Perhaps one difference in the self-presentation of theory qua theory now is that much of it seems to have moved in the direction of highly formal analysis, incorporating sophisticated mathematical models.  So perhaps the "core" of theory now largely consists of formal modeling on the one hand and analyses of the writings and ideas of the original generations of grand theorists, on the other. I suppose one way to look at the phenomenon that Michèle has identified is to ask the key question: how do we sociologists recognize theory? What counts as theory? And, what are the professional and epistemological consequences of something counting as theory? It may be as simple as some research putting the theoretical frameworks and claims in the foreground and the empirical cases and examples in the background and some research emphasizing the significance of the cases or phenomena themselves. The former will be more likely to be recognized as theoretical works than the latter.  Sociologists who have migrated toward the culture and comparative-historical sections may not be satisfied with this de facto method of conceptualizing theory. They may prefer a more "alternating current" approach to the case(s) and the abstract framework, bringing now one, now the other into the foreground.

One final thought about Michèle's method of surveying the discipline via a focus on the top ten graduate programs. Many graduate students come to sociology after cutting their teeth on some form of it in the liberal arts undergraduate college context. I may be extrapolating wildly from my experience at Swarthmore College, but I would guess that these students may have a more theoretical orientation toward sociology at the outset of their graduate educations. How might this inflect their own incorporation of the more satellite (or hybridized) model of theorizing as they enter the profession? 

In short, after Alvin Gouldner's "coming crisis" (1970), developments in sociology and the human sciences more broadly pulled theory in multiple directions, away from its high modernist dispensation, perhaps precisely because it is so central to the sociological imagination. Yet perhaps that high modernist moment was itself an aberration and a conceit. As I have argued in Cultures of Inquiry, theory finds its way into even the supposedly most antitheoretical historicism (and by implication, ethnography, etc.), and its relation to those frames of inquiry is manifold. Yet this is hardly a new circumstance; rather, it is a general condition of sociohistorical inquiry as a discursive domain. Thus, in the wake of high modernism, the task of people interested in theory as an enterprise is to (re-)constitute the domain of theory as a Hydra to which many different forms of social thought are connected.

In programmatic terms, it doesn't surprise me that much of the teaching of social theory has been taken up by sociologists who identify themselves first and foremost in different subdisciplinary terms. The collapse of high modernism has not (yet) been worked through in terms of how sociologists orient themselves to theoretical discourse in relation to research practice. But what does seem clear is that, as Robin says, many sociologists do theory in the course of research, rather than thinking of theory as an independent or relatively autonomous enterprise. What's interesting here is, that as Michèle shows so well, some subdisciplinary specializations have much more couched their enterprises in ways that engage theory, while others -e.g., sociology of the family, criminology, etc., etc.- have not. What I suspect is going on here is that those other subdisciplines are more oriented as theory consumers (i.e., they will draw on Foucault, or Bourdieu, or Smith, or Bauman, etc.), whereas culture, comparative/historical, and gender are zones where scholars think that core epistemological and theoretical issues are at stake in how their enterprises (and in turn sociology) are constituted. Yet here we are faced with a great puzzle about the moment of high modernism: the scholars often invoked as great theorists have themselves typically been involved in substantive analyses centrally concerned with culture, and yes, history. Which begs two sorts of questions. First, is the image of a high modernist era of grand theory itself a social memory of a Chimera? And second, what ASA section(s) would Marx (or _____ [fill in the blank]) have joined?

It is appropriate that Michèle starts the conversation by taking a kind of national sample of department theory teachers, Robin writes about the hybridization or shifting constructions of the term theory, and John writes about the historical transformation of modernist conceptions of theory. It is clear why these three are good section chairs, speaking articulately in the languages of their subfields. But the very differences in accounts seems to contradict the starting point of the conversation set up my Michèle. There remain differences among the subfields that are meaningful, if not profound barriers for entry, for people wanting to think conceptually about their work. The statements by Robin and John represent different forms of analysis, theoretically animated but with distinct qualities. Their essays demonstrate that there is not one way to be "interested" in theory, or one project of theorizing in sociology. This said, there cannot be a service sector of thinkers serving as satellite ateliers for "theory." What constitutes theory is different among those who teach it and seek it out in their research whether they come to it from the sociology of knowledge, cultural sociology , historical sociology, gender studies, or any other subfield. Even if they teach Marx, Weber, and Durkheim, how they do it is different. The pattern of differences may be a result of the balkanization of subfields in sociology that John describes, but what it creates is neither a disorganized mish-mash or a hierarchy of linked fields. It is a pattern of distribution of the common problem of thinking theoretically, where subfields concentrate on parts of the task and theory junkies read across fields to become sophisticated.

I find this heartening both because I am one of those theory junkies who loves to read across fields, and because the balkanization of sociology has been bemoaned for much of my professional career. I'm happy to be able to point to the virtues of distributed cognition that sociologists can use as well as understand in their work. It is a socially-organized capacity to think better than a Mannheimian individual or Haraway's situated knower, linking differences in expertise and points of view through conversation and yielding a social capacity to think. The debate that Michèle has started points precisely to this possibility, and can allow the Theory section not to integrate theory from its satellites, but to make the distribution of theorizing both more visible and visibly useful for sociologists. 

UNC has a reputation --deserved, at least in part-- as a "powerhouse" in two areas: sophisticated quantitative methodology and demography. Graduate students often come to me anxious about what they see as a deficiency in "qualitative sociology" and a desire to do a qualitative project. I have developed a stump speech, akin to those currently plaguing the country's political landscape, about the poverty of the qualitative-quantitative distinction. I suggest that they replace it with a new way of evaluating sociological work: the extent to which it is deeply theoretical. In the main, students take to this argument, although I suspect they take it with the grain of salt appropriate to the fact that they are hearing it from the guy who taught their introductory (and, in general, only) social theory seminar. The next question is simple: what makes a given work of sociology theoretical?  That's precisely the question raised by Michèle Lamont's thoughtful investigation into the teaching and practice of theory. She makes two interesting, related points: (1) that theory is often (and perhaps increasingly) taught by non-"theorists"; and (2) that theory section members often maintain memberships in other sections, including most prominently Culture and Comparative/Historical Sociology.  Interestingly enough, as Robin Wagner-Pacifici notes in her response, there are overlaps among section leadership too; Michèle is a major figure in cultural sociology, and as I write this I am peering at the new volume on my shelf by John Hall et al., "Sociology on Culture."  One might be forgiven for supposing, given those crossovers, that it is culture around which the others are in orbit! (It's worth noting that, in the scheme of an ASA membership, section memberships are very low in cost, and may appease friends and colleagues seeking to boost section coffers, which may artificially inflate the degree of overlap.)  Let me suggest, returning to my claim that all (good) sociology is theoretical, that the reason for this overlap is that these are the areas of sociology in which explicit theorizing is important. Not wanting to put too fine a point on it, I'm suggesting that culturalists, social psychologists, comparative historicists, etc., are "into" theory because they study pressing, complex questions that demand a sophisticated theoretical infrastructure. Because of these thorny issues, theory has not been routinized into method as it has in some other subfields such as, say, demography.  That leads me to my next point: theory's relegation to self-identifying "theorists" is bad both for theory and for sociology. Again taking UNC as an example, the graduate theory seminar is viewed as a combination of a ritualistic paean to the presumed progenitors of sociology (one cannot be allowed a Ph.D. in sociology without "knowing" Marx, Weber, Durkheim, Simmel, and whomever else the instructor thinks worthy) and an introduction to the epistemological problems to be solved by the upcoming methods sequence. 

What, then, makes theory theory? Is it, as Robin asks, essentially about the (re)production of "grand theory"? Alternatively, is it constituted by formal, testable models ("theory construction") that provide hypotheses for empirical testing? Or is it the task of extrapolation and comparison commonly referred to as "theorizing" or "interpreting," in which an analyst suggests meanings for empirical findings suggesting further work? My contention is that theory should be understood as all three of these, but that its most exciting possibilities lie in the third: the creative task of interpreting events, ideas, and practices. As Michèle's piece suggests at the end, crossovers between theory and other sections is "as it should be." These crossovers, though, are not so much indications of satellite status as they are contributions of different analytic resources to vexing sociological questions. 

I have a quite "analytical", philosophy of social science-based approach to teaching theory, and am beginning to use simulations and more models-based reasoning in theory teaching. This for me is where the exciting theoretical action is nowadays in American universities. Via networks and the new science of complexity, all of the other social sciences are rediscovering social structure, and "sociological" ways of thinking -but unfortunately (and ironically) not too many sociologists are actually participating in this.

What I dislike most about the predominant ways of teaching contemporary sociological or social theory in the US, is the way nearly all textbooks and syllabi, whether by traditionalists or postmodernists, focus entirely on a "canon" of individual theorists and texts -usually with a lot of biographical clutter, and reverential textual exegesis. The fact they now introduce Judith Butler, Jürgen Habermas or Jacques Derrida in amongst the usual suspects (Parsons and co), hasn't changed the fact that this kind of presentation of theory is undigestable to students, boring, and irrelevant, both to empirical sociologists, and more importantly, in terms of what is going on across most of the social sciences. For my money, it is indicative of the increasingly esoteric self-isolation of American sociology from the leading currents of social science, as much as an ill-advised slippage towards "theory" in the humanities. And yet with rational choice, game theory, evolutionary theory, networks, institutionalism, social capital, human ecology, and some of the more analytically retrievable parts of the interpretative/phenomenological tradition (such as self/other interactionism, ethnomethodology, theories of power, reproduction, field/habitus, cognitive theory, etc) -there are actually lots of cross-disciplinary theoretical tools out there. With these, it is possible to build a comprehensive overview of contemporary social science theory, that can work at both basic and advanced levels. This is what I think theory should be all about: teaching students to think in abstract, hypothetical, yet sim-ple analytical terms about how society "might" work. 

Graduate Training in Theory
Those who teach theory in top departments are likely to be those who received training in theory while in graduate school. In an era when virtually all American sociologists are encouraged to focus on substantive empirical research, I suspect that most of those who teach theory received their degrees from programs in which they were required to study theory intensively in addition to their particular substantive areas of specialization. In short, when few just do theory, theory teachers come from schools where all graduate students get a strong grounding in theory.  To examine this supply-side explanation of the instruction of theory in top departments in this country, I located the PhD institutions of the individuals Lamont named. (I added my current home, Indiana University, to the list of departments that Lamont listed.) The table below shows what I found. 

Berkeley, Harvard, Chicago, and Princeton PhDs are over-represented among theory instructors relative to all other departments. Given the numbers of PhDs produced and the reputation of the departments, one might expect to see more theory teaching being done by scholars from Wisconsin, Michigan, Stanford, UCLA, and North Carolina. To investigate the theory training at the most over-represented and under-represented of these schools, I asked someone familiar with the graduate curriculum at each of these schools questions about their graduate training, and verified it by checking departmental websites. 

Berkeley, Chicago, and Harvard seem to be unique in requiring a theory component to the field or qualifying exam. My observations of the differences between graduate training at Berkeley and IU and the email I received back from Berkeley alums describing graduate theory training at their new institutions suggests that the formal requirements indicate dramatic differences between departments in the role of theory in sociology. While I know less about Chicago and Harvard's graduate training, the theory component of the qualifying exam at Berkeley was only a piece of an environment that suggested, implicitly and explicitly, that theoretical competence is fundamental to being a sociologist. Developing a research agenda that engaged in broad theoretical questions was the central task of graduate school. Acquiring the methods to pursue these questions was treated as secondary. This attitude is quite different than the theory as "ritualistic paean" that Perrin has found at UNC. Several of the Berkeley alums I corresponded with -without prompting- mentioned that the departments they were currently at specialized in methodology instead of theory. These departments develop cultures in which it is assumed that universally high levels of statistical competence are expected as a matter of course. While all of these departments produce PhDs who are theoretically and methodologically sophisticated, few departments seem to emphasize theoretical and methodological training equally. Thus, it is not surprising that a few schools are over-represented in the teaching of theory. If more top programs were to increase the baseline level of theoretical training required of all graduate students, this might diversify and increase the pipeline of future instructors of theory. 

The Feminization of Sociology
Lamont suggests that a less than hospitable climate for feminist scholarship in the Theory section may partially explain the shifting locus of theoretical scholarship. Pushing this point further, I'd like to suggest that this pattern might be related to the gender of scholars as well as the content of the scholarship. In December 2001 Footnotes published a short article on the gender composition of ASA sections. This article reported that the Theory section is the most disproportionately male of all large sections of the ASA, while the Culture section is balanced with respect to gender. The Theory section is not among the sections with high number of student members, while the Culture section is. This suggests that women entering the profession are not joining the Theory section in high numbers. The article concludes by suggesting that "sections that do not attract new women members may lose numerical standing." I suspect that many women entering the profession who do theoretical work identify with the Culture section instead of the Theory section. My sense is that the Culture section is perceived and experienced as being more hospitable to women in no small part because its leaders have included high visibility women scholars such as Wendy Griswold, Ann Swidler, and Michèle Lamont. Culture section chairs have been women slightly more than half of the years since the founding of the section in 1988. Women interested in theoretical work have found an intellectually engaging home in the Culture section instead of trying to break into established networks of scholars within the theory section. Thus, the feminization of sociology may have contributed to this shift in the location of theoretical development in the field of sociology. 

This all helps me make sense of a puzzle I encounter whenever I teach a graduate culture seminar. I always wonder what to title it. "Sociology of culture" is not telling enough. I've taken to calling it "cultural theory and cultural analysis," all the while uncomfortable that other seminars do not often get this double duty. In the seminar, we read "theorists" who in very different ways constructed "culture" to begin with --Adorno, Gramsci, Foucault, and Bourdieu are usual suspects. We also read people who have asked empirical questions innovatively, refining what sociologists mean by "culture" in relation to particular social realms. The latter works represent "cultural analysis," not "theory" per se, but are no less theoretical; they help us see *culture* in the empirical relationships, institutions, selves, which other parts of sociology already taught us to see. Identifying the theorist authors distinct from others, I hope, helps place the significance of their work inside the subfield. Fifty years from now there may be different reasons, or no special reason, to begin a culture seminar (as distinct from a theory seminar) with theorists. For now, the subfield has only relatively recently defined the basic terms for our conversations. 

To distinguish theory from theoretical is not to reify current theory genres as exclusively theory, for all time. On the contrary, some theoretical, empirical studies will become horizon-setting works, and thus theory as I mean the term. I agree with Michèle that grand theory for theory's sake seems like a less promising enterprise right now than middle-range theoretical works that are informed by theory. At the same time I look forward to discovering from among those works in cultural sociology the ones that are horizon-setting, not necessarily grand ones that generalize about wide swaths of human experience. By identifying many works as theoretical, we keep ourselves open to the potentially changing construction of "theory" as a set of genres--the hybridization which Robin mentioned. I take that as a good thing both for theory and sociology. 

People with an interest in Theory have had a tough time in Sociology since the second World War. In the prewar period, figures such as Sorokin and MacIver played prominent roles in major university Sociology departments, and leading sociologists, such as F. Stuart Chapin at Minnesota, taught the Theory courses. Standards of scholarship, and particularly of knowledge of the broad range of historical social theory, were high. All this changed significantly for the worse in the post-war period, when the then young turks, Merton and Parsons, dismissed the older style of theory writing, which was explicitly historical, with a new model of "systematic theory," which was explicitly anti-historical. In a very short time, the theory jobs in key departments, of which there were always very few, were taken over by their students, who shared many of their prejudices. To get a job in the post WWII period people were commonly advised to label their interests "Social Change." Yet in the forties in such departments as Wisconsin theory was still lively, and theory had prestige in Sociology, and students who trained at these places, such as Arthur Vidich, kept the tradition alive, though they did not do so in the major departments. So theory of the older kind went down-market.  Demographic factors were very important. Parsons and Merton were young when they came to influence. They were fueled by grant money -Parsons had a massive Carnegie grant, for example, with few strings attached. They had an expanding field in which to place their students. And they had good students, as well. Some of these students did pay their dues as "theorists" and master the classic texts, despite the narrowness of the Parsons-Merton approach to the classics (and its systematic misrepresentations of them as prototypes of their own views). But the representatives of this approach were in power for an extraordinarily long period, from the early fifties to the early eighties. Whatever else may be said about this post-Parsons-Merton generation -and several of its members were good sociologists and serious, decent academics- they were not original thinkers, or serious historical scholars. Nor were they intellectually tolerant. It was difficult to publish in the US without kow-towing to them, and impossible for their younger critics and rivals to get jobs in the departments whose theory teaching they controlled. The sixties produced bitter generational conflict. One indirect consequence of the conflict was that the flow of new European developments in theory, including new developments in history of theory, was choked off, with the following results: US and world sociology and social theory radically diverged, students in theory could not get access to world-class theory in major American Sociology departments.  Sociology itself, reeling from the enrollment (and student quality) drops of the early seventies, became less intellectual, less theoretical, and more gendered. The great debates in "theory" of the late 20th century were between Foucault and Habermas. US academics generally were eager importers of these ideas, but, especially in Sociology, contributed little to the discussion -we had no Giddens or Baumann, as the UK did. Habermas took Rawls and Rorty seriously, but not the sociologists of the same age. So the intellectual action in the US in "theory" shifted away from Sociology at the same time as the field was going through a key generational (and of course also gender) change.  When this generation was replaced as they retired, theory positions vanished or were turned into places that non-theorists, including people in cultural sociology, could be appointed, at a time when gender was a critical issue in many academic appointments. Yet in a sense, there were important intellectual continuities between Parsons and cultural approaches anyway -Geertz (who was the probably the best of this generation) was a Parsonsian who had revamped the basic Parsonsian ideas about central values and repackaged them very successfully as a kind of postmodernism. This became a surrogate for theory which had the advantage that "cultural studies" provided opportunities to produce standardized research products -books- that could be made the basis of a normal career. But these people were not theorists in the older sense. They had not paid their dues in the study of the classic texts, they had no interest in the philosophical side of the tradition, and they tended to be very disciplinary in their orientation. 

The reproduction of theorists at major departments thus came to an end, and the community of discourse that still existed as late as the sixties came to an end, with the astonishing result, as we have seen from the contributions to this discussion, that in major departments, the people who teach theory do not even represent themselves as theorists. 

And (this paragraph kind of invites Andreas and Marjorie DeVault into the discussion, since they're asking similar questions to Michèle's original ones) what would "real theory" be like if it were always open to infusions by ethnographic (and other process-oriented) visions? Would it dissolve into nothing but ethnomethodology-nothing BUT process, that is? I think of myself as doing the "extended case method," but have the sneaking suspicion that the only way that that method works is if you have a fairly rigid, schematic theory to "reconstruct." That is, for example, you can use ethnography to reconstruct "Marx" or "Weber," but not "Dewey," who suffers from too little structural analysis, too much of a sense of constant motion. This seems to be pointing me in the direction of saying that we need overly static theories, but that can't be right either -it's like poaching.  Back to Michèle's original questions: In a really empirically oriented department like mine (University of Wisconsin), everyone says that their work is "informed by theory" and "challenges pre-existing theory." No one claims to do work that is a-theoretical. And that, in a lot of people's minds, is why it is possible NOT to have a specific set-aside sequence for "real theory"-because theory is "everywhere." It's sort of parallel to arguments about setting aside a department for African American Studies-teaching "theory" as a separate object implies that you can teach the rest of sociology without theory; but, the "theory" in the courses about fertility rates is pretty slim. Not requiring, e.g., two semesters of theory, or a theory prelim, risks placing theory not "everywhere," but "nowhere."

Partly for this, I think we need grand theory for shared reference points-collective representations-in order to think together as sociologists. A good grand theory raises (gee, this is brilliant hahah) important questions about how we got here and how to live and what to struggle for. Maybe the reason for the tremendous overlap between the theory and culture sections is that a lot of us culturistas became sociologists when we realized that you can't change "structures" without having a "people" who want to change them. Entering sociology with this political/theoretical mission/question, we inherited the reflexive mission of sociology that most everyone else could more easily forget. 

I start this commentary with an empirical hypothesis that is both comparative-historical and cultural. This is proper because the trend that Michèle Lamont's essay describes is that comparative-historical and cultural sociologists are getting the jobs that once would have been reserved for theorists. Consequently, if I wish to address those who are employed in "theory" jobs today, I must begin with a comparative-historical and/or cultural angle. I should say in regard to the hypothesis, that I certainly do not mean to suggest that there is anything surprising or unpredictable about comparative-historical sociologists or cultural sociologists welcoming the opening up of career positions that were once largely reserved for those specializing in theory. What makes the hypothesis interesting is that the organization created to serve those who specialize in theory would seem to welcome the loss of jobs for those who specialize in theory.  One of the contributors to the loss of jobs for those who specialize in theory is the trend toward seeing theory's importance mainly in its ability to generate researchable hypothesis. Unfortunately, although I have tried to disguise the issue of concern here as a researchable hypothesis, it really is not. Frankly, I don't care whether this is the first time in history, or one of the few times in history, or if there is a long and consistent chronicle of the leaders of a occupational organizations welcoming the loss of career opportunities for those who are the organization's purported constituency. Under the guise of proposing a hypothesis, I am advocating a position.  I feel that I must be perfectly clear about my position because the point seems to have become lost in the first wave of comments. The question is not whether some comparative-historical or cultural sociologists should teach theory. The question is whether those who specialize in theory should be able to get a job or whether it is better to discourage our students from specializing in theory and instead to encourage them to disguise what Mukerji calls their "theory addiction" in the more acceptable form of comparative-historical and cultural sociology. I do not in any way suggest that there is anything wrong with what Lamont calls theory satellites. All theorists are thrilled when our work is read and used and modified and critiqued by those who do not specialize in theory. We are even more thrilled when those who did not originally specialize in theory become "theory junkies."  That really is not the point in contention. The question is whether the satellites should now completely colonize the planet that they once orbited around. Perrin, in his comment argues that, "theory's relegation to self-identifying 'theorists' is bad both for theory and for sociology." One might ask whether it is equally bad if quantitative methods is taught by self-identified "quantitative people?"  Indeed, is it bad for sociology if any sociology class is taught by a self-identifying "sociologist?" Nevertheless, I am not arguing that theory courses should only be taught by those who are pure in theory. It is certainly to be expected that those who are able to productively mix theory and empirical research will be both better teachers and better researchers. But why is the cross-over that produces this mix only valued in one direction? If it is good that some start out in a substantive area and become interested in theory, why is it not equally good for some to start out as theorists and to let their theoretical interests lead them to substantive research areas? While there are many career paths open to the former, there are few for the latter.  I believe that theory as a specialization is an important part of sociology. Immersion in a theoretical tradition provides the possibility for the alternate worldviews that are the mark of grand theories. These worldviews are what connect sociology to a wider intellectual public. The popularity of such writers as Foucault, Bourdieu, Habermas, Bauman and Giddens is precisely that they are social theorists. In addition, it is the tension between these worldviews that provides the common core of sociology. As Mukerji points out in her response, if the balkanization of subfields in sociology does not result in a disorganized mish-mash or a hierarchy of linked fields, it is because theory junkies read across subfields. But it is not just that a few insomniacs with not enough papers to grade and little taste for popular fiction like to read across the subfields. The commonality is not created simply by the act of reading, but the purpose of that reading. We read across the subfields to see what alternate worldviews are being produced. The common core of our discipline is created through an engagement with those worldviews that substantive theories project.  One of the most remarkable aspects of the disappearance of theorists from American sociology is how little it has been noticed. Lamont's 'exploratory' research seems to be about 40 years too late. Apparently, the top schools have not been hiring theorists for that long and one can imagine that other schools only hire theorists because the top schools have snatched up all of the good comparative-historical and cultural sociologists. I believe that the decline of theory in sociology can be traced to a historical event which established a departmental structure that now continues the decline.  This is, of course, a researchable historical question and I will therefore be only suggestive here. The historical event was the attack upon Talcott Parsons. I take Stephen Turner's point in his comment that Parsonians dominated theory positions, but I think the decline in theory can be traced to the post-Parsons' era. If grand theory, such as Parsons', inherently contains a political standpoint and a worldview, then attacks upon any dominant theory are quite predictable in a society where there is a plurality of political standpoints and worldviews. The particular form that the attack upon Parsons took was an unlikely alliance between positivists and politically engaged leftists against Parsons' grand theory. Between these two schools of critics, there was very little agreement on the grounds of the criticism. For the positivists he was not empirical enough; for the leftists he was too conservative. But both implicitly agreed that it was the creation of grand theory that was objectionable. For the positivist, it was unscientific; for the leftists, it was likely to interfere with the need for immediate political action. These two approaches combined in their attack on grand theory and fractured the discipline of sociology after Parsons.  Although they have both begun to fade, the decline of theory in sociology continues because of the criticisms' effect upon the structure of departments. Since the successful attack on grand theory, most sociology departments have seen the role of the specialized theorist as mainly historical exegesis. If the task of a theorist is to create new theories, then a good argument can be made for having more than one theorist in a department, so that interaction can promote creativity. Once the role of the theorist is mainly to pass on a history, each department needs only one. That change alone limits the number of jobs for theorists, but the next predictable step virtually wiped them out. In making a hiring decision, a department is guided by many criteria. Once the particular specialty has been decided upon, these criteria boil down to two types. The first is how well a particular candidate fits the internal criteria of the specialty. For theorists, that means their ability to create new theories, their command of the tradition, their ability to connect theories to empirical work, their publications, their status in the community of theorists, etc. The second criterion is the candidate's fit in the alliances that make up the department. As the first set of criteria lost their importance for theorists, the second became more prominent. Alliances in the department rarely included other theorists, since the new theorist was being hired only because the department's one theorist was leaving. Consequently, other factions in the department preferred hiring an ally who could also teach theory. For example, those doing cultural work prefer to hire another cultural sociologist (who could also teach theory) rather than a strong theorist. This is what I surmise has led to the situation today, but whatever the historical and structural causes, the result is clear: the United States is not producing young, world-class theorists. Few graduate students want to specialize in theory since they cannot get a job.  Instead most theoretically minded students go into cultural or comparative historical sociology. This situation is far from irreversible. This is the positive side of theory-satellites. Theorists now have natural allies embedded in many departments. As indicated, these allies first instinct will be to hire someone in their same specialty for a theory position, but I believe that the case could be made for hiring a theorist. What cultural sociologist would not hire a Bourdieu and what comparative-historical sociologist would not hire a Habermas? It is a short step to ask them to hire a potential Bourdieu or Habermas. Even the positivists and politically engaged sociologists can be convinced of the value of someone who specializes in theory. At one time positivists might have thought that their middle-range theories would be eagerly used by policy makers to shape a better society. Few can have such a naïve belief today. To the extent that positivists wish to communicate their findings to a wider intellectual public, they need their findings to be taken up as part of the sort of grand theory that theorists write. Similarly politically engaged sociologists might have once believed that the revolution was imminent and that the creation of theoretical alternatives only interfered with the required action. Today, all those who wish for social change must agree that the most significant obstacle is the lack of alternate worldviews.  In an era when it could be asserted that history has ended, it may be that the sort of alternate worldviews created by grand theory have never been more necessary.

What seems to missing is not the means to reverse the trend, but out failure to provide an argument for it. Perhaps we have not noticed this trend until now. Perhaps some of us identify too closely with our marginal status. Perhaps we think that it is only natural that our best social theorists are European and increasingly not even sociologists. Whatever the reasons, we now need to make the case for hiring those who specialize in theory, and we must ask that those who represent our interests take up this cause and use their connections in the theory-satellites to spread the message. 

Why are sociological theorists such an endangered species? In part, we are collectively to blame. Those of us who regard ourselves as theorists cannot agree on the nature of our enterprise. Much of what passes for sociological theory is intellectual history; some is nakedly normative; other 'theoretical' work has no straightforward empirical implications. This lack of agreement is reflected in the reactions to Michèle Lamont's comments. Whatever pleasure we might gain from participating in sharp theoretical debates, fundamental scholarly dissensus detracts from the prestige of sociological theory. By increasing the probability that any given theoretical article will receive split reviews, it minimizes the chance of publication in the top journals. Dissensus also makes it difficult for theorists to be hired in academic posts and to gain grants and competitive fellowships. Given the lack of consensus about theory among theorists, why should the subject continue to be taught to our students?  For me, theory consists of a set of internally consistent propositions specifying the mechanisms responsible for the production of social outcomes. As such, anything that qualifies as a theory must yield empirical implications. What does this kind of theory offer to sociologists? There are at least two answers. The first answer is that theory provides a small set of enduring general questions. These questions lead to research from any number of perspectives utilizing a wide variety of relevant evidence. Fundamental questions structure the discipline because they permit people in quite different and specialized fields -scholars who on this account otherwise would have nothing in common- to understand, value, and learn from one another's research. The most fundamental questions in sociology are two. How does social order come about? And, once such an order is established, how does it change? I submit that attempts to answer these two questions have the unintended effect of structuring the discipline.  The second answer is that theory offers a ready source of causal mechanisms. Adequate social explanations must entail two separate components, causal relations and causal mechanisms. If a causal relation asserts that X causes Y, then the relevant causal mechanism states just how it is that X causes Y. Whereas much effort in sociology has been devoted to the conditions under which causality can justifiably be inferred (particularly in quantitative data), considerably less attention has been paid to the role of causal mechanisms. Although there is little agreement about what causal mechanisms really are, they make claims about a process occurring in units of analysis at some lower level (such as individuals) that affect outcomes at a higher level (such as groups or societies). To this simple description, it is necessary to add an important proviso, namely, that causal mechanisms are frequently unobservable. This means, of course, that often they cannot be directly inferred from evidence.  Therefore, theory supplies the key questions in the field as well as the mechanisms animating the proposed solutions to these questions. Note that these two tasks lie well beyond the scope of empirical research. Theory has to be judged somewhat differently from empirical research. 

The value of a theoretical work does not directly depend on its empirical adequacy. Durkheim's Suicide doesn't offer the last word on the determinants of suicide rates, but despite this we continue to assign it in our classes. Ditto with The Protestant Ethic and most of the other classic works in the field. Many attacks on these works' empirical adequacy have been published, but the classics seem to have been remarkably resilient. These and other classical theories are far from historical relics admired by a small set of scholars with antiquarian tastes. The continued relevance of these works rests on the questions they pose and the mechanisms they propose.

Our graduate students need to be exposed to theory to learn about these fundamental questions and social mechanisms. This serves at least two different purposes. In the first place, it helps to determine the relative importance of research topics. By and large, the more directly a given piece of research relates to an important theoretical question, the more important it is. In the second place, familiarity with the uses and abuses of existing causal mechanisms can inspire students to develop entirely new ones. If these benefits of theory were more widely appreciated, I have no doubt that the quality of sociological research would improve. And it would be clearer that theory has an undeniable place in the graduate curriculum.

Since World War II, Stephen Turner observes in the 2004 Chair's Corner exchange, "people with an interest in Theory had a tough time in Sociology." Case in point-trying to shape a career in theory in the 1950s, I heard from all sides: no future for me in sociology if I did a theory dissertation.  Turner's account of Parsons and Merton throws no light upon my predicament, however.  Dismissive glosses on those two pioneers of a professional approach to theory work constitute part of our problem. Such responses have long precluded clear understanding of just what Parsons and Merton accomplished. Systematic, yes, but why decry them and their closest junior colleagues as "explicitly anti-historical" (Merton on the history of science, Parsons on fascism and social evolution, Bellah on Tokugawa Japan, Smelser on the industrial revolution, Geertz on Javanese traditions, Eisenstadt on the history of everything?) Selective, yes, who isn't? Yet why call "exclusionary" two of the most broad-gauged and polymathically literate homegrown American sociologists of the past century? At the outset they engaged seriously with (then largely untranslated) texts of Durkheim, Pareto, Simmel, and Weber. Habermas, another polymath who nevertheless read Durkheim, Weber, Mead, and Parsons only late in his career, felt obliged to conclude that Parsons's work was "without equal in... its social-theoretical scope," and in other respects as well.  It was not how those theorists redirected the field that gave me a hard time then; it was the radically empiricist tendencies of the 1950s. Theory work actually picked up notably in the 1960s-the Theory Section began in 1968, after all-and Goodman plausibly dates its decline to a couple decades later. He notes that it was not the narrowness of Parsons's work and influence but the attack on him and "grand theory" by radical empiricists and politically engaged leftists that did the damage. (While I'm at it, Ronald Glassman's dates are also a bit off: Parsons did not learn systems theory in Heidelberg in the mid-1920s but at Harvard with Henderson in the mid-1930s; Merton did not push middle-range theory due to the "red scare" of the late fifties, but promulgated his famous position contra Parsons at the ASA meetings a decade earlier.)  Whatever, those who promote the vocation of autonomous theorizing need to make the case for it, as TP and RM did repeatedly. Michele Lamont's provocative lines invite us to do so. Her provocations include the advice that autonomous theorizing had best be left in abeyance- thereby, so Goodman, putting herself as Section Chair in the anomalous position of celebrating the reduction of career opportunities for the guild's own constituency.  I find precious little in the exchanges, however, to challenge Lamont's formulations as they bear on the rationale for what some call 'autonomous theorizing.'  Her assertions include the claims that 

• some sociologists do "'grand theory' entirely disembodied from empirical research"-without indicating what that might mean or who the culprits might be;  whatever non-empirical theory might be, sociologists should be encouraged to stop doing it [grand theory' disembodied from empirical research seems increasingly to be a thing of the past. ...And from where I stand, this is as it should be."];
• it should nevertheless be retained for the important task of "complementing" empirical research ["grand theory certainly remains a crucial part of a good sociological education, and of intellectual literacy more generally"]. 

So I shall defend two counter-claims: that 

• sociology badly needs more, not less, autonomous theorizing; and 
• the ASA Theory Section has been remiss in not heeding to the guideline formulated by Michael Hechter: that a major raison d'être of the Section is to nourish and sustain autonomous theorizing. 

When the Section engaged a comparable dialogue in 1996-7, I sought as chair to sidestep such difficulties ("Social Theory as a Vocation," Perspectives 19:2). To be sure, I did suggest that autonomous theory work in sociology generically yields certain benefits, like enhancing intellectual mastery and instilling discipline. But it seemed obvious to me that we could make little headway in clarifying the vocation of social theory without essaying a differentiated understanding of the forms and functions of theory work. So I proceeded, not by discussing what theory is, but by talking about what theorists do. I asked: what exactly do we profess when we profess Social Theory? Then I pointed to exemplars of valuable theory work, of both the pure and mixed kinds, and sketched a way of presenting them that might help to orient those inclined to profess social theory as a vocation.  The argument proceeded by making three sets of distinctions-between custodial functions (recovery) and heuristic functions (invention); between heuristic work internal to the discipline and heuristic work external to the discipline (Simmel's three extra-disciplinary domains for sociology); and, within the internal heuristic, among a number of discrete functions. These distinctions generated a paradigm of nineteen distinct kinds of theory work, each of which had been realized in acclaimed exemplars, and each of which remained, and remains, worthy of additional serious inquiry. Space permits me here to reproduce only the paradigm itself. 

An Agenda of Theory Work in Sociology 

1. Custodial work
    1.1. recovering the heritage: critical editions, translations, exegeses
    1.2. correcting and reinterpreting accepted recoveries 
    1.3. forming inventories and codifying ideas
    1.4. constituting and sustaining dialogues 2. Heuristic work: internal 
    2.1. conceptual articulation, disambiguation, and reformulation
    2.2. construction of conceptual frameworks and typologies
    2.3. model-building and formalization
    2.4. problem-finding and problem-justifying
    2.5. redirecting current scholarship
    2.6. theorizing new areas, through
        2.6.1. extrapolation from earlier work
        2.6.2. attending to emergent phenomena
        2.6.3. constructing new analytic angles

3. Heuristic work: external
    3.1. foundational
        3.1.1. ontological
        3.1.2. epistemological
    3.2. representational syntheses
        3.2.1. syntheses with other disciplines and perspectives
        3.2.2. grand descriptive syntheses
    3.3. ethical syntheses
        3.3.1. ethics of scholarship
        3.3.2. theory-practice relations 
    3.4. social diagnosis and criticism.

Rather than rehearse the dozens of exemplars I cited in connection with this schema, let me make the case afresh by encouraging others to engage dialogues on substantive issues in the field  in addition to career issues. We could:

• Identify pieces of good theory work that signally improved the course of sociological investigation or practical applications, and instances of flawed or misdirected work that could have been signally improved by the adduction of some well-crafted theory work.  Reflect on some of the exemplary achievements of theory work with an eye to formulating criteria of excellence and calling attention to those exemplars. 

 
• Use something like the above paradigm as a basis for engaging in collegial deliberations about the character of theory curricula and training programs. (Myself, I refuse to consider simply teaching a set of given canonical works to constitute a meaningful course in theory. At present, my theory curricula consist of units such as Finding Problems, Justifying Problems, Disambiguating Concepts, Constructing Narratives, Constructing Models, Linking Assumptions to Models, Critical Codification, Legitimating a Discipline, Comparing Formulations, and the like.* Such work should, I believe, be required of all sociology students, as resources for enhancing intellectual mastery and distilling discipline and thereby improving the often shoddy quality of conceptual work and theoretic argument in research projects.) 

 
• Attempt to formulate significant cutting-edge problems, both those raised by current complications in the literature and those that ought to be addressed but that, for one reason or another, have been ignored or neglected. Here is a sample list. 

 

Why do we tolerate the widespread assumption that equality is the sole or hegemonic value to guide research?  What would happen if one assumed other kinds of value relevance instead, such as freedom, solidarity, excellence, courage, justice, spirituality, or others?

What do different ways of visualizing Simmel's "geometry" of social forms accomplish respectively? Does Scheler's concept of Ressentiment exhaust the content of Nietzsche's earlier version? Is its content exhausted by the concept of relative deprivation?  What might be gained by applying it systematically to phenomena of the 20th century? What benefit might the emerging field of "complexity sciences" have for sociological analysis? Where does the distinction between culture and social system/structure stand now?  How does its semantic field shift as you move from one notion of culture or of social structure to another? What are the different ways of conceptualizing the relation of value and norms, and what is to be gained/lost by using one or the other? What implications does the emerging field of neuro-phsyiological psychology have for action theory? Does it short-circuit the largely untheorized nexus between behavioral system-personality system?  What might be gained by applying the ideas of Comte, Mosca, and Pareto on the circulation of elites to current trends in modern societies? How might Mannheim's concept of generational structures be utilized in analyzing social systems of varying levels of size and complexity? Would Summer's ideas on crescive and enactive institutions add anything to Jim Coleman's work on two cognate types of organizations? Why do sociologists fail to incorporate what have long been established as the several independently variable meanings of rationality in Weber's corpus? What can be done today to exploit the media of interchange paradigm?  Why do theorists confound the agency/structure dichotomy with that of freedom/determinism? Why do investigators get hung up on what I have called the pathos of ambiguity-the notion that a certain concept is so hopelessly multivocal that one had better fix a univocal meaning for it or stop using the concept altogether? 

What are the defensible criteria of good problem selection?

*For my current theory course syllabus, write dlok@uchicago.edu, subject line: 
send Social Theory syllabus. 

1. Theory defined reflexively as what theorists study, what theorists do, what theorists teach, what theory courses contain, what theory exams test. Resistance to defining theory as theory texts.

2. Theory defined as distinct from other subfields: if theory is taken as a center, other subfields are its satellites, or vice versa. Subfields each have their barriers to entry. Theory as an autonomous discipline, or a vocation. 3. Theory defined as constitutive of sociology, or of its subfields, or vice versa. Subfields overlap, or are hybridizing, or have never really been so separate as advertised. Or theory takes different form in different subfields. 4. Theory defined (favorably or unfavorably) as distinct from method, process, empiricism, research, science, normative politics, philosophical orientation, ethnographic orientation, history, or --conversely-- as suffused with one or more of these, or as fundamental to one or more of these. 5. Theory as a set of common fundamental questions or reference points, or as common fundamental powers of analysis or inquiry.

6. Theory has its own internal boundaries, such as grand theory vs. middle range theory, historical vs. systematic.

LAMONT: theory as the work of those who self-define as theorists or have jobs as theorists; autonomy vs. overlap vs. integration in areas of research; category distinctions among grand theory, middle range theory, and empirical research; if theory is taken as a center, other fields are its satellites.

WAGNER-PACIFICI: that which sociologists count as recognizably "theory"; hybridization (if distinct areas merge) vs. the idea that there was never a real split between grand theory and empirical work. HALL: theory as ways of theorizing; different frames of inquiry; if theory is a Hydra, its different heads connect to different forms of thought; theory made constitutive of some subdisciplines, or simply consumed for use in others. MUKERJI: theory as distributed among subfields, taking different form in each; theory as a subfield with barriers to entry. PERRIN: that which defines a work as "theoretical"; if culture is taken as a center, theory is one of its satellites; theory as infrastructure (constitutive) of subfields, vs. a method; theory as one among many 
analytic resources. FAVELL: theory as analytic thinking, vs. any canon of texts. ARMSTRONG: that which schools teach as theory; engaging in broad theoretical questions vs. using methodology; if the teacher/researcher is taken as a center, theory is one among many competences. LICHTERMAN: theory sets horizons of discussion (often akin to grand theory), while theoretical work focuses on objects within those horizons (akin to middle-range theory), and each subfield has its theory and its theoretical work; sometimes theoretical work becomes theory by setting a new horizon. TURNER: historical theory vs. systematic theory (Parsons, Merton); Parsonian-generation theory = classic texts, interest in the philosophical/intellectual/theoretical, nondisciplinary, vs. newer, more gendered cultural approaches = theory without classic texts, less interest in the philosophical, with disciplinary orientation. ELIASOPH: theory not always open to ethnographic or other process-oriented visions; empirical work said to be "informed by theory" or to support/challenge existing theory; theory is everywhere--nobody claims that their work is a-theoretical--but it deserves particular study; grand theory offers shared reference points, collective representations and questions. GOODMAN: if theory is taken as a center, should it accept colonization from its own satellites, or full hybridization?; grand theory's alternative worldviews provide both the core of sociology (basic debates) and links to a public outside sociology; grand theory has been seen as inevitably political: = unscientific vs. positivists, = Parsonian-conservative-politically inactive vs. politically active leftists.

HECHTER: social theory vs. intellectual history; theory vs. the normative; "theoretical" work actually has empirical implications; theory vs. research (theoretical questions lead to research); theory provides internally consistent propositions, sources of causal mechanisms, and enduring general questions shared across sociology--these fundamental questions are the only things all sociologists share; importance of research is determined by relationship to important theoretical questions.

LEVINE: that which theorists do (19 kinds; theory as vocation); theory not anti-historical (not even Parsons et al.); theory as a body of arguments, not a body of texts; famous theory texts no less empirical than other texts; autonomous theory work enhances intellectual mastery and instills discipline (Finding Problems, Justifying Problems, etc.); theory as pluralistic set of components required for creative substantive research.