To the Editor:
Mr. Petersen, in his article “Prejudice in American Society” (October), first states that social scientists tend to define prejudice primarily as a hostile attitude, then criticizes rather strongly this presumed tendency, and finally argues that prejudice should be viewed as a prejudgment based upon insufficient facts. Whether or not the majority of social scientists hold that prejudice is best conceived of as a hostile attitude I do not [know]. Further, I suspect that Mr. Petersen does not either. He quotes Gordon W. Allport of Harvard in a context which suggests that the latter accepts the purported majority view; it has been my impression, based to a large extent upon a reading of Dr. Allport’s well-known book, The Nature of Prejudice, that he is one of the most distinguished upholders of the view that prejudice should be defined as a prejudgment. . . . [But] Mr. Petersen’s major thesis [is] the conceptualization of prejudice. . . .
Mr. Petersen points out, first, that prejudice literally means prejudgment, and second, that a number of social scientists have forsaken the pristine purity of this definition. Although Mr. Petersen’s observation is undoubtedly accurate, it is also irrelevant. Words . . . arise out of the need of men to summarize some aspect of life which is of relevance to them; that which may be relevant to men in general is not necessarily relevant to social scientists in particular. It is only natural for social scientists to give new meanings to the words they adopt. (One might plausibly argue that Mr. Petersen’s discussion of the meaning of the term “minority group” is based upon his unwillingness to recognize such almost inevitable transformations.)
The basic question, however, is whether or not it is best, for purposes of scientific inquiry, to define prejudice in terms of prejudgment. Ultimately, of course, the justification for any definition rests upon its empirical and theoretical usefulness, a characteristic which is indeed difficult to predict. Yet to define prejudice as a judgment based upon incomplete knowledge is to imply that there are judgments founded upon complete knowledge. Merely to state the proposition thus is to reveal its absurdity. Mr. Petersen notes in his article that many of the generalizations made by the physical sciences, and all of those made by the social sciences, are probability statements. Thus an intellectual elite, possessing both sophisticated techniques for acquiring knowledge and the time to apply them, is far from basing its judgments upon the complete facts. If this is true of the elite, it is certainly true of others, uninterested in abstractions, with limited access to knowledge, and often faced with the necessity of making immediate decisions. The truth of these observations can be illustrated by use of one of Mr. Petersen’s own examples of a relatively specific stimulus—trade unions. I would defy him to obtain any consensus as to either the function or consequences of trade unions in American society from, say, an economist, a sociologist, a psychologist, an employer, and a trade union official. In short, to define prejudice as prejudgment is to define it so broadly that the concept loses all meaning.
I would further argue that even if most judgments could be evaluated in terms of their factual bases, Mr. Petersen’s definition would be of little use. A statement that Negroes are neither as intelligent nor as moral as whites is a good example of a probability generalization. Yet this statement, which if made by a Southern housewife would be indicative of prejudice, would indicate nothing of the sort if made by a Northern social scientist. In short, a statement, whether true or false, is meaningless by itself, and only acquires meaning when seen in the context of the personality making it. . . .
I would agree with Mr. Petersen that the definitions of prejudice prevalent in the social sciences are far from adequate. Even more, I would agree with some of his specific criticisms. At the same time an inadequate definition is not necessarily a useless one. Coherent, logical, and systematic definitions are the prerogative of relatively advanced disciplines. At present the social sciences can afford to, and indeed must, be satisfied with definitions that effectively, if crudely, outline phenomena, the study of which will prove to have both practical and scientific usefulness. It is my feeling that a definition of prejudice in terms of hostile attitudes grasps the essentials of the problem far better than one which emphasizes prejudgment. The recent notable advances which have been made in the study of prejudice by use of the former definition—of which The Authoritarian Personality is perhaps the outstanding example—would suggest that this feeling is not without justification. How warranted it is, of course, only time will tell.
Bureau of Applied Social Research
New York City
Mr. Petersen writes :
Let me suggest that Mr. Budner reread my article in the light of these further comments below, and perhaps he will find that we disagree less than it would seem.
- I did not say, and I did not mean to imply, that all words should be maintained in the “pristine purity” of their etymological meaning. . . . [But I would not] propose that social scientists cry halt to their efforts to delineate more precisely the vocabulary of their profession. Such key words as institution, middle class, culture, and a host of others which are loosely used in common speech, do need to be specified in professional writings. This has nothing to do with whether we define prejudice as hostility. There is a good word to denote a person who feels hostility, namely, hostile; and if we define prejudiced as a synonym, for clarity’s sake we might better dispense with it altogether. To use a single term to denote two concepts, or either one of them, is not a great improvement over lay speech.
- I did not say, and I did not mean to imply, that “a majority” of American social scientists define prejudice as a hostile attitude. . . . I speak of a “tendency,” and cite three examples and say that more could be cited. Other analysts of prejudice, indeed, define it as a pre-judgment. Professor Allport in The Nature of Prejudice is one: but, in the report of the conference I referred to, Professor Allport said what I quoted him as having said.
- “To define prejudice as a judgment based upon incomplete knowledge,” Mr. Budner writes, “is to imply that there are other judgments founded upon complete knowledge.” Several times I made the point that “the” facts are hard to come by, particularly in the social sciences; and if “complete” knowledge is to be interpreted literally, then it is unattainable in any discipline. Shall we, therefore, abandon the empirical base of modern science? How would Mr. Budner react to a physicist or a biologist who informed him that, because human knowledge cannot be perfect, therefore he will use a single term to denote both an empirical and a moral quality? We sociologists call ourselves scientists. If the designation is not a mere honorific, it means that the truths we establish in our field, partial and tentative though they may be, are our authority.
- Mr. Budner writes: “A statement, whether true or false, is meaningless by itself, and only acquires meaning when seen in the context of the personality making it.” . . . If a Southern housewife demonstrates certain attitudes toward Negroes, this is prejudice; but if a social scientist makes the same statement, it would indicate “nothing of the sort.” I would say, in the contrary, that a social scientist can also exhibit prejudice, and a housewife is not inevitably prejudiced. The difference is not innate, but based on the social scientist’s training to look at the facts and report them accurately. I propose that he continue to apply this training also to the area of group prejudice. . . .
- I would agree that the justification of any definition “rests upon its empirical and theoretical usefulness,” but I do not think that in this case one has to try to “predict” this characteristic. One need only look at the record, as to some degree I did in my article, to find that a tool is not useful if it confuses the empirical with the moral. We live in an age of totalitarian horrors, and the stance that any hostile attitude is bad has blinded some social scientists to a portion of the real world. Suppose the Southern housewife believes (as one might expect from such a person) that the Soviet Union is a totalitarian country, not significantly different from Nazi Germany; while social scientists tell him that it is a kind of extension of a social-welfare state. Must we accept the word of this “elite,” or may we look at the facts of Soviet society?
- If I considered The Authoritarian Personality too complex a subject to discuss in the article, I cannot make up this lack here. However, I would strongly urge Mr. Budner to read, or perhaps reread, the brilliant critique of it by Paul Kecskemeti in this journal (“Prejudice in the Catastrophic Perspective,” March 1951).
- In conclusion, let me spell out again the moral point I intended to make in my article. I regard prejudice as bad. I think it is the duty of social scientists to try both to establish empirical truths and to teach their students and the general public to base their judgments on as many of these truths as have been established. On the other hand, I do not feel that hostility is in itself a bad thing. Some persons, some movements, are evil.