What Is Wrong with Social Theory[1]
Herbert Blumer
University of California
MY concern is limited to that form of social theory which stands or presumes to stand as a part of empirical science.[2]
The aim of theory in empirical science is to develop analytical schemes of the empirical world with which the given science is concerned. This is done by conceiving the world abstractly, that is in terms of classes of objects and of relations between such classes. Theoretical schemes are essentially proposals as to the nature of such classes and of their relations where this nature is problematic or unknown. Such proposals become guides to investigation to see whether they or their implications are true. Thus, theory exercises compelling influence on research—setting problems, staking out objects and leading inquiry into asserted relations. In turn, findings of fact test theories, and in suggesting new problems invite the formulation of new proposals. Theory, inquiry and empirical fact are interwoven in a texture of operation with theory guiding inquiry, inquiry seeking and isolating facts, and facts affecting theory. The fruitfulness of their interplay is the means by which an empirical science develops.
Compared with this brief sketch of theory in empirical science, social theory in general shows grave shortcomings. Its divorcement from the empirical world is glaring. To a preponderant extent it is compartmentalized into a world of its own, inside of which it feeds on itself. We usually localize it in separate courses and separate fields. For the most part it has its own literature. Its life-line is primarily exegesis—a critical exami-
( 4) -nation of prior theoretical schemes, the compounding of portions of them into new arrangements, the translation of old ideas into a new vocabulary, and the occasional addition of a new notion as a result of reflection on other theories. It is remarkably susceptible to the importation of schemes from outside its own empirical field, as in the case of the organic analogy, the evolutionary doctrine, physicalism, the instinct doctrine, behaviorism, psychoanalysis, and the doctrine of the conditioned reflex. Further, when applied to the empirical world social theory is primarily an interpretation which orders the world into its mold, not a studious cultivation of empirical facts to see if the theory fits. In terms of both origin and use social theory seems in general not to be geared into its empirical world.
Next, social theory is conspicuously defective in its guidance of research inquiry. It is rarely couched in such form as to facilitate or allow directed investigation to see whether it or its implications are true. Thus, it is gravely restricted in setting research problems, in suggesting kinds of empirical data to be sought, and in connecting these data to one another. Its divorcement from research is as great as its divorcement from its empirical world.
Finally, it benefits little from the vast and ever growing accumulation of "facts" that come from empirical observation and research inquiry. While this may be due to an intrinsic uselessness of such facts for theoretic purposes, it also may be due to deficiency in theory.
These three lines of deficiency in social theory suggest that all that is needed is to correct improper preoccupations and bad working practices in theorizing. We hear repeatedly recommendations and injunctions to this effect. Get social theorists to reduce drastically their preoccupation with the literature of social theory and instead get in touch with the empirical social world. Let them renounce their practice of taking in each other's washing and instead work with empirical data. Let them develop their own conceptual capital through the cultivation of their own empirical field instead of importing spurious currency from alien realms. Get them to abandon the practice of merely interpreting things to fit their theories and instead test their theories. Above all, get them to cast their theory into forms which are testable. Have them orient their theory to the vast bodies of accumulated research findings and develop theory in the light of such findings.
These are nice injunctions to which all of us would subscribe. They do have a limited order of merit. But they neither isolate the problem of what is basically wrong with social theory nor do they provide means of correcting the difficulties. The problem continues to remain in the wake of studies made with due respect to the injunctions. There have been and there are many able and conscientious people in our field, alone, who have sought and are seeking to develop social theory through careful, sometimes meticulous preoccupation with empirical data—Robert E. Park, W. I. Thomas, Florian Znaniecki, Edwin Sutherland, Stuart Dodd, E. W. Burgess, Samuel Stouffer, Paul Lazarsfeld, Robert Merton, Louis Wirth, Robin Williams, Robert Bales and dozens of others who equally merit mention. All of these people are empirically minded. All have sought in their respective ways to guide research by theory and to assess their theoretical propositions in the light of empirical data. Practically all of them are familiar with the textbook canons of empirical research. We cannot correctly accuse such people of indifference to the empirical world, or of procedural naivete, or of professional incompetence. Yet their theories and their work are held suspect and found wanting, some theories by some, other theories by others. Indeed, the criticisms and counter-criticisms directed to their respective work are severe and box the compass. It is obvious that we have to probe deeper than the level of the above injunctions.
In my judgment the appropriate line of probing is with regard to the concept. Theory is of value in empirical science only to the extent to which it connects fruitfully with the empirical world. Concepts are the means, and the only means of establishing such connection, for it is the concept that points to the empirical instances about which a theoretical proposal is made. If the concept is clear as to what it refers, then sure identification of the empirical instances may be made. With their identification, they
( 5) can be studied carefully, used to test theoretical proposals and exploited for suggestions as to new proposals. Thus, with clear concepts theoretical statements can be brought into close and self-correcting relations with the empirical world. Contrariwise, vague concepts deter the identification of appropriate empirical instances, and obscure the detection of what is relevant in the empirical instances that are chosen. Thus, they block connection between theory and its empirical world and prevent their effective interplay.
A recognition of the crucial position of concepts in theory in empirical science does not mean that other matters are of no importance. Obviously, the significance of intellectual abilities in theorizing, such as originality and disciplined imagination, requires no highlighting. Similarly, techniques of study are of clear importance. Also, bodies of fact are necessary. Yet, profound and brilliant thought, an arsenal of the most precise and ingenious instruments, and an extensive array of facts are meaningless in empirical science without the empirical relevance, guidance and analytical order that can come only through concepts. Since in empirical science everything depends on how fruitfully and faithfully thinking intertwines with the empirical world of study, and since concepts are the gateway to that world, the effective functioning of concepts is a matter of decisive importance.
Now, it should be evident that concepts in social theory are distressingly vague. Representative terms like mores, social institutions, attitudes, social class, value, cultural norm, personality, reference group, social structure, primary group, social process, social system, urbanization, accommodation, differential discrimination and social control do not discriminate cleanly their empirical instances. At best they allow only rough identification, and in what is so roughly identified they do not permit a determination of what is covered by the concept and what is not. Definitions which are provided to such terms are usually no clearer than the concepts which they seek to define. Careful scrutinizing of our concepts forces one to recognize that they rest on vague sense and not on precise specification of attributes. We see this in our common experience in explaining concepts to our students or outsiders. Formal definitions are of little use. Instead, if we are good teachers we seek to give the sense of the concept by the use of a few apt illustrations. This initial sense, in time, becomes entrenched through the sheer experience of sharing in a common universe of discourse. Our concepts come to be taken for granted on the basis of such a sense. It is such a sense and not precise specifications that guides us in our discipline in transactions with our empirical world.
This ambiguous nature of concepts is the basic deficiency in social theory. It hinders us in coming to close grips with our empirical world, for we are not sure what to grip. Our uncertainty as to what we are referring obstructs us from asking pertinent questions and setting relevant problems for research. The vague sense dulls our perception and thus vitiates directed empirical observation. It subjects our reflection on possible relations between concepts to wide bands of error. It encourages our theorizing to revolve in a separate world of its own with only a tenuous connection with the empirical world. It limits severely the clarification and growth that concepts may derive from the findings of research. It leads to the undisciplined theorizing that is bad theorizing.
If the crucial deficiency of social theory, and for that matter of our discipline, is the ambiguous nature of our concepts, why not proceed to make our concepts clear and definite? This is the nub of the problem. The question is how to do this. The possible lines of answer can be reduced a lot by recognizing that a great deal of endeavor, otherwise conscientious and zealous, does not touch the problem. The clarification of concepts is not achieved by introducing a new vocabulary of terms or substituting new terms—the task is not one of lexicography. It is not achieved by extensive reflection on theories to show their logical weaknesses and pitfalls. It is not accomplished by forming or importing new theories. It is not achieved by inventing new technical instruments or by improving the reliability of old techniques—such instruments and techniques are neutral to the concepts on behalf of which they may be used. The clarification of concepts does not come from piling up
( 6) mountains of research findings. As just one illustration I would point to the hundreds of studies of attitudes and the thousands of items they have yielded; these thousands of items of finding have not contributed one iota of clarification to the concept of attitudes. By the same token, the mere extension of research in scope and direction does not offer in itself assurance of leading to clarification of concepts. These various lines of endeavor, as the results themselves seem abundantly to testify, do not meet the problem of the ambiguous concept.
The most serious attempts to grapple with this problem in our field take the form of developing fixed and specific procedures designed to isolate a stable and definitive empirical content, with this content constituting the definition or the reference of the concept. The better known of these attempts are the formation of operational definitions, the experimental construction of concepts, factoral analysis, the formation of deductive mathematical systems and, although slightly different, the construction of reliable quantitative indexes. Although these attempts vary as to the kind of specific procedure that is used, they are alike in that the procedure is designed to yield through repeated performances a stable and definitive finding. A definition of intelligence as being the intelligence quotient is a convenient illustration of what is common to these approaches. The intelligence quotient is a stable and discriminating finding that can be checked through a repetition of clearly specified procedures. Ignoring questions as to the differential merit and the differential level of penetration between these approaches, it would seem that in yielding a specific and discriminating content they are the answer to the problem of the ambiguous concept in social theory. Many hold that resolute employment of one or the other of these methods will yield definitive concepts with the consequence that theory can be applied decisively to the empirical world and tested effectively in research inquiry.
So far, the suitability of these precision endeavors to solving the problem of the ambiguous concept remains in the realm of claim and promise. They encounter three pronounced difficulties in striving to produce genuine concepts related to our empirical world.
First, insofar as the definitive empirical content that is isolated is regarded as constituting by itself the concept (as in the statement that, "X is the intelligence quotient") it is lacking in theoretic possibilities and cannot be regarded as yielding a genuine concept. It does not have the abstract character of a class with specifiable attributes. What is "intelligence quotient" as a class and what are its properties? While one can say that "intelligence quotient" is a class made up of a series of specific intelligence quotients, can one or does one point out common features of this series—features which, of course, would characterize the class? Until the specific instances of empirical content isolated by a given procedure are brought together in a class with common distinguishing features of content, no concept with theoretic character is formed. One cannot make proposals about the class or abstraction or relate it to other abstractions.
Second, insofar as the definitive empirical content that is isolated is regarded as qualifying something beyond itself (as in the statement that, "Intelligence is the intelligence quotient" wherein intelligence would now be conceived as including a variety of common sense references such as ability to solve business problems, plan campaigns, invent, exercise diplomatic ingenuity, etc.), the concept is constituted by this something which is beyond the definitive empirical content. But since this "something beyond" is not dealt with by the procedure yielding the definitive empirical content, the concept remains in the ambiguous position that originally set the problem. In other words, the concept continues to be constituted by general sense or understanding and not by specification.
Third, a pertinent question has to be faced as to the relation of the definitive empirical content that is isolated, to the empirical world that is the concern of the discipline. One has to have the possibilities of establishing the place and role of the specific content, in the empirical world in order for the empirical content to enter into theory about the world. A specific procedure may yield a stable finding, sometimes necessarily so by the internal mechanics of the pro-
( 7) -cedure. Unless this finding is shown to have a relevant place in the empirical world under study, it has no value for theory. The showing of such relevancy is a critical difficulty confronting efforts to establish definitive concepts by isolating stable empirical contents through precise procedures. Incidentally, the establishment of such relevancy is not accomplished by making correlations. While classes of objects or items covered by concepts may be correlated, the mere establishment of correlations between items does not form concepts or, in other words, does not give an item as an instance of a class, a place or a function. Further, the relevance of an isolated empirical content to the empirical world is not established merely by using the concept to label given occurrences in that empirical world. This is a semantic pit into which scores of workers fall, particularly those working with operational definitions of concepts or with experimental construction of concepts. For example, a careful study of "morale" made in a restricted experiment may yield a stable finding; however, the mere fact that we customarily label many instances in our empirical world with the term, "morale," gives no assurance, whatsoever, that such an experimental construct of "morale" fits them. Such a relation has to be established and not presumed.
Perhaps these three difficulties I have mentioned may be successfully solved so that genuine definitive concepts of theoretic use can be formed out of the type of efforts I have been considering. There still remains what I am forced to recognize as the most important question of all, namely whether definitive concepts are suited to the study of our empirical social world. To pose such a question at this point seems to move in a reverse direction, to contradict all that I have said above about the logical need for definitive concepts to overcome the basic source of deficiency in social theory. Even though the question be heretical I do not see how it can be avoided. I wish to explain why the question is very much in order.
I think that thoughtful study shows conclusively that the concepts of our discipline are fundamentally sensitizing instruments. Hence, I call them "sensitizing concepts" and put them in contrast with definitive concepts such as I have been referring to in the foregoing discussion. A definitive concept refers precisely to what is common to a class of objects, by the aid of a clear definition in terms of attributes or fixed bench marks. This definition, or the bench marks, serve as a means of clearly identifying the individual instance of the class and the make-up of that instance that is covered by the concept. A sensitizing concept lacks such specification of attributes or bench marks and consequently it does not enable the user to move directly to the instance and its relevant content. Instead, it gives the user a general sense of reference and guidance in approaching empirical instances. Whereas definitive concepts provide prescriptions of what to see, sensitizing concepts merely suggest directions along which to look. The hundreds of our concepts—like culture, institutions, social structure, mores, and personality—are not definitive concepts but are sensitizing in nature. They lack precise reference and have no bench marks which allow a clean-cut identification of a specific instance and of its content. Instead, they rest on a general sense of what is relevant. There can scarcely be any dispute over this characterization.
Now, we should not assume too readily that our concepts are sensitizing and not definitive merely because of immaturity and lack of scientific sophistication. We should consider whether there are other reasons for this condition and ask particularly whether it is due to the nature of the empirical world which we are seeking to study and analyze.
I take it that the empirical world of our discipline is the natural social world of every-day experience. In this natural world every object of our consideration—whether a person, group, institution, practice or what not—has a distinctive, particular or unique character and lies in a context of a similar distinctive character. I think that it is this distinctive character of the empirical instance and of its setting which explains why our concepts are sensitizing and not definitive. In handling an empirical instance of a concept for purposes of study or analysis we do not, and apparently cannot meaningfully, confine our consideration of it strictly to what is covered by the abstract reference
( 8) of the concept. We do not cleave aside what gives each instance its peculiar character and restrict ourselves to what it has in common with the other instances in the class covered by the concept. To the contrary, we seem forced to reach what is common by accepting and using what is distinctive to the given empirical instance. In other words, what is common (i.e. what the concept refers to) is expressed in a distinctive manner in each empirical instance and can be got at only by accepting and working through the distinctive expression. All of us recognize this when we commonly ask, for instance, what form does social structure take in a Chinese peasant community or in an American labor union, or how does assimilation take place in a Jewish rabbi from Poland or a peasant from Mexico. I believe that you will find that this is true in applying any of our concepts to our natural empirical world, whether it be social structure, assimilation, custom, institution, anomie, value, role, stratification or any of the other hundreds of our concepts. We recognize that what we are referring to by any given concept shapes up in a different way in each empirical instance. We have to accept, develop and use the distinctive expression in order to detect and study the common.
This apparent need of having to make one's study of what the concept refers to, by working with and through the distinctive or unique nature of the empirical instance, instead of casting this unique nature aside calls, seemingly by necessity, for a sensitizing concept. Since the immediate data of observation in the form of the distinctive expression in the separate instances of study are different, in approaching the empirical instances one cannot rely on bench marks or fixed, objective traits of expression. Instead, the concept must guide one in developing a picture of the distinctive expression, as in studying the assimilation of the Jewish rabbi. One moves out from the concept to the concrete distinctiveness of the instance instead of embracing the instance in the abstract framework of the concept. This is a matter of filling out a new situation or of picking one's way in an unknown terrain. The concept sensitizes one to this task, providing clues and suggestions. If our empirical world presents itself in the form of distinctive and unique happenings or situations and if we seek through the direct study of this world to establish classes of objects and relations between classes, we are, I think, forced to work with sensitizing concepts.
The point that I am considering may be put in another way, by stating that seemingly we have to infer that any given instance in our natural empirical world and its content are covered by one of our concepts. We have to make the inference from the concrete expression of the instance. Because of the varying nature of the concrete expression from instance to instance we have to rely, apparently, on general guides and not on fixed objective traits or modes of expression. To invert the matter, since what we infer does not express itself in the same fixed way, we are not able to rely on fixed objective expressions to make the inference.
Given current fashions of thought, a conclusion that concepts of social theory are intrinsically sensitizing and not definitive will be summarily dismissed as sheer nonsense by most people in our field. Others who are led to pause and give consideration to such a conclusion may be appropriately disquieted by what it implies. Does it mean that our field is to remain forever in its present state of vagueness and to forego the possibilities of improving its concepts, its propositions, its theory and its knowledge? This is not implied. Sensitizing concepts can be tested, improved and refined. Their validity can be assayed through careful study of empirical instances which they are presumed to cover. Relevant features of such instances, which one finds not to be covered adequately by what the concept asserts and implies, become the means of revising the concept. To be true, this is more difficult with sensitizing concepts than with definitive concepts precisely because one must work with variable instead of fixed forms of expression. Such greater difficulty does not preclude progressive refinement of sensitizing concepts through careful and imaginative study of the stubborn world to which such concepts are addressed. The concepts of assimilation and social disorganization, for instance, have gained more fitting abstraction and keener discrimination through insightful and realistic studies, such as those of W. I. Thomas and Robert E. Park. Actually, all
( 9) that I am saying here is that careful and probing study of occurrences in our natural social world provide the means of bringing sensitizing concepts more and more in line with what such study reveals. In short, there is nothing esoteric or basically unusual in correcting and refining sensitizing concepts in the light of stubborn empirical findings.
It should be pointed out, also, that sensitizing concepts, even though they are grounded on sense instead of on explicit objective traits, can be formulated and communicated. This is done little by formal definition and certainly not by setting bench marks. It is accomplished instead by exposition which yields a meaningful picture, abetted by apt illustrations which enable one to grasp the reference in terms of one's own experience. This is how we come to see meaning and sense in our concepts. Such exposition, it should be added, may be good or poor—and by the same token it may be improved.
Deficiency in sensitizing concepts, then, is not inevitable nor irremediable. Indeed, the admitted deficiency in our concepts, which certainly are used these days as sensitizing concepts, is to be ascribed to inadequacy of study of the empirical instances to which they refer, and to inadequacy of their exposition. Inadequate study and poor exposition usually go together. The great vice, and the enormously widespread vice, in the use of sensitizing concepts is to take them for granted—to rest content with whatever element of plausibility they possess. Under such circumstances, the concept takes the form of a vague stereotype and it becomes only a device for ordering or arranging empirical instances. As such it is not tested and assayed against the empirical instances and thus forfeits the only means of its improvement as an analytical tool. But this merely indicates inadequate, slovenly or lazy work and need not be. If varied empirical instances are chosen for study, and if that study is careful, probing and imaginative, with an ever alert eye on whether, or how far, the concept fits, full means are provided for the progressive refinement of sensitizing concepts.
Enough has been said to set the problem of what is wrong with social theory. I have ignored a host of minor deficiencies or touched them only lightly. I have sought to pin-point the basic source of deficiency. This consists in the difficulty of bringing social theory into a close and self-correcting relation with its empirical world so that its proposals about that world can be tested, refined and enriched by the data of that world. This difficulty, in turn, centers in the concepts of theory, since the concept is the pivot of reference, or the gateway, to that world. Ambiguity in concepts blocks or frustrates contact with the empirical world and keeps theory apart in a corresponding unrealistic realm. Such a condition of ambiguity seems in general to be true of concepts of social theory.
How to correct this condition is the most important problem of our discipline insofar as we seek to develop it into an empirical science. A great part, if not most, of what we do these days does not touch the problem. Reflective cogitation on existing theory, the formulation of new theory, the execution of research without conceptual guidance or of research in which concepts are accepted uncritically, the amassing of quantities of disparate findings, and the devising and use of new technical instruments—all these detour around the problem.
It seems clear that there are two fundamental lines of attack on the problem. The first seeks to develop precise and fixed procedures that will yield a stable and definitive empirical content. It relies on neat and standardized techniques, on experimental arrangements, on mathematical categories. Its immediate world of data is not the natural social world of our experience but specialized abstractions out of it or substitutes for it. The aim is to return to the natural social world with definitive concepts based on precisely specified procedures. While such procedures may be useful and valuable in many ways, their ability to establish genuine concepts related to the natural world is confronted by three serious difficulties which so far have not been met successfully.
The other line of attack accepts our concepts as being intrinsically sensitizing and not definitive. It is spared the logical difficulties confronting the first line of attack but at the expense of forfeiting the achievement of definitive concepts with specific, objective bench marks. It seeks to improve
( 10) concepts by naturalistic research,[3] that is by direct study of our natural social world wherein empirical instances are accepted in their concrete and distinctive form. It depends on faithful reportorial depiction of the instances and on analytical probing into their character. As such its procedure is markedly different from that employed in the effort to develop definitive concepts. Its success depends on patient, careful and imaginative life study, not on quick shortcuts or technical instruments. While its progress may be slow and tedious, it has the virtue of remaining in close and continuing relations with the natural social world.
The opposition which I have sketched between these two modes of attack sets, I believe, the problem of how the basic deficiency of social theory is to be addressed. It also poses, I suspect, the primary line of issue in our discipline with regard to becoming an empirical science of our natural social world.
