reports that their laboratory instructors considered them superior. The control subjects were not given such favorable evaluations. The results show that those favorably evaluated accepted their partnersí descriptions of the three photos to a greater extent than the control subjects. As Theorem 5 predicted, more social influence on the participantís cognitions was accepted in the action systems where the members received favorable evaluations. We have repeated this experiment Uppsala and replicated this result.


So far, we have dealt with influence from descriptions and evaluations. Let us now turn to influence from prescriptions.

To do this effectively we have to go outside the scope of this theory for a while and consider one origin of evaluations. In the Bank Wiring Room study of the Western Electric Research Program several informal social norms were found about production restriction. In a re-analysis of the findings of the study Homans was able to point out that the workers who conformed to such norms were more popular than others.45 In a similar re-analysis of a study of a street corner gang Homans could demonstrate that those who conformed to the gang norms tended to be more appreciated by the others.46 As a formal postulate this might be formulated the following way:

The more an actorís actions conform to the uniform prescriptions (expectations and/or social norms) in his action system, the more favored the evaluation of him and/or his actions tend to become in the system.47

[p 195]

This principle Ė here called the Homans Postulate Ė has later been documented in a re-analysis of a large number of studies.48

The evaluations that according to the Homans Postulate occur when an individual conforms or deviates from a norm are reflected in his self-attitudes according to the First Cooley Hypothesis. We can therefore write:

Theorem 6. The more an actorís actions conform to the uniform prescriptions in his action system, the more favored his self-attitudes tend to become.

Thus conformity in an action system is rewarded by high self-esteem while deviance is followed by low self-esteem. An illustration is provided in the following case history:

A college student was driving to a distant city to attend a football game. It was the Big Game of the season and represented an important event in the seasonís social festivities. He was accompanied by a girl whose good opinion he valued highly and whom he wished to impress with his extensive plans for a weekend of parties and amusement. They became very gay and hilarious during the course of the drive and he was silently congratulating himself on the successful arrangements he had made. Suddenly a siren sounded behind him and, when he stopped the traffic officer reprimanded him severely and in a very insulting manner for driving like a high-school kid. The sound of the siren and the officerís intrusion immediately destroyed both his rapport with the girl and the happy anticipations he had had.49

When our self-evaluation has been threatened by such accusations of deviance from norms we are likely to resort to defense mechanisms or to evasion. The student in the above case resorted to "aggression"; he berated the manners of the officer and told the girl that the police in that state were notorious for their bullying methods. [p 196] In other cases we might resort to "rationalizations". When we leave a party early we do not say that we are bored when the hostess asks us to stay another short while. We say that we have another engagement, and thus describe ourselves as faith adherents to the norm that appointments shall be kept.

Theorem 6 suggests that conformity leads to more favored self-attitudes. But the Postulate of Ego-Needs maintains that we change our actions to achieve favored self-attitudes. Hence we also change our actions to achieve conformity. This then becomes a new and important Theorem of Normative Compliance:

Theorem 7. An actorís actions have a tendency to become dispositions conforming to the uniform prescriptions of his action system.

This hypothesis is entirely in line with an old but not very explicit sociologic tradition. Most clearly it has been formulated by Professor Segerstedt in his dictum: "Uniform behavior must be regarded as a result of social norms acting as causes."50 To appreciate its significance, one needs to know the historical fact that as late as the 1920ís and 1930ís most social psychologists of stature assumed that an imitation instinct accounted for what we now account for in terms of norms.51

The relationship expressed in Theorem 7 is implicit in much research reporting, but explicit tests are rare. Hall, for example, studied 40 bomber crews that had somewhat different norms as to the behavior of their aircraft commanders, and he found that the commanders tended to behave according to these norms.52 Several investigations of compliance to informal norms designed to restrict production have appeared since the publication of the previously mentioned Bank Wiring Room experiment, for example, a study of a machine shop by Donald Roy.53 In a study of political participation in Uppsala correlations of .57 and .26 were found between a scale measuring perceived expectations from parents that children participate in politics and two scales measuring different political activities of the children.54 In a study of adult education in New York City positive relationships were also found between expectations of college enrollment among family members, friends, high school teachers, and the subjectsí registration in college.55 A test of the assumptions that Theorem 7 holds in an action system is found in a study by Peter Blau.56 In an employment agency new performance statistics were intro-[p 197]duced for interviewers. This meant that the social visibility of the conformity of the interviewers to the rules of the organization was changed. It was found that the interviewers changed their behavior so that those actions that became visible through the new statistical indices conformed more closely to the norm.

In an effort to check the Theorem of Normative Compliance against the older instinct hypothesis, we gathered data from Congregational and Lutheran students at the University of Minnesota concerning their participation in religious classes and clubs during their high school years. A scale of the Guttman-type was constructed to measure the extent to which social norms about religious participation existed in some relevant action systems. Other Guttman scales were constructed to measure the religious participation habits of the studentsí family members and the studentsí friends. The religious participation habits of the subjects and those of their parents correlate most moderately. The habits of the subjects and their friends barely correlate in the case of the Congregationalists (N = 144) and do not correlate at all in the case of the Lutherans (N = 176). The occurrence of this negative case coupled with the nominal correlations in the other case support our decision not to include the imitation hypothesis in our theory. The social norms about participation correlate better with the actual participation. This is also true when the participation habits of the studentsí families and friends are held constant. With a dual covariance adjustment partialling out the effect of family customs and the friendsí customs, the social norms about participation are significantly associated with the subjectsí participation habits (F = 10.73 with 171 x 2 degrees of freedom and E = 13.56 with 138 x 2 degrees of freedom for Congregationalists and Lutherans, respectively).

The process of conformity described in Theorem 7 can be amplified by additional rewards of the kind indicated in the Postulate of Ego-Needs and in Theorem 1. The Postulate of Ego-Needs and Theorem 7 combine into:

Theorem 8. An actorís tendency to have his actions conform to the uniform prescriptions of his action system increases to the extent that he obtains favored self-attitudes.

Theorems 1 and 7 combine into:

Theorem 9. An actorís tendency to have his actions conform to the uniform prescriptions of his action system increases to the extent that favored uniform evaluations of him occur in the system.

An experiment supporting the latter hypothesis has been published by Schachter, Ellerton, McBride, and Gregory.57 They arranged for groups of girls to work on the manufacture of checker boards. Each group consisted of three persons, [p 198] one to cut cardboards, one to paste them on wooden bases, and one to paint the squares on the boards. The cutters were the subjects and each cutter worked in a separate room. The experimenter gave every cutter messages which the cutter believed were sent by her group mates. When these messages were interspersed with speed-up norms, production increased, and when they contained slow-down norms production decreased; this is in accordance with the Theorem of Normative Compliance as we already know. However, Schachter and his co-workers let every second group experience more favorable evaluations from the members and the experimenter according to the technique developed by Bach. We find that favorably evaluated groups receiving speed-up norms increased their production 111 per cent while less favorably evaluated groups receiving the same norm increased their production 83 per cent. The favorably evaluated groups receiving slow-down norms decreased their production 41 per cent, and the less favorably evaluated groups receiving same instructions decreased their production 7 per cent. A replication of the experiment has confirmed the results.58 We see, as suggested in Theorem 9, that favorably evaluated members conform more to the norms than less favorably evaluated members.


If the theory we have presented is of some value it should predict future occurrences of social influence and account for past instances of social influence. The most dramatic account of compliance available in the social science literature is Professor Mertonís study Mass Persuasion.59 It deals with a war bond drive presented by the Columbia Broadcasting System on September 21, 1943. During a span of 18 hours Kate Smith, the radio star, spoke some 60 times begging, cajoling, and demanding that her listeners call their radio stations pledging the purchase of a total of 39 million dollars worth of new bonds. This amazing incident of social influence is described in detail in Mertonís monograph and it provides an acid test for any theory of compliant actions.

At the time of this propaganda marathon, Americans had already been exposed to social norms to buy bonds and most had already done so. The majority felt that they had already conformed to the norm. The first task for Miss Smith was to change this belief. She did this by describing persons who had made far greater sacrifices than her listeners:

Early yesterday morning a man who had lost both legs in the war called in and said he wanted to buy a bond. He wanted to buy a bond with the money he had been saving for years to buy himself a pair of artificial limbs ... making the supreme sacrifice, giving up the dream he had cherished for years, the dream of walking once again. As he said himself: My limbs can wait, [p 199] but this war canít ... What sacrifice are you or I or any of us making that would in any way compare with the self-sacrifice of this magnificent person?60

No less than 20 per cent of the themes of Miss Smith dealt with such civilian sacrifices. Her listeners, we predict from the Postulate of Descriptive Compliance, had a tendency to change their cognitions from "my bond purchases conform to the norm" toward "my bond purchases do not conform to the norm".

When we learn that we do not conform to the norm, we know from Theorem 6 that our self-attitudes become less favorable. Kate Smith in a variety of additional ways further lowered the self-attitudes of her audience. Themes of the sacrifice of servicemen constitute 26 per cent of her appeals:

Now they are braving the swamps and jungles, risking illness and wounds, pain and death. . . staking their lives so that you and I may never know the horrors of a blitz or a bombing. . . nor the tragedy of torture and deliberate starvation ...

Are you backing the attack? Are you really, honestly now, in your heart seeing to it that [servicemen] have the best fighting equipment and plenty of it? Are you buying luxuries or are you turning those dollars you donít need into war bonds?61

This is a common theme in war time. Servicemen receive a most favorable evaluation, civilians a more dubious evaluation. The First Cooley Hypothesis suggests that these were accepted as self-evaluations: Miss Smithís audience was mostly civilian and her reminder of this gave them of lower self-attitude.

However, when our self-attitudes are lowered we become defensive, as we noted in the discussion of Theorem 6 and in considering the problem of evasion. There is a strong temptation to turn off the radio, or to think in terms of projections or reaction-formations: "I have done my share for the war effort; what has the radio star herself done? I have certainly done more than she?" But Kate Smith stressed her own sacrifice in being on the radio for 18 hours in a row:

Hello, everybody, this is Kate Smith again ... Let me tell you it has been a long grind ... sitting here since eight oíclock yesterday morning urging each and every American to join Columbiaís great war bond drive. . .62

Miss Smith interspersed statements about her own sacrifice at different points; 5 per cent of her themes deal with it. Quite unwittingly it was likely to serve as an effective antidote toward the aggression that she aroused against herself when [p 120] she told the audience that they did not conform to the norm or that they ranked lower than servicemen.

Lowering the self-attitudes of the audience provided the immediate motivation for action. Many remaining themes deal with the channeling of this motivation into bond purchases. They consist of prescriptions as to what to do:

Surely, if a legless veteran can give up his dream of a life-time, then we can give a little extra money to buy another bond. Donít delay, call Circle 6-4343 and give WABC [sic] your order for the highest bond you can afford, or even more than you can afford. Will you buy a bond?63

This simple prescription, to go to the telephone and call in the order, constitutes 7 per cent of the themes. We expect the Theorem of Normative Compliance to operate here, and the response was, as previously mentioned, resounding: 39 million dollars worth of bond pledges. The telephone pledge permitted action at the very moment when the listener was most fully motivated.

After the purchases, the acts conforming to the norm, there was strong evidence of catharsis. One interviewee reports:

After I called up, I felt good. I felt I had done something real on the phone.64

As Theorem 6 indicates: conformity to the norms enhances the actorís self-esteem. More could be said about the relevance of other data from Professor Mertonís book for our theory. The above might, however, suffice to illustrate our point: in the theory we have presented each hypothesis taken by itself may seem trivial, but the joint implications of several hypotheses render predictions and explanations that were previously unknown.

Further considerations as to the diffusion of descriptions, evaluations, and prescribed actions need to take structural factors into account. First among these is the extent of an action system or a chain of such systems. This sets the ultimate limits for diffusion, as the anthropologists early discovered.65 Second, types of contacts in the system determine the rate of diffusion. For example, the type of relationship a doctor has to his colleagues will affect his rate of acceptance of a new drug.66 Two simple mathematical models illustrate the range of possibilities of diffusion due to types of contact. If an item reaches a constant number of persons per time unit, independent of the number who have already accepted the new item, we obtain the simple case of linear diffusion (provided all rewards equal). If, on the other hand, all persons in the system have an equal probability of communicating the item to one another per time unit we have the case of logistic diffusion (provided all rewards [p 201] are equal). This is the limiting case of diffusion in a public in which all members talk to a random selection of other members. Third, the rate of interaction will be an indispensable factor in predicting the speed of the compliance. All three facts are primarily determined by the social structure at hand. The theory of compliance will perhaps advance most when it eventually can be tied to a theory of social structure.

In conclusion it might be said that most of the time, we find our hypotheses and their implications supported by findings from several studies. Each study might have its defects, but considered as cumulative evidence67 for our two Postulates of Compliance, the Postulate of Ego-Needs and the Homans Postulate of the origin of evaluations, the studies reviewed give considerable empirical backing to the theory.


1. Clerical assistance for the preparation of this article was furnished through a grant from the Council for Research in the Social Sciences at Columbia University. The original research data reported in connection with the presentation of Theorems 1 and 7 were gathered with the support of the Laboratory for Research i Social Relations at the University of Minnesota, which also financed the replication of the experiment in support of Theorem 5. Earlier formulations of the theory of this paper have been made in the author's M A Thesis at the University of Minnesota entitled "A Semantic role Theory" (1951) and revised in his F L Thesis at Uppsala University with the same title (1952). An impetus to take up the problem again was given by a grant from the Ford Foundation to the Bureau of Applied Social Research at Columbia University to explore the utilization of propositions in social theory in the solution of practical problems, that is, the extent to which we can make valid recommendations without conducting original research. In reviewing this issue, it appeared that what is at present considered "social theory" does not normally contain explicit propositions and summaries of research in their support. An exception to this is found in social psychological theory (see references in footnote 5). This paper presents one area of findings in social psychology in a format believed to render reasonably precise and reasonably well confirmed propositions. The hope is that a similar summary of findings concerning structural processes in organizations and societies will eventually be possible. The difficult problem as to how propositions of this kind (and of the kinds found in other formats of social theory) are to be translated into practical recommendations is not treated in this paper.

2. Morris, C., Signs, Language, and Behaviour. New York: Prentice-Hall, 1946, p 62.

3. A fourth possible topic is the effect of emotional actions in our environment, for example during crowd conditions. However, at present we seem to know too little about such phenomena to include them in any stringent theory. Cf Brown, R W., "Mass phenomena", in Handbook of Social Psychology (ed G Lindzey). Cambridge: Addison-Wesley, 1954, pp 833-876.

4. Tarde, G., Les lois de líimitation (rev ed). Paris: Felix Alcan, 1895.

5. For example, Cartwright, D., "Some Principles of Mass Persuasion", in Human Relations, 2, p 253-267. 1949; Festinger, L., "Informal Social Communication", in Psychological Review, 57. 1950 p 271-282; Festinger, L.,"A Theory of Social Comparison Processes", Human Relations, 7 (1954), pp 117-140; E Katz and P F Lazarsfeld, Personal Influences, Glencoe, Free Press, 1955, pp 43-133; R K Merton and A Kitt, "Contributions to the Theory of Reference Group Behaviour", in R K Merton and P F Lazarsfeld, eds, Continuities in Social Reseach: Studies in the Scope and Method of "The American Soldier", Glencoe, Free Press, 1950, pp 40-105; T W Newcomb, "An Approach to the Study of Communicative Acts". Psychological Review, 60 (1953), pp 393-404; H W Riecken and G C Homans, "Psychological Aspects of Social Structure", in G Lindzey op cit pp 786-832; R Rommetweig, Social Norms and Roles, Oslo, Akademisk Forlag, 1953, pp 11-94; and I Sarnoff, D Katz, and C McClintock, "Attitude-Change Procedures and Motivating Patterns", in D Katz et al, eds, Public Opinion and Propaganda, New York, Dryden, 1954, pp 305-312.

6. H L Zetterberg, On Theory and Verification in Sociology, Stockholm, Almquist och Wiksell, and New York, Tressler Press, 1954, ch 2.

7. T T Segerstedt, Die Macht des Wortes, ZŁrich, Pan-Verlag, 1947, pp 17-25.

8. J Dewey, Human Nature and Conduct, New York, holt, 1922, part I, sec 2.

9. T T Segerstedt, Social Control as a Sociological Concept, Uppsala, Lundequistska, 1948, p 3.

10. T T Segerstedt, "Some Assumptions in Attitude Research, Theoria, 17 (1951), pp 226-239, and H L Zetterberg, op cit, p 41.

11. For example, H Blumer, "The Mass, the Public, and Public Opinion", in A M Lee, ed, New Outline of the Principles of Sociology, New York, Barnes and Noble, 1946, pp 185-193.

12. A very explicit statement is found in P Lecky, Self-Consistency: A Theory of Personality, New York, Island Press, 1945, p 82.

13. Cf J Olds, The growth and Structure of Motives, Glencoe, Free Press, 1956.

14. Cf A Freud, Das Ich und die Abwehrmechanismen, London, Imago, 1936.

15. R R Sears, Survey of Objective Studies of Psychoanalytical Concepts, New York, Social Science Research Council, 1943, ch 7.

16. D Krech and R Crutchfield, Theory and Problems of Social Psychology, New York, McGraw-Hill, 1948, pp 87-94, or, L Postman, "Toward a General Theory of Cognition", in J H Rohrer and M Sherif, eds, Social Psychology at the Cross Roads, New York, Harper, 1951, pp 242-272.

17. E Cooper and M Jahoda, "The Evasion of Propaganda: How Prejudiced People Respond to Anti-Prejudice Propaganda", Journal of Psychology, 23 (1947), pp 15-25.

18. P F Lazarsfeld, B Berelson and H Gaudet, The Peopleís Choice, 2nd ed, New York, Columbia University Press, 1948, pp 90-91.

19. H H Hyman and P B Sheatsley, "Some Reasons Why Information Campaigns Fail", Public Opinion Quarterly, 11 (1947), pp 413-423.

20. V M Sims and J R Patrick, "Attitude toward the Negro of Northern and Southern College Students", Journal of Social Psychology, 7 (1936), pp 192-204.

21. R L Orden, "Interaction between Attitude and the Definition of the Situation in the Expression of Opinion", American Sociological Review, 17 (1952), pp 50-58.

22. M B Smith, "The Combat Replacement", in S Stouffer et al, The American Soldier: Combat and its Aftermath, Princeton University Press, 1949, pp 242-289.

23. C H Cooley, Human Nature and the Social Order, New York, Scribner, 1902, pp 183-185.

24. K Lewin, Resolving Social Conflicts, New York, harper, 1948, ch 12.

25. S F Miayamoto and S M Dornbusch, "A Test of Interactionists Hypotheses of Self-Conception", American Journal of Sociology, 61 (1956), pp 399-403.

26. J Israel, Self-Evaluation and Rejection in Groups, Stockholm, Almqvist och Wiksell, 1956, p 233.

27. R S Ort, "A Study of Role Conflicts as Related to Happiness in Marriage", Journal of Abnormal and Social Psychology, 45 (1950), pp 691-699.

28. G Karlsson, Adaptability and Communication in Marriage, Uppsala, Almqvist och Wiksell, 1951.

29. It should not be overlooked that a person may be much more eager to receive a favorable evaluation from one quarter to another. Praise from the expert is more pleasing than praise from the layman. The group whose norms of evaluation are used when an actor formulates his self-attitudes seems to be what is usually called his "reference group". See, H H Hyman, The Psychology of Status, Archives of Psychology, no 269, 1942, and R K Merton and A S Kitt, op cit. (Sometimes there is a tendency to call any influential action system a reference group. For a discussion of this, see H H Kelley, "Two Functions of Reference Groups", in G Swanson et al, Readings in Social Psychology, New York, Holt, 1952, pp 410-414. The idea of reference groups introduces an important additional consideration to Theorem 1: the norms for assessing increases or decreases in self-evaluation may not be norms of the action system which accords our subjects a more or less favored valutaion. The immigrant to the United States who finds that his job is so highly evaluated and paid for that he can afford to have a car may experience a sizeable boost in self-esteem. His fellow worker who receives the same wage, and also buys a car experiences only a moderate boost in self-esteem upon this purchase. The reason is obvious: the immigrant assesses the increment the car brought to his self-esteem according to norms of evaluation from the old country where the car was an upper class symbol. His native fellow worker assesses the increment in terms of norms of evaluation in a country in which practically every family has a car. It becomes an important question, hence, to spell out in specific hypotheses the factors which determine a personís reference group. This is a task in which social theory so far has not been successful. For a discussion of some partial problems involved in this task - the problems of dimensions and of points of anchorage for scales of evaluation - see, R Rommetweig, op cit, pp 11-18.

30. H H Hyman, "The Value Systems of Different Classes: A Social Psychological Contribution to the Analysis of Stratification", in R Bendix and S M Lipset, eds, Class, Status, and Power, Glencoe, Free Press, 1953, pp 426-442, and Yearbook of Education, London, Evans, 1950, p 635.

31. A M Rose, Studies in the Reduction of Prejudice, Chicago, American Council on Race Relations, 1948.

32. I L Janis and T B Kings, "The Influence of Role Playing on Opinion Change", Journal of Abnormal and Social Psychology, 49 (1954), pp 211-218.

33. T M Newcomb, Personality and Social Change, New York, Dryden, 1943.

34. S E Asch, "The Doctrine of Suggestion, Prestige, and Imitation in Social Psychology", Psychological Review, 55 (1948), pp 250-276, and H B Lewis, Studies in the Principles of Judgments and Attitudes IV: The Operation of "Prestige Suggestion", Journal of Social Psychology, 14 (1941), pp 229-256.

35. C I Hovland, A A Lumsdaine, and F D Sheffield, Experiments on Mass Communication, Princeton, Princeton University Press, 1949, pp 188-200.

36. M Sherif, The Psychology of Social Norms, Haper, 1936. It is typical of the conceptual confusion in this field that Sherif uses the term "norm" to denote converging descriptions.

37. S E Asch, "Effects of Group Pressure upon the Modification and Distortion of Judgments", in G E Swanson et al, op cit, pp 2-11.

38. R W Berenda, The Influence of the Group on the Judgments of Children, New York, Kingís Crown, 1950.

39. E W Boward, Jr, "Group Structure and Perception", Journal of Abnormal and Social Psychology, 46 (1951), pp 398-405.

40. H Cantril, H Gaudet, and H Herzog, The Invasion from Mars, Princeton, Princeton University Press, 1940.

41. A Jennes, "The Role of Discussion in Changing Opinion Regarding a Matter of Fact", Journal of Abnormal and Social Psychology, 27 (1932), pp 279-296.

42. A S Luchins, "Social Influences on Perception of Complex Drawings", Journal of Social Psychology, 21 (1945), pp 257-273.

43. M J Huntington, "The Development of a Professional Self-Image among Medical Students", Paper read at the Meeting of the American Sociological Society in Washington D C in September 1955.

44. K Back, "Influence through Social Communications", Journal of Abnormal and Social Psychology, 46 (1951), pp 9-23.

45. G C Homans, The Human Group, New York, Harcourt Brace, 1950, pp 140-144.

46. Ibid, pp 179-181.

47. The Homans Postulate carries an implication that might be stated in a way that might be of interest to philosophers. In any action system in which there are uniform prescriptions there also tend to be evaluations which are favorable to the conformist or conforming action, and unfavorable to the deviant and deviating action. In short, wherever people say "Do X!" they also say "Those who do X are good" or "X is good" and "These who donít do X are bad" or "Non-X is bad". We make a point of mentioning this only because it is of relevance to the long debate in philosophy s to the nature of value judgments. Some have held that prescriptions and value judgments are identical. For example, Carnap once held that the value statement "Killing is evil" is equivalent to the prescription "Do not kill!" (R Carnap, Philosophy and Logical Syntax, London, Kegan Paul, 1935, pp 23-24). Such view, in our opinion, is not to be settled by philosophical fiats but is best considered as a hypothesis to be tested. The Homans Postulate represents such a hypothesis that predicts a tendency for such prescriptions and value judgments to occur jointly in actions systems. This, however, does not mean that prescriptions and evaluations are equivalent, synonymous, or identical. They might well have different causes and consequences. For example, there is no reason to believe that prescriptions cause self-attitudes in the same way that the First Cooley Hypothesis assumes that evaluations cause self-attitudes.

48. H W Riecken and G C Homans, op cit. In some instances Riecken and Homans seem to force the hypothesis upon data that deal with things other than conformity to norms. For example, the findings by Lundberg and Lawsing showing a positive correlation between popularity and sociometric rank have no direct relevance to the hypothesis, nor do Whyteís findings that higher ranking kitchen workers in a restaurant work on more prestigious food.

49. J Dollard, et al, Frustrations and Aggression, New Have, Yale University Press, 1939, p 12.

50. T T Segerstedt, Social Control as a Sociological Concept, op cit, p 23.

51. N E Miller and J Dollard, Social Learning and Imitation, New Haven, Yale University Press, 1941, pp 289-318.

52. R L Hall, "Social Influence on the Aircraft Commanderís Role", American Sociological Review, 20 (1955), pp 292-299.

53. D Roy, "Quota Restriction and Goldbricking in a Machine Shop", American Journal of Sociology, 52 (1952), pp 427-442.

54. H L Zetterberg, "Some Institutional Interrelations in Sweden", unpublished paper.

55. H L Zetterberg, "A College for Adults", (mimeographed paper) Bureau of Applied Social Research, New York, 1956, pp 19-30.

56. P M Blau, The Dynamics of Bureaucracy, Chicago, University of Chicago Press, 1955, pp 34-44.

57. S Schachter, N Ellerton, D McBride, and D Gregory, "An Experimental Study of Cohesiveness and Productivity", Human Relations, 4 (1951), pp 229-238.

58. L Berkowitz, "Group Standards, Cohesiveness, and Productivity", Human Relations, 7 (1954), pp509-519.

59. R K Merton, Mass Persuasion, New York, Harper, 1946.

60. Ibid p 53.

61. Ibid pp 52-53.

62. Ibid p 54.

63. Ibid p 53.

64. Ibid p 55.

65. A Goldenweiser, "The Principle of Limited Possibilities in the Development of Culture", Journal of American Folklore, 26 (1913), pp 259-290.

66. H Mentzel and J Coleman, "The Flow of Scientific Information in the Medical Profession", (mimeographed paper) Bureau of Applied Social Research, New York, 1954 and 1955.

67. Cf H L Zetterberg, On Theory and Verification in Sociology, op cit pp 20-21.