On the Making of Textbooks in Social Psychology

Luther Lee Bernard

A recent article on current social psychology [1] emphasizes the supposed non-research character of textbook writing in this field. It also raises some questions regarding the character of texts which I believe the author has not settled. Perhaps, as one of those who have actually written social psychologies within the last decade, I can throw some light upon the question that will not fall wholly within the field of speculative discussion. In the first place, I should not like to have anything I say interpreted as an argument against the experimental method in any field of science. I believe thoroughly in all forms of experimental and quantitative research and it appears to me to be perfectly clear that the most fruitful development of all science lies in these directions. Certainly all of this experimental material that is usable should be ingested and assimilated by the textbook writers in the field of social psychology. But there are also other factors that should not be neglected in the production of textbooks.


In the first place there are two ways of preparing a handbook in social psychology, as in any other science. One is to produce a laboratory manual, citing the more important experimental studies and selecting the most typical simple experiments for repetition in the laboratory. The other method is to summarize the findings of recent and earlier work in social psychology, or as much of it as can be presented or as seems pertinent to the purpose of the author, in a general textbook on the subject. In between these two typical methods of presenting materials in a handbook may be as many intermediate methods as authors may care to utilize. The same possibilities of

( 68) choice are of course also open to teachers of the subject. Which method, or what variation of these two methods, is actually employed by the writer or teacher will of course depend upon (1) the purpose he has in view and (2) whether he is more interested in the exposition of the technique of the method or in the presentation of the results of the methods; that is, of the relevant findings of the science or of a section of the science. In every science we find these two types of workers-those who are concerned primarily, or even exclusively, with the details of methodology, and those who are primarily, or even wholly, interested in the findings of the workers in their field.

It must be recognized that these types are extremes and that each, probably as much as the other, suffers from limitation of viewpoint. The methodologist, especially if he is a strict interpretationist and therefore a rigid experimentalist, is very likely to lose sight of the objectives and products of his science in their larger aspects. He becomes narrowed and perhaps ever more narrowed in his viewpoint, with the result that he is in danger of becoming in the end a mere technician instead of a scientist. I have in mind a biologist who early in his career did some important experimental work on the cultivation of tissues, but who apparently lost the wider view of his work, with the result that he has been almost forgotten as a leader in that field. Other men have been better able to plan research and adjust it to the needs of the various related sciences than he. They have the "outlook" which he lacked or lost. I have been told that this same man, when asked to write a book on the cell declined on the ground that it was too broad a subject for a research man to handle. This is, of course, an extreme case, but it points to a moral. The same moral is indicated by a less extreme example, well known to all teachers of elementary courses in laboratory sciences. Since emphasis upon research methods in the sciences became so strong about thirty years ago, the first courses in most sciences have been turned

( 69) into laboratory courses with some explanatory lectures and reading. The result, it is entirely safe to say, is that the average person who comes through such a laboratory course today does not know enough of the general principles and theory of his subject to use it as a basis for further study in some other field, and most students cannot take more than one year in a basic science. My own experience as a teacher of sociology and of social psychology has been that it is necessary for me to teach nearly all of my students the facts of biology that appear to be essential for work in sociology and social psychology, in addition to teaching them those subjects also. They know how to dissect a frog, they have learned a few names and classifications, but they do not know biology. Apparently the laboratory method has not delivered the goods. My purpose here is not to cast aspersions upon laboratory methods and experimental work. I believe in both, but I wonder if they are not more effective in producing research results than in teaching students who are handicapped by a life limited to three score and ten years the principles of a science.

On the other hand, the teacher who desires to get the basic and general principles of his subject over to students in the limited time at his disposal, finds that he can use verbal conditioning more effectively than laboratory demonstration for this purpose. Of course there is danger that what he gains in the amount of material covered and in scope and organization of principles by the use of verbal or logical conditioning may in part be compensated by loss of vividness and of detailed insight and understanding. This compensating loss does, to some extent, occur and marks the greatest weakness of the old ideological methods of teaching. It may, however, be avoided to some extent by mixing the two methods, using the laboratory demonstration largely for illustration and to make concrete the subject matter.

Essentially the same problems arise in textbook writing as in teaching. Shall the book be written primarily as a

( 70) demonstration in methodology, or primarily for the purpose of getting over to the student the most important facts and principles of social psychology? The answer to this question depends upon the purpose of the writer. If he wishes to train investigators, he may well emphasize methodology. But even with such a purpose in view, he should not neglect to get over to the student a minimum of consistently organized content or findings, a general survey of the field, in order that the future investigator may have sufficient perspective to recognize an important problem in the field and plan a worth-while investigation. On the other hand, the writer may reflect that only very few students are going to be investigators in the science of social psychology, but that perhaps ninety-nine out of each hundred will take the course for the purpose of understanding human nature and functional human and group relationships. In such a case he may prefer to prepare a textbook that sets forth the results of investigation without much reference to the methods by which these results were achieved. As I look over the recent textbooks in social psychology apparently written from this point of view, I do not get the impression, which the writer referred to above evidently has, that most of them have been constructed in entire or even in relative ignorance of research done in the field. I have before me now four such handbooks written by social psychologists with sociological leanings, and I see everywhere abundant evidence of close familiarity with the literature of investigation and constant reference to the newest findings in the field.


I wonder also if this critic may not be laboring under something like a psychological illusion in another respect. Has he allowed himself to judge the extent to which the social psychologists use research materials by the extent to which he finds the experiments reported individually by chapter and page? Here again we must recur to the two ways of writing a textbook referred to above. . . .

( 71) Some writers, especially in a new science, appear to consider the textbook as the defender of a point of view. Others regard it as a summary organization of findings that can be relied upon with a reasonable degree of certainty, or at least tentatively. The former procedure will of course require citations at the bottom of the page and numerous supplementary citations in the text. The latter outlook will usually be content with bibliography at the end of the chapters and at the end of the volume, with such references in the text as are necessary to give credit for specific contributions or to recognize variant viewpoints. It may very easily happen that a textbook constructed along the lines of the second procedure mentioned will be the result of much greater familiarity with investigation in the subject, and will be a much fairer and more representative presentation of the field than will a work constructed along the lines first indicated, although the appearance to the superficial reader, or to one who glances at the pages instead of reading them, may be exactly the reverse.


Another important consideration that does not appear in the criticism referred to is that an adequate social psychology of the second type mentioned above could not now be written from the materials of experimental research, and perhaps never can be written entirely from such sources. A laboratory manual of the first type could of course be constructed at any time from the experimental materials in existence. . . . There is a vast difference between a treatise based on a limited set of data and one that must cover the whole field of behavior which the experimental data only partly represent. A textbook that is intended to be a treatise on social psychology should, it seems to me, cover the whole range of psychosocial processes or behavior adjustments in society. It so happens that as yet there are experimental data for only a small part of that field. Of course, it is conceivable that a strict par-

(72) -tisan of the experimental method might, and probably would, wish to rule out data not obtained by the experimental method, and would prefer to construct a very incomplete textbook, which would from his standpoint be "dependable," rather than seek to have it cover the whole ground at the expense of "experimental accuracy." . . . In the case of social psychology, such a procedure would result in a ludicrously fragmentary treatment.

On the other hand, the opposite procedure of covering the whole field with data of unequal value may expose the writer to the charge of system making at the expense of scientific accuracy. But that is exactly what life does. It cannot wait on the experimental method before it is lived. When experimental data are available for guidance, well and good. If they are not available, the next best data must be used. It would be ridiculous to expect people not to live because they have not experimental data for guidance. Life demands a system of some sort for guidance and completeness. That is why philosophy arose and it is why it will continue until there is something better---experimental data, perhaps-to take its place. It is the business of social science, including social psychology, to explain a phase of life and to give guidance in living, and people who turn to it for such guidance have a right to ask for the best there is available at any time, even though it may not be the best possible interpretation of the field of human behavior. . . .

If one accepts the narrower alternative of putting into a text only those things which conform to one special criterion of validation-say the experimental method-he lays himself open to serious difficulties in addition to the limited character of his product. Who shall decide what is the proper criterion of validation? The experimentalist may naïvely answer that the experimental method is its own justification, because its results can be duplicated by any other worker under the same conditions. I am inclined to think that those who would urge a Methodist, a Catholic, or a Freudian criterion of validation could

(73) make the same argument. If one will put himself in the same attitude of mind as one who urges one or the other of these criteria, he will unquestionably be convinced by the same evidence. Their tests by this token become subjective or traditional, instead of objective and naturalistically measurable. But it must be admitted that they often reach good results through the use of their criteria, and that the experimentalists often draw erroneous conclusions. For myself, I prefer the experimental method, whenever it can be used, because it is nearer to corrected or weighted sensory verification, but I hope I am not naïve enough to think that only experimental sources of guidance to behavior should be used, and that other sources should be discarded, when experimental aid is absent. But the point is that, since the choice of criteria is after all partly a matter of the human equation and never wholly as simple a matter as following the lead of a scientific law, for the experimentalist dogmatically to erect the experimental criterion as final and exclusive is simply in many cases to invite the advocates of the other criteria to erect their own as absolutes. Thus we are developing in this country Catholic science, fundamentalist science, spiritualistic science, Christian science, and, according to some indications, Jewish social science. Possibly such a segmentation of the field of science along lines of cleavage between criteria of validation is to be desired. Possibly it may lead to a practical test of relative utility and promote the survival of the fittest criterion. History, however, does not give much comfort in such matters. If such a theory of the survival of the fittest in the larger sense were true, how could the experimentalist explain the survival of so many cultural absurdities in our day, including the very criteria of validation that oppose his own?

Two direct counts in particular may be brought against a great deal of strictly experimental work as viewed by the social psychologist. Often of necessity the scope of experimental work is too limited to throw much light upon the larger psychosocial processes. Those who would de-

( 74) -pend upon it almost exclusively show a marked tendency to disregard the wider group aspects of behavior, or what we might call collective behavior processes, and to concentrate upon the mechanisms of individual responses in social situations. Obviously, the experimental method is easier to employ in making data for the individual psychologist's type of social psychology than in covering the subject matter of the social psychology that is more affected by the sociologist. The former sees his subject matter through the individual, while the latter must also look at collective behavior from the angle of the group. The difficulty of subjecting groups, and especially indirect contact groups, to experimental controls is obvious enough.

In the second place, the conditions of the experiment are often necessarily artificial and the results obtained warped or distorted. The control of stimulus-response objects, and especially of people, in an experiment is not the same as the control of inanimate objects. The very control process changes their personalities and hence their responses. The most careful checking and computing can not remove this fundamental difficulty. The process of the experiment brings a new and powerful set of conditioning factors into the situation, with the result that the responses of the person who is the subject of the experiment are not necessarily made to the stimuli set for him, but to those set about him as controls. One need only reflect on how differently he behaves in "private situations," when he realizes that he may be observed by strangers, from the way he responds when alone or surrounded only by friends, in order to realize the truth of this observation. Indeed, the fact appears to be well enough known to everybody except those who believe that experimental results should always and under all circumstances be taken at their full face value.

Two other secondary difficulties are also very commonly to be met with in connection with experimental data. Much of the so-called experimental work now offered in the field of social psychology is not such at all, but is in the nature

(75) of hypothetical tests, scales, and measures designed to serve as technique in doing experimental work. Some of the journals are literally filled with such measuring devices which are recommended to give dependable results in testing native I.Q.'s, emotional types, racial differences, musical ability, and various other special abilities, bents, skills, personality traits, etc. Those who have observed these measuring devices at work during the past few years or who have endeavored to use them are not unacquainted with their limitations, or at least should not be. In fact, the literature on this very matter of limitations is not inconspicuous, and some of the most important and most convincing criticisms of these techniques have not themselves been experimental, but logical, observational, and statistical. . . .

The second incidental point here follows directly from the first. In many cases the results of experimental measurements and tests in the field of personality and behavior have little control value until they are interpreted. Do the tests of Negroes and of whites by the same scale indicate "native" differences, different habit patterns, or different environments? Or do they indicate something else? The tests themselves will not tell you, however much you prod them or employ them. The answer to your question is always in your interpretation of your numerical results, and your interpretation is not an experimental procedure. It goes back almost always to common sense, to general observation, or to statistical data. In other words, there are other methods besides experimental ones that must be used by any social psychologist, or in fact by any social scientist, who wishes to present conclusions to students or to the public which may be of use to them.

The three checks upon experimentally obtained results mentioned above are not wholly distinct but are closely related. Common sense is just a term to cover accumulated experience, which is likely to have a wider comparative base to rest upon than any single experiment, but which may of course be very defective in accuracy of in-

(76) -formation or technique. It is based upon observation and statistical computations and generalizations, as well as upon less reliable elements of tradition, belief, rumor, etc. It is not urged that "common sense" is more accurate, in its technique of judgment, even at its best, than experimentation, but it looks at the problem in hand from a vastly wider angle than the conditions of the experiment. Observation and statistical generalization are but different aspects of the same thing, or we would better say statistical generalization is a refined, quantitative form of observation.[2] All of these methods, as well as a judicious application of observation by means of the case method technique, and not infrequently the use of carefully controlled logical interpretation and extension, must be employed in securing data and results for social psychology. They must be used both to check the interpretations of the experimenter and re-interpret his data, and to secure data-especially with regard to the wider or collective aspects of behavior---which the experimenter cannot secure by experiment alone, because he cannot produce an adequate technique. It will do no good for the extreme partisan of the experimental method to rule out of the scope of social psychology those problems and subject matters which cannot be handled by experimentation. Life is broader than a single method, and it is the business of science, including social psychology, to offer the best solutions it is able to the questions posed by life. Otherwise science becomes a relatively meaningless and esoteric thing apart from life.


The charge of the critic that the social psychologies still deal with out-of-date themes, such as mobs, crowds, propaganda, public opinion, suggestion, imitation, and instinct, I take to indicate a failure to recognize that there may be two viewpoints in social psychology, each perhaps equally legitimate. Perhaps the critic is here speaking from the standpoint of the psychologist rather than of the social

( 77) scientist; but I suspect that the social psychology which is being developed by the sociologists and which can be used by the social sciences will ultimately prove the more important and the more widely welcomed of the two brands. . . .

Along with this growing interest in social psychology on the part of the psychologists has gone naturally something of a shifting of interest in the subject matter of such courses. The psychologist, as would be expected, deals with the behavior of the individual, and some psychologists who have manifested an interest in social psychology---F. H. Allport, for example---do not seem to be able to see society, groups, institutions, i.e., collective behavior. Their attention is so fully concentrated upon the individual trees that they not only fail to perceive the forest, but they make elaborate arguments that there is no forest, but only trees. Now this display of myopia, which perhaps is partly the result of too close concentration in experimentation upon the individual, serves very well to illustrate the difference in viewpoint-always present, if not always so extreme-between the psychologist and sociologist or social scientist when looking at the field of social psychology. The psychologist has attempted to make social psychology, like individual psychology, a science of individual behavior. He has shown himself rather intolerant of---or perhaps he has merely failed to understand---the social-science viewpoint, which must necessarily take into account collective behavior.

To the social scientist crowds, mobs, propaganda, public opinion, imitation, suggestion, are very decidedly realities. They are social phenomena that must be taken into consideration and dealt with. The social scientist who specializes in social psychology is not unaware that all of these mass or collective behavior phenomena may be dealt with from the standpoint of the individual behavior mechanisms involved, when the individual unit in the collective behavior process is being considered. Every recent writer of a textbook in social psychology from the sociological or social-

( 77) science standpoint with whose work I am familiar not only understands these individual behavior processes, but also describes them in his book. But, unlike the individual psychologist who writes on social psychology, he does not stop here. He goes on to describe also the collective behavior patterns which are peculiarly the field of interest of social science. Take, for example, the categories "suggestion" and "imitation," which some of the individual psychologists would banish from the vocabularies of the science of social psychology. It is perfectly possible to explain everything that happens in connection with suggestion and imitation in terms of the conditioned response and the conditioning process, as far as the responding individual is concerned. In my own textbook on social psychology I do so explain these behavior processes, but I also recognize that these conditioned responses, which are nothing more to the individual psychologist, frequently occur in such patterns of collective behavior that it becomes necessary to name these collective behavior patterns and to describe this collective behavior. The individual psychologist, with his attention concentrated upon the behaving individual, may fail to see the collective behavior pattern and there fore may deny that it exists. But, really, it is scarcely a good argument to assert that the elephant has no ears because to the blind man, who is in contact only with the trunk, he seems so like a rope.

My surprise is even greater to learn that the subject of instinct is passé in social psychology.[3] I admit that it should be. In fact, it occurs to me that nothing short of the rather patent inability of one of our water-tight-compartment sciences to learn anything of importance from another could explain why any social psychologist could still take seriously a classification of human instincts when a discussion of collective behavior is involved. Here again it was the sociologists, finding it necessary to explain collective behavior rather than be content with speculating about individual behavior, who were foremost in breaking

(78) down this biological superstition inherited from the metaphysical "mental scientists." Curiously, it is the psychologists who have defended the concept even when it was obvious to almost every one else that it was a lost cause. I have been much amused by the deprecatory remarks made by various psychologists commenting on my book, Instinct, and various articles criticizing the concept of instinct appearing since 1921. These remarks range from one sarcastic "Bernard knows" to a disdainful "He is a sociologist." Letters from psychologists whose theories of instinct I had criticized were frequently as interesting. When I see McDougall's more recent books extending, rather than contracting, his list of instincts, and when I see the latest social psychology by a psychologist still explaining human behavior on the basis of a hoary theory of instincts as if nothing had been done in the field, I cannot be convinced that the subject is passé, at least among the psychologists. Even yet most of the textbooks in psychology give the concept good standing. On the other hand, I do not know of a single sociologist of marked reputation who uses the concept to explain collective behavior, unless peradventure through a slip of the pen back to an earlier pattern of thinking. Must we lay this difference between the sociologists and the psychologists in dealing with the concept of instinct to the difference in their orientation towards the fields of social psychology, the one trying to explain everything in terms of individual response and the other insisting upon the necessity of viewing the behavior processes from the collective standpoint and from that of environmental conditioning?

This difference of orientation towards the field and subject matter of social psychology is interestingly illustrated by the experience of a social scientist who decided to add a man to his staff to make psychosocial interpretations of community life. This social scientist is a very deliberate man and endeavors to plan his procedure beforehand with much forethought. He consulted a considerable number of people, including the writer, as to whether he should employ

( 80) a social psychologist trained from the standpoint of the social sciences as well as in psychology, or whether he should entrust his program to a psychologist pure and simple. Of course he received conflicting advice, but finally the influence of the psychologists prevailed, and he announced that he was convinced that, since the studies he wished to have made were psychological, only a man trained as a psychologist could make them successfully. He employed a man recommended to him by the psychologists. After some years of trial he is not recommending the reappointment of the psychologist, not because he is a poor psychologist, but because his training and viewpoint have not prepared him to grasp the investigational problems of collective behavior that are of significance to the social scientists. The solution of difficulties of this type seems to be in the recognition of the legitimacy and the necessity of two types of social psychology, one of them developed from the standpoint exclusively of individual behavior and the other with regard to both individual and collective behavior.

Finally, it seems to me to be expected that the new research data, whether obtained experimentally or by means of case studies, statistical generalization, or even by more general forms of observation and induction, whether interpreted or uninterpreted, should first be utilized in special treatises dealing with such fields as political organization, the press, child welfare, boys' gangs, neighborhood groups, and the like. It is not alone the magnitude of the task which renders it difficult for a textbook in a science to embody all the results of research in a rapidly growing field, but the necessary conservatism of the text prevents an undue hastening of the process. The teachers of textbooks are usually behind the writers. A textbook must not be too far ahead or out of the reach of the teachers. I have seen excellent textbooks fail simply because they were too good, not because they were too poor. It is also necessary not to be too precipitate in embodying research results which soon may be contradicted by new

( 81) data. The difficulty of controlling research processes in social psychology renders the turnover of "fact" in that field somewhat rapid.

That these special treatises should not be called social psychology also is not strange. In the first place, they do not cover the whole field of social psychology. They are named, therefore, after the parts of the field they do cover. There is also a second reason of importance. The newspaper, the popular magazine, the movie, and the radio have so popularized the adulterated thought in our day that the stiffness of a textbook excites a feeling of dread in the emotions of the average intelligent person. Social psychology is a textbook term, and discerning authors keep as far away from textbook titles as possible. I have noted a tendency on the part of textbook writers, editors, and publishers even, to sugar-coat the textbook pills by giving them romantic and dissimulated titles. It is well known that the poor textbooks succeed best. Perhaps those that appear not to be textbooks at all will succeed better---financially speaking.


  1. Gardner Murphy, “A Review of Current Social Psychology,” The Journal of Philosophy, XXVII (1930),pp 435-438.
  2. See "The Development of Methods in Sociology," The Monist, April, 1928, pp. 292-320.
  3. Murphy, op. cit., p. 436.

Valid HTML 4.01 Strict Valid CSS2