In these essays I identify some central problematics of the discipline of sociology as I have come to view them over a lifetime of reflecting, reading, and writing in and on that intellectual field.

By "problematics" I mean those generic, recurrent, never-resolved and never-completely-resolvable issues that shape how we pursue our work, how we generate theoretical tensions and conflicts in that work, how we converse and debate with one another, anti how we engage in that complex counterpoint of simultaneous advance, retreat, and repetition in our scholarship. The word "generic" also requires specification. I will not analyze the contents, internal tensions, and shortcomings of the work of any single sociologist or sociological point of view. Rather, I will focus on the philosophical, theoretical, methodological, and (occasionally) ideological issues that pervade sociological work, conversation, and controversy.

In carrying out this assignment, I will move through four successive sociological levels—the micro, involving the analysis of the person and personal interaction; the meso (or middle, or intermediate), connoting structural but subsocietal phenomena such as formal groups, organizations, social movements, and some aspects of institutions; the macro (or societal); and the global (or multisocietal). I assure the reader immediately, lest you suspect, that my choice of these four levels was not determined by the fact that I was asked to give four lectures. I chose them because they reflect commonly made distinctions in sociology, because each level presents some distinctive problematics, and because I myself have done sociological work that touches each level.

That being said, it should be recorded that I do not regard these four levels as embodying necessarily valid distinctions, or as reflecting some readily identifiable social reality. In fact, distinguishing among the four levels is analytically convenient at best and analytically mischievous at worst. On the convenient side, the distinctions yield a reasonable way of organizing a discussion of problematics; even here, however, there is a difficulty, because some problematics appear at more than one analytical level. On the mischievous side, the fourfold distinction lends itself to reification, to the view that the levels are separable and separate kinds of social reality. By this time we should know better than that. For example, the long-standing distinction between gemeinschaft and gesellschaft (a micro-macro distinction) has proved as troublesome as it has worthy in sociology, both because no totally satisfactory definition of either idea has materialized and because social forces emanating from both supposed levels constantly pervade the seamless social process. The troubles do not disappear, moreover, by reconceptualizing, as Jürgen Habermas has done, gemeinschaft as life-world and gesellschaft as the rationalized world of economy-bureaucracy-state. Moreover, I will note from time to time that a number of sociological problematics arise in attempting to define the relations and transitions among the different levels.

A Generic Problematic: Sociology's Intellectual Identity

I begin with a problematic that has, does, and will infuse all of sociology: its intellectual identity. Without hesitating, we normally refer to sociology as a social science. That is a misleadingly simple designation. Sociology, created out of and in the context of already-established humanistic traditions (especially history and philosophy), scientific traditions (both physical and life sciences), and aesthetic traditions, has never been able to make up its mind whether it is primarily scientific, humanistic, or artistic in orientation. Appreciating this, we can also understand the basis for many outside criticisms and internal divisions of the field. Let me elaborate.

•     By the scientific orientation I refer to inquiry that focuses on natural laws and logically closed theoretical formulations; on causal, even deterministic analysis; on a dispassionate, objective, and nonevaluative attitude toward the subject matter under study; on empirical investigation; on precision and measurement; and on a method of inquiry that isolates and controls as many causes as possible to arrive at the decisive ones.

•     By the humanistic orientation I have in mind inquiry that focuses on the human being; entails a preoccupation with the human condition (including human welfare, justice, equity, and suffering); does not hesitate to evaluate; and deals above all with human meanings, systems of which constitute culture.

•     By the artistic orientation I refer to two separate connotations—first, an aesthetic posture toward subject matter, or an emphasis on pattern; and second, an emphasis on the application of knowledge, as in the "art of medicine" or the "art of the possible."

All three orientations constitute both the significant moral/intellectual environments of sociology and parts of the sociological enterprise itself. With this in mind, we can appreciate why sociology typically enjoys—better, suffers from—two types of experiences.

First, from outside, critics representing these orientations in their "purer" forms may react selectively to—that is, recognize some but not all parts or—sociology and assail the field for aspiring to what they represent, but failing to achieve it. Natural scientists frequently take on a bemused or hostile posture because sociology—or social science in general—pretends to be but is not truly scientific, that is, is "soft," which is shorthand for qualitative, imprecise, humanistic, anti artistic. Humanists or those in the humanities may either find sociology territorially offensive, an intrusion on their traditional turf, or see it as arid and inhumane. Those who arc artistically oriented find sociology ugly or useless, according to which of the two orientations of the artistic is invoked.

Second, from inside, sociology's complex composition—denying from its neighboring and penetrating orientations from science, the humanities, and the arts—leads sociologists to raise doubts about their field's mission, unity, anti identity and to foster recurrent controversies. Among these arc the following familiar, overlapping examples.

•     Sociology as value-free (scientific orientation) versus sociology as value-relevant (humanistic orientation).

•     Sociology as fount of basic knowledge (scientific orientation) versus sociology as applied knowledge (artistic orientation).

•     Sociology as agent of knowledge creation (scientific orientation) versus sociology as agent of ameliorative or revolutionary improvement of society (humanistic and artistic orientations).

•     Experimental-aggregative-causal modes of analysis (scientific orientation) versus configurational-clinical modes of analysis (artistic orientation).

•     Emphasis on positive facts and behavior (scientific orientation) versus emphasis on phenomenology or individual meaning (humanistic orientation).

•     Quantitative analysis (scientific orientation) versus qualitative analysis (humanistic anti artistic orientations).

In the American sociological tradition the scientific sides of these polarities have dominated. In fact, I can submit a working definition of that imprecise term, "mainstream sociology," as composed of those who, in one way or another, have tended to vote for the scientific side of the polarities. As a rule, this mainstream side enjoys a disproportionate share of support and resources from university administrators and external funding agencies, given the general dominance of the scientific ethos in American society. At the same time, there is no academic department and no sociological convention or congress in which the larger scientific-humanistic and scientific-artistic tensions do not surface in overt or covert ways. And, of course, different national and regional traditions of sociology manifest different combinations and balances among the several polarities.

Microsociological Analysis: The Problem of Other Minds

The microsociological level includes sociology's version of social psychology, or the study of the person as oriented to the external, especially the social, world; processes of personal interaction; and the study of small groups that typically but not always involve face-to-face interaction. It is important not to reify this definition. The micro level shades into the higher levels: for example, the family is simultaneously a primary group and an institution, and persons anti interpersonal interaction make a difference at all levels of social organization. We ignore such admonitions for the moment, however, and move forward on the assumption that the micro level constitutes a legitimate analytic focus.

The micro level involves, above all, human beings (social science investigators) directly studying other human beings (as they interact with one another). This means that, under all but the most radical of behaviorist approaches, we, the investigators, use our minds to study other creatures with minds. (Even radical behaviorists do not escape the assumption that investigators have minds, if they are to investigate!) A corollary is that there must be at least minimal communication between the investigator and others. This feature is evident in the experimental study of humans, interviewing, participant observation, and even in direct observation; it disappears only under conditions of completely unobtrusive observation of behavior and the study of recorded precipitates of behavior. Even in those cases we are often said to "converse" with our subject matter.

This observation implies that we cannot proceed with study without immediately exciting an enduring philosophical issue: the problem of "other minds." The problem is a logical offshoot of skeptical philosophy, rooted in the works of Berkeley and Hume, who raised fundamental questions about our ability to assume the independent, enduring existence of all external reality, including the minds of others. The problem was reactivated in the 1940s in a forceful statement by the English philosopher John Austin ([1946] 1979).When the problem of other minds is extended to the sociological investigation of persons—both individually and in interaction with one another—it divides into several subquestions:

•     How can we know that others, including other minds, exist? This is the issue of skeptical philosophy proper.

•     Even if we know or assume that other minds exist, how can an individual know about the nature and contents (thoughts, images, sensations, emotions) of minds other than his or her own? This is the problem of verstehen that pervades the Weberian and related traditions of sociology.

•     On what basis (observation, imputation, empathy, projection) do we inter or attribute mental states to others? How can we have confidence in these inferences and attributions?

•     What is the influence of our own minds (as investigators) on the minds of others, and vice versa, in the process of investigation?

•     On what basis do interacting others know and take account of the minds of one another?

Many variations of and controversies in microsociology—and to some degree in sociology as a whole—emanate from the different ways in which these questions are answered.

For present purposes I will ignore the first question—the impossibility of the existence of other minds. For one thing, l do not have the patience to address, one more time, the question, Flow do we really know? when there is always room for enough doubt in contemplating the human condition to permit us to ask that question a hundred or a thousand more times without coming to a point of final philosophical certainty. Perhaps more important, it is fair to say that by becoming social scientists, we adopt an affirmative working answer to that question as a matter of occupational commitment; if we did not, we would be forever packing our philosophical luggage and never stepping on the train to take the sociological journey.

With respect to the role of the investigator's mind and the minds of others in the generation of sociological knowledge, the positions of Émile Durkheim and Max Weber—as expressed in their sociological manifestos—constitute a vivid point of reference. Durkheim's sociological positivism represents an extremely simple solution, in the sense that he attempted to define both as methodologically nonproblematical. He argued that if the sociologist approaches reality with preconceptions in mind, he or she distorts that reality. Instead, the investigator should cast aside such preconceptions and regard social phenomena as things, that is, as "distinct from the consciously formed impressions of them in the mind." The most important characteristic of a "thing," moreover, is "the impossibility of its modification by a simple effort of the will" (Durkheim [1895] 1958: 28). The investigator should free his or her mind of all preconceptions, take a passive relationship to social reality, and deal with phenomena "in terms of their inherent properties" (ibid.) and their "common external characteristics" (ibid., 35).

On the side of those being investigated, Durkheim took an equally positivist stance. He regarded individuals' "internal states"—such as motives, meanings, and emotions—as inaccessible to observation, and demanded that we put in their place some "external" or behavioral index that can be studied scientifically (Durkheim [1893] 1949: 64). He was hostile to the practice of appealing to psychological forces to explain social facts, going so far as to assert that "every time a social phenomenon is directly explained by a psychological phenomenon, we may be sure that the explanation is false" (ibid., 104). There is reason to doubt that Durkheim's own sociology was in fact presupposition-less, anti there is clear evidence that he himself had ready recourse to psychological explanations, for example, in his analyses of suicide rates. However, his methodological position is clear: neither the mind of the investigator nor the mind of the investigated should—and, in the best of worlds, does—play an active role in the generation of sociological knowledge.

Weber contrasted with Durkheim on both counts. He insisted that scientific reality was not given in nature but was the product of a series of selections based on the investigator's interests and values. In investigating we select only those parts of reality that are "interesting and significant to us, because only [those parts are] . . . related to the cultural values with which we approach reality" (Weber 1969: 78). In thus asserting that any attempt to develop a sociology "without presuppositions" is "not only practically impossible—it is simply nonsense" (ibid.), Weber was saying, in our terms, that the investigator's mind must be regarded as active in the generation of scientific knowledge. Similarly, in understanding sociological reality, the investigator must also take into account the minds of the investigated persons. To appreciate this, one need only consult his definition of human action: "the acting individual attaches a subjective meaning to his behavior— be it overt, covert, omission or acquiescence" (Weber 1968, 1:4; emphasis added).

In making these methodological commitments, Weber generated two derived and fundamental problems.

•     How can the investigator grasp the mind of the actor being investigated? To respond, Weber developed his idea of empathic understanding, or verstehen. Such a problem did not arise in Durkheim's methodological outlook, since neither the mind of the investigator nor the mind of the actor being investigated was thought to be problematical.

•     How can the (presumably idiosyncratic) subjective meaning systems of different individuals be compared with one another, so that general statements, if not laws, can be generated in sociological investigation? To respond, Weber developed his notion of the ideal type, which entailed, in effect, assuming sufficient likeness or comparability of others' minds so that general constructions, such as "the Protestant ethic" and "rational bourgeois capitalism," could be characterized as sociological reality. Again, such a problem of comparing other minds with one another did not arise in Durkheim's positivism, because he regarded aggregation into general social facts as a matter of simply recognizing like items of behavior, or "things."

In a word, Durkheim solved the other minds issue by ignoring both these problems but in the meantime created a philosophically vulnerable methodology. Weber acknowledged the independent significance of the minds of the investigator and the investigated but in the meantime forced upon himself the need to develop formulations that would address the philosophical issues that he created by that acknowledgment.

Durkheim's and Weber's formulations represent two possible solutions of the other minds issue, namely, ignoring it and confronting it. While versions of the Durkheimian solution still remain in microsociology under the heading of social behaviorism (e.g., Homans 1974), most approaches and debates in microsociology confront the problem: as a result, differences in perspectives emerge in terms of how anti with what theoretical assumptions to deal with the problem of other minds. The following types of "solutions" are evident in the literature.

•     The utilitarian solution, found in classical economics, endows the actors being investigated with a material, self-interested motivational orientation and, in addition, asserts that this endowment coincides with reality, that is, that individuals are universally materialistic and egoistic.

•     The "heuristic" utilitarian solution acknowledges that the rational pursuit of economic interest is not a psychological universal but assumes that that orientation is a powerful theoretical device in explaining market and other behavior.

•     The "radical pragmatic" utilitarian solution, associated with the position of Milton Friedman (1953), allows that the postulate of rationality may be erroneous or nonsensical, but so long as it "works" in predicting economic results, it is justified.

•     The phenomenological solution, considered generally, involves the claims that the meaning systems of investigated others do indeed constitute sociological reality, and that it is essential to discern these meanings to understand and explain that reality. There are a number of variants of the phenomenological approach. Weber's formulation of verstehen is one. I now call attention to several other related formulations.

(1)     The ethnographic approach in anthropology and sociology is committed, in one way or another, to take into account the reports and accounts of informants and other actors in describing the culture and behavior of the society or group under investigation.

(2)     The symbolic interactionist approach rejects the idea that the individual person is a passive vessel through which various social and psychological forces work, and insists that human behavior cannot be understood without taking into account how individual persons actively endow their internal and external environments with meaning and act on the basis of that meaning. The methodological implication of this position is that the sociological investigator must grasp, appreciate, and incorporate those aspects of meaning in any explanation of human behavior.

(3)     The ethnomethodological approach also rejects the idea of the causal significance of social structure and social roles, and insists that the behavior be understood as the product of continuous reciprocal monitoring of meanings and accounts of action on the part of interacting individuals.

(4)     The dramaturgical approach exemplified by Erring Goffman (1959) entails a view of the individual actor as continuously manipulating meanings in social situations as a way of presenting himself or herself. While phenomenological in the sense that an understanding of this process is essential, Goffman also endows individuals with certain motives, such as status-striving, maintaining esteem, and maintaining consistency of self-image. In that respect his approach can be likened to that of rational choice theorists, who assume that certain preference schedules exist in the actors they investigate.

(5)     Pierre Bourdieu's (1984) formulation of habitus' as the individual's meaning-orientation system is an interesting phenomenological variant. On the one hand, Bourdieu criticizes economists' distortions of actors' orientations because they force upon actors their (i.e., the economists') preferred worldview. In this criticism Bourdieu suggests the desirability of a more appreciative approach. His own formulation of habitus is a multifaceted orientation that includes motivation, past experience, memory, and information.

(6)     The methodological position voiced by some in the feminist and ethnic studies literature (sometimes called sexual and racial essentialism) holds that social investigation involves the sympathetic appreciation of others' situations and outlooks, and only those of their kind (women and minorities, respectively) are capable of this appreciation.

(7)     An opposing methodological position argues that foreigners to a group have a special advantage in understanding its situations and outlooks, because they stand outside the taken-for-granted assumptions of those being investigated. ("The last creature in the world to discover water is a fish.")

•     For purposes of completeness, one might mention "radical phenomenology," a position that maintains that reality inheres in others' meanings, but these are so inaccessible that they defy understanding by investigators. This approach is a completely skeptical solution to the problem of other minds, and, it must be acknowledged frankly, leads to a kind of methodological paralysis that appears to rule out sociological investigation entirely.

This map of approaches to the problem of other minds is sufficiently comprehensive to lay out the central methodological dilemmas involved. At one extreme, radical positivism solves the problem of other minds by denying its importance; the evident cost of this strategy is to distort reality by ruling out essential sources of determination in human behavior. At the other extreme, radical phenomenology solves the problem of other minds by submitting to it; the evident cost of this strategy is to render scientific generalization impossible. With respect to the intermediate strategies, the key issue is whether and to what degree the investigator actively endows meanings (including cognition, affect, and motive) to others anti whether and to what degree the investigator attempts to appreciate or grasp others' meanings as they experience them.

In their turn, endowment and appreciation generate their own methodological dilemmas. Endowers provide themselves with the opportunity to generalize about others, but risk distorting others' meanings. Appreciators claim that they represent human reality more faithfully, but risk being caught in an ideographic trap, unable to break out of the idiosyncrasies of individuals' meaning systems and to generalize about them.

This dimension of endowment-appreciation is loosely—but not precisely—correlated with other aspects of social scientists' worldviews, though it must be realized that the following observations are impressionistic and speculative. Endowers tend to have a "hard" (scientific, objective, frequently quantitative) approach to their subject matter; appreciators tend to have a "soft" (humanistic, subjective, frequently qualitative) approach to theirs. Endowers tend to "analyze"; appreciators tend to "identify with" and "respect." Endowers may be accused of "arrogance"; appreciators may be accused of "sentimentality." And, most speculatively, endowers tend toward a conservative mentality on a variety of dimensions, appreciators toward a liberal or radical mentality. Exceptions—such as the scientific Marxist approach, which is both endowing and radical—may test these assertions, but they do constitute food for thought.

At this point I would not be surprised if readers are not experiencing a certain impatience. It is all very well, you might be asking, for me to lay out this conceptual geography and note the difficulties of each of the solutions to the problem of other minds. That is the luxury of the critic: to identify problems in others' thought without oneself taking a stand on the resolution of those problems. Imagining your discontent, I will now lay out a series of assertions that I believe to be the correct philosophical and methodological guidelines for sociologists to follow in the study of individuals and their interaction with others.

First, we cannot adopt the radical skeptical position (either that other minds do not exist or that we cannot know them) or the radical phenomenological position (that other minds can be known only by those investigated). If we adopt either, we may as well turn in our identity cards as sociologists, because both positions involve, in effect, a renunciation of the possibility of knowledge about others. Moreover, in adopting either, the only role that remains for us is that of the negatively minded philosopher.

Second, we cannot fully externalize or deprivatize other minds by embracing a behaviorism that denies, ignores, or freezes the independent significance of humans' perceptions, affects, intentions, and evaluations. Under behaviorism I include both stimulus-response theories and rational choice formulations that rest on assumptions of fixed and stable preferences. The latter are, in effect, stimulus-response theories, because they explain behavior on the basis of knowledge of individuals' external circumstances (price, income, etc.). We need not embrace fully either symbolic interactionism or agency theory in acknowledging the necessity of taking into account the independent role of "internal" human processing of perceptions, sensations, affects, and intentions, as well as the adaptive alteration of behavior based on this processing.

Third, we should acknowledge that, as social investigators, we are agents, anti that we must endow others with generalized motives, orientations, and capacities. This is a theoretical and methodological necessity in my estimation, for two reasons. (1) It seems a philosophical impossibility to reflect others' representations of their own minds without some independent act of interpretation; and that act, however minimal, entails the further act of endowing. (2) Unless we acknowledge the necessity for assigning general orientations to others, we are caught in a trap of methodological particularism and cannot hope to strive for general statements about our chosen subject matter. In other words, as social investigators we are forced, by theoretical and methodological necessity, to take the analytic step from the appreciation of the idiosyncratic to the typification of the general.

Fourth, in typifying others' orientations, we should not press the typifications beyond what they are—namely, constructed and admittedly distorted simplifications, necessary to proceed with investigation and analysis. Put another way, we should not reify or essentialize. Furthermore, the typifications must always be regarded as tentative anti open to empirical investigation and conceptual manipulation. Suppose, for example, in studying social mobility, we assume, as an analytic starting point, that actors are guided primarily by orientations of status-striving—preferring a higher rather than a lower place in a status hierarchy. Some such typification is essential for analytic purposes. However, that typification should not be simply executed then forgotten. Independent empirical investigations (observation, interview, survey) can throw light on when such a typification is likely to be valid and useful and when it should be altered or abandoned. In a word, we should regard the act of analytic typification as a sociological problematic, to be investigated in the same way as any other sociological problem.

Finally, there are two reasons generic to the sociological enterprise why we should take a direct scientific interest in the typifications with which we endow others' minds.

(1) There is no formulation in sociology—micro, meso, macro, even global—that does not contain at least implicit assumptions and attributions about actors' assessments, knowledge, emotions, and motives. Durkheim, in Suicide, attempted heroically to live up to his methodological dictum that social facts are caused and explained by reference to other social facts. In illustrating this he linked types of social integration to different rates of suicide. Yet in case after case, we find Durkheim making theoretical sense of these links by referring to the putative psychological effects of, say, anomie, and to the putative behavioral effects of those psychological effects. Similarly, analyses of international finance strategies rest on assumptions about individual or corporate actors' motives or goals (to maximize profits, to secure conditions of monetary or political stability); and analyses of international politics consistently endow heads of state and foreign ministers with explicit game theoretical goals and strategy preferences or with some mix of motives of national self-interest, aggression, and peace seeking.

(2) One of the main vulnerabilities of every social science is that many explanatory efforts may degenerate into arbitrary or post factum formulations and accounts because investigators have at their disposal a range of possible psychological orientations (typifications) that can be attributed to actors. If "findings" do not seem to fit an explanation based on one assumed orientation, then the investigator might replace it with another, which presumably makes better sense of them. In other words, the range of assumptions about other minds constitutes a suitcase of possibilities, and the investigator is forever tempted to pick different items from the suitcase, as the occasion demands, to make sense of anomalous or contradictory results of empirical research. All this is to underscore that the most careful and self-conscious attention be given to the kinds of psychological endowments that sociologists (and other social scientists) attribute to the minds of those they study.

How to Endow: Limitations on the Model of the Stable, Adult, Informed, Literal Actor

Having established the necessity for the social investigator to endow the actor with psychological characteristics—all the while keeping that endowment open to revision in light of theoretical and empirical considerations—we turn to the next logical question: What should be the content or substance that we attribute to those we investigate? In other terms, what kinds of assumptions about human nature should we adopt to generate the most effective explanatory models of behavior and interaction?

We begin our response to this question with a familiar and identifiable image—the rational economic actor in the classical utilitarian tradition. There are two reasons for choosing this model: its simplicity and its radicalness. By the latter I mean that utilitarian theorists imposed very extreme conditions on the image in the process of making it simple.

The ingredients of the utilitarian model in classical economics are the following:

•     By way of motivational assumptions, tastes are "given" for purposes of analysis; actors strive to maximize their pleasure—in this case their economic well-being—and act in accord with a few assumed psychological principles, such as that of diminishing marginal utility.

•     The individual possesses complete information about the market.

•     The individual operates in an environment with only a few identifiable elements, namely, the price and quantity of goods available and the level of his or her own resources.

•     The individual reacts to information literally, that is, does not make mistakes about it, does not elaborate it into complicated symbolic systems or otherwise distort it.

•     Equipped with tastes, preference schedules, resources, and information about the market, the individual calculates correctly and behaves consistently.

•     Others behave predictably and interact peacefully with the actor; actors do not coerce or defraud one another, anti all occupy equally powerless positions with respect to their capacity to influence conditions of the market.

Such simplifying assumptions also enter into models that sociologists employ. The "role conformity" model of the actor found in some versions of role theory, for example, regards the individual as a socialized person, one who understands the norms and sanctions as they are presented to him or her, does not distort information, and is motivated, other things being equal, to follow the dictates of the normative system in which he or she is implicated.

We now understand enough about the process of endowing the actor with typified orientations to set aside the objection that those orientations do not constitute an accurate or adequate psychology. No typifications ever do. One can ask, however, about the conditions under which an assigned typification is useful as part of an explanatory model or theory. With reference to the assumption of the rational economic actor, my answer is a simple but unfamiliar and controversial one. It goes as follows: such a model is most useful under those social conditions that institutionalize its characteristics and conditions. A typical market for commodities is such an institution: it makes price levels and wage levels public, not secret; it institutionalizes choice and rewards calculation, in that it provides actors who calculate effectively with valued and disvalued sanctions (money, goods, commercial failure); actors in the market are protected, more or less, by institutionalized laws against fraud and coercion. All this is to say that the validity of the typifications assigned by the economist to the actor is assured by the institutional conditions of the actor-in-situation, and, for that reason, predictions of behavior based on typifications under those conditions are likely to be powerful, because they reflect the realities of institutionalization.

In the history of their discipline, economists and others have realized that highly typified assumptions are not always valid, even in institutionalized market conditions. Correspondingly, much of the history of economics has been marked by relaxing the highly simplified typifications and then reconstructing models based on new typifications.

To choose a few examples of this: the theory of imperfect competition relaxes the assumption that individual actors cannot influence production and prices; economics as a whole has moved away from its earlier materialism and has introduced a whole new variety of utilities (prestige, power, self-esteem, etc.) that constitute preference schedules; many models of market behavior based on lack of information, uncertainty, and risk have been generated; and recent explanations are based on the assumption that when the costs of information and transaction become too high, economic actors invent systems of hierarchy (authority relations in organizations) and trust (in contracts) to minimize those costs.

What has given economics its theoretical continuity is its insistence on reincorporating the typification of rationality (including purposive-ness, reasonableness, calculation, and self-interest), even after important parametric conditions have been relaxed. Enthusiasts of such typification, such as Gary Becker (1976), have argued for its universality, that is, its applicability to all kinds of institutional conditions (systems of justice and crime control, racial discrimination in labor markets, mate selection and family formation, fertility and other demographic behavior). That principle of rationality, even watered-down rationality, is the primary article of faith of economic analysis, and that principle survives even when the model of the rational actor is incorporated into the analysis of political behavior (Downs 1957) and the analysis of sociological problems such as conformity to authority and participation in collective behavior (Coleman 1990).

It is to economists' and others' credit that the limitations of the classical typification of the economic actor have been recognized and that relaxations and reformulations have proceeded apace. Those modifications have given greater flexibility and applicability to economics, though perhaps at the cost of theoretical determinacy. However, the continuing insistence on incorporating rationality as a typification has actually discouraged certain other lines of relaxation of the central postulates of economics. These lines concern mainly the non-rational and irrational sides of life, which, it can be argued, pervade all behavior including economic behavior in the purest of markets. The following examples of omitted relaxations come to mind.

•     Active distortion of information on the part of actors. Revised economic models, as indicated, take account of lack of information, risk, and uncertainty, but not rationalization, projection, displacement, and other forms of distortion that deviate from the assumptions of actors' assessment of economic and social reality that are built into the economic models.

•     The process of symbolization of commodities, work, and other economic phenomena, which endow them with systems of meaning above and beyond their reference to assumed utility preferences.

•     The place of affect in interaction. In one sense this is an odd omission, because the original summum bonum of the utilitarian tradition was the seeking of pleasure and the avoidance of pain, which are, of course, matters of affect. In fact, however, the affects of anxiety, rage, love (especially blind love), neurotic conflict, psychosis, and addiction (except when it can be explained as rational action; see Becker and Murphy 1988) do not find a place in formal economic analysis.

To notice these systemic distortions in the tradition of economics and elsewhere directs us toward ways of modifying the micro-level attributions we impose on our subject matter. To that topic I now turn.

The Incorporation of Cognition, Meaning, and Affect into Sociological Typifications

I have just summarized and assessed the effects of the analytic bias toward rationality in the utilitarian tradition in the economic sciences. Traditions other than economics have also contributed to diminishing the affective, nonrational side of life. Marx inherited much of the utilitarian tradition and tended to subordinate all moral and affective sides of life to the status of by-products of the objective forces of history, though indirect references to affects—the misery of proletarianization and the proletarian rage—are implied in his work. Durkheim, also a thoroughgoing positivist, rejected "internal states," though his analysis of ritual and collective effervescence in religious celebrations takes account of the vivid emotionality of such occasions. Weber's work concentrated above all on rationality and rationalization (though not in the economists' sense of the term). He admitted the "affectual" as one of his four fundamental types of action (Weber 1968), but aside from its appearance in the analysis of charisma, the affective aspects of the Protestant religion, and his remarks on disenchantment, the emotional side of life occupied a peripheral place. In general, then, as Alan Sica (1988: 32) has concluded, Western theorists have not greeted the notion of the irrational warmly "as a concept or as the root of an ideology . . . for some time."

A major exception to the rationalistic bias in social thought is found in the late nineteenth and early twentieth centuries in the work of Nietzsche, Freud, Le Bon, Pareto, Mosca, and Michels, all of whom stressed the nonrational and irrational sides of human life in different ways. But as far as long-term impact on the sociological tradition, Nietzsche has had little place, Freud must be considered marginal (particularly in comparison with his influence on the "culture-and-personality" approach in anthropology), Le Bon has been passé for decades, Mosca and Michels are remembered mainly for their contribution to the distribution of power in society, and Pareto is famous not for his residues and derivations (the emotional and ideological dimensions of society) but for his "optimum," a rational principle of economic welfare and social policy. Another major exception is found m the work of Georg Simmel, the only classical sociologist who even approached a sociology of emotion. While he insisted on a level of sociological reality (sociological forms) that is independent of psychological impulses, his own work gave open acknowledgment to the salience of the erotic and the emotional in many interpersonal relations (Simmel 1984).

The second half of the twentieth century has, if anything, accelerated the trend toward rationalist psychology and rational control in society. Consider the following developments as illustrations.

•     In economics, the continued vitality of the tradition of rational choice. This perspective has also accomplished a major invasion of political science and has made minor incursions into sociology and anthropology as well.

•     In psychology, the overwhelming success of the "cognitive revolution," with offshoots into cognitive science anti information science.

•     In psychoanalysis, the shift from drive psychology toward ego psychology and object relations theory, anti the general decline of psychoanalysis and its insistence on the irrational role of the unconscious.

•     The development of theories of rational management of the economy through monetary and fiscal policy, as well as the rationalization of approaches to business in "management science."

•     The vast rationalization accomplished by the computer revolution in all its facets.

•     Shifts in more specific fields of study, such as social movements, with the diminution of emphasis on affect and ideology toward the more rational emphasis on resource mobilization and strategies of social movement organizations (see chapter 2).

This family of tendencies in the social sciences—and more illustrations could be produced—has continued apace in the late twentieth century, despite the evident vitality of the nonrational in the postmodern world, which appears in new versions of alienation and disenchantment, mental disorders, conflict, violence, and a resurgence of primordialism in group attachments and political life.

I conclude this chapter by suggesting a corrective to the individualistic, rational approach—an alternative methodology for the study of social psychology and personal interaction.

Supraindividual Constructions in Microsociology: The Example of Trust

At the beginning I defined the microsociological level as focusing on the individual person and personal interaction. Even within this circumscribed range, however, it is essential to distinguish three levels of analysis: (a) the psychological; (b) the intersubjective; and (c) the sociological, or systemic. In this closing section I explicate these distinctions, argue for their independent significance, and point out how all three are essential for explaining orientations and behavior at the microscopic level. Throughout I will use the idea of trust as a running example.

Trust has appeared in several lines of literature in the past two decades. In sociology, Niklas Luhmann (1979) and Bernard Barber (1983) produced major, if preliminary, theoretical statements; economists have interpreted trust as a generalized way of reducing transaction costs (especially the costs of securing information and establishing the conditions of exchange) in market settings (Williamson 1993); in economic sociology trust has entered into the analysis of market networks (Granovetter 1985), the ethnic economy (Light and Karageorgis 1994), and the informal economy (Portes 1994); and empirical analyses of trust have been ventured in areas such as the family, monetary attitudes, and litigation (see Lewis and Weigart 1985).

Trust is an evident and familiar psychological phenomenon, as revealed by the notion of a "trusting person." The attribute of trust connotes cognitive dispositions (expecting consistent behavior on the part of others), attitudes toward others ("people are basically good"), emotional dispositions (low levels of anxiety and hostility in interpersonal relations), and an openness of behavior that emanates from these dispositions. The typification "rational economic actor" implies a trusting person—one who accepts offered prices as honest prices and one who does not expect theft, violence, or fraud on the part of others. The idea of a "distrusting person" connotes outlooks, emotions, and behavior opposite to that of the trusting person. Goffman's (1959) typified actor appears to be something of a distrusting, even paranoid, person, always on the lookout for feint, sham, phoniness, conning, and "presented" rather than authentic impressions.

A number of lines of microsociological analysis focus on the problem of intersubjectivity, including strategies by which interacting individuals sustain predictable interpersonal relations ("trust," though it is not always named that) and repair those relations when they threaten to break down. The main image of interaction in the symbolic interactionist literature is of individuals engaged in giving off signals, interpreting and reinterpreting meanings associated with those signals, and mutually informing and correcting one another (Blumer 1969). The same model of monitoring taken-for-granted understandings and meanings is the focus of ethnomethodological analysis, with special emphasis on "repair work" that is done when conversation and other kinds of interaction break down (Garfinkel 1967; Schlegloff 1987). Goffman's dramaturgical games accomplish the same purposes, as do the processes of "frame alignment" (Goffman 1974; Snow et al. 1986), or the bringing of different persons' interpretive frameworks into agreement as a condition for interaction and the pursuit of collective goals. All these lines of analysis represent investigators' efforts to take account of the intersubjective processes that deal with the problem of other minds in interaction. Attitudes of mutual trust constitute "successful" outcomes of this kind of interaction, though other outcomes, including distrust, can be envisioned when the processes of trust generation break down.

At both the psychological and intersubjective process levels, the unit of analysis remains the individual, even when interaction is involved. Yet the analysis of trust and other aspects of interaction cannot end at this point. Processes of interaction—including two-person interaction—also have a sociological element, a systemic quality that cannot be generated by referring to persons and their psychological characteristics and cannot be reduced to or derived from these. In a word, trust becomes institutionalized; as such it has a sustained and reproduced reality of its own, independent of the psychological states of trust or distrust experienced by interacting persons. For this reason it is erroneous to treat trust only in terms of psychological expectations, "repeated games," or a condition sustained only so long as it serves the purposes of persons in interaction—for example, to reduce transaction costs—to be given up when it no longer serves those persons.

How should we characterize the sociological level of trust? The most evident instance is found in fiduciary roles, in which it is normatively expected, sometimes legally mandated, that people act in a relationship of trust to one another, even though they may not trust one another from a psychological or intersubjective point of view. But that is only the most evident example. Virtually all human interaction— even between blank strangers and between enemies—involves some level of institutionalized trust or distrust. Put differently, interactive relationships involve expectations about the following elements of trust.

•     What is the range or scope of activities in which those interacting may expect predictable behavior on the part of others? In relations between strangers approaching one another on the street, the list is minimal and would include only expectations that the other keep a certain distance and not behave menacingly or as if out of control. Two drivers approaching one another on the road share expectations that are more complex, for example, that the other will obey the rules of the road as embodied in the highway code and will not make unpredictable or indecipherable moves with the vehicle. In neither example is there any expectation that the other will experience any specific affects: it is perhaps desirable to remain calm, but if the other driver is boiling with anger and that anger does not spill over into breaking the specific expectations, the affect is not relevant. The scope of activities to be trusted in more enduring relationships (among friends, lovers, or kin) is greater, and often calls for helping behavior, "understanding," psychological support, going out of one's way, and experiencing relevant affects.

•     What affects are appropriate in the relationship? Some relationships (e.g., between cashier and customer) are neutral on this score; others (e.g., between physician and patient) call for the active suppression of emotion on the side of the one and are more permissive on the side of the other: still others (e.g., between spouses) call for the active expression of mutual respect, sympathy, and love.

•     What is the mix of trust and distrust in a relationship? The institution of the market provides interesting mixes. Certainly, as Simmel ([1900] 1978) demonstrated, any market transaction is marked by a trust in the validity and value of the money exchanged (rules of "legal tender"). If this trust breaks down, substitute systems of trust (e.g., barter) may arise to take their place. At the same time, the idea of caveat emptor and the proliferation of practices such as giving receipts as proof of purchase, guaranteeing refunds, providing title deeds for property, requiring "truth in advertising," and affording legal recourse for cheating also indicate that distrust is institutionalized as well. Multiple marriages institutionalize "infidelity" without making it a matter of trust and distrust; monogamous marriages make sexual fidelity a matter of trust and often make infidelity a principal basis for marital dissolution; open marriages, at least in principle, institutionalize the denial of trust and distrust as marital issues.

•     What are the rules of evidence that justify the inference that trust has been broken and legitimize expressions of suspicion and distrust, as well as the emotions of anxiety, shame, rage, and revenge? These affects, like all other aspects of institutionalized relations, are issues of normative regulation, and demonstrate the truth that no feature of social life, however private, is beyond social interest.

These aspects of institutionalized trust set the stage for interaction and constitute important determinants of how individuals define situations and react to them. Raising the question of institutionalization, moreover, calls out for the development of a complex classification of types of trust, based on the types of sociological relationships into which people enter—buyer-seller, politician-voter, parent-child, teacher-student, employer-employee, friend-friend. It also calls for the analysis of types of trust and associated expectations that informal roles—hero, fool, villain, scapegoat—generate. It also calls for modeling of benign and vicious circles of intersubjective trust and distrust, respectively, and of how these result in the cementing, alteration, or breakdown of institutionalized relations of trust.

Many interesting sociological situations arise under conditions of discontinuity among the three levels—psychological, intersubjective, and institutionalized—of trust. We capture such discontinuities in daily discourse by identifying the "gullible," who trusts more than the social situation merits; the "paranoid" (or what the French capture in the phrase mefiez-vous ), who trusts less; and the "realist," who both trusts and distrusts and keeps an eye out for evidence of both. Democracy institutionalizes extended relations of trust and distrust between citizens and politicians. Under certain conditions (exposure of malfeasance, swings in public mood, shifts in standards of morality) the ground rules for what constitutes trustworthiness and untrustworthiness may shift and redefine the political process. Under such circumstances, moreover, intersubjective trust between citizen and politician forever threatens to break down into mutual distrust.

The objective of this exploration of the different levels of psychological and sociological trust is to demonstrate that (a) the sociological mode of analysis is not different at the microsociological level than it is at higher levels of social organization; indeed it penetrates the most intimate levels of interaction; (b) regularities of behavior cannot be understood or explained without reference to the sociological dimension; and (c) the psychological, intersubjective, and institutional levels must incorporate affective and other "nonrational" ingredients. We might even say that the model of sociologically naive actors—as in rational choice and game theoretical models—are misguided for almost all occasions. Our typifications and explanations must involve the continuous interaction of institutionalized expectations, perceptions, interpretations, affects, distortions, and behavior.