The triple crisis of sociology
Alvin Gouldner published The Coming Crisis of Western Sociology just 45 years ago, in 1970. Like most great theories it was the Owl of Minerva, which came with the dusk. Gouldner predicted the fall of Parsonion (quasi)-positivist functionalism and the raise of a more reflexive sociology. Well, by 1970 Parsonion sociology was dead and sociology was entering its most exciting epoch. People like Alvin Gouldner, the young Marty Lipset, C. Wright Mills, S.M. Miller, Lee Rainwater, Pierre Bourdieu, David Lockwood, Ralph Miliband, Claus Offe, the young Ralf Dahrendorf – not a Lord yet – and many others offered a refreshingly reflexive and critical sociology (I could add a few names from then-socialist Eastern Europe, like Zygmunt Bauman, Leszek Kolakowski, the Praxis group in Yugoslavia). This work during the 1960s and 1970s attracted undergraduate and graduate students not only in exceptionally high numbers but also of unusually high quality. At the time Gouldner predicted the coming crisis in sociology it looked like the crisis was over, the discipline found its way out of the dead-end street of structural-functionalism, it was blossoming, it was the Mecca of radical students – and those who tended to be radical during the 1960s and 1970s tended to be smart too.
Sociology 101 used to be a boring telephone directory of impenetrable, empirically untestable concepts, now it became the ground of political mobilization. Suddenly enrollments jumped. The number of sociology majors increased almost 350% between 1950 and 1972. At its peak, in 1971-72, there were 4 sociological degrees for every 100 bachelor’s degrees conferred – more than twice the current level (see figure).
By the turn of the century the crisis had arrived. Social sciences, starting with economics, followed by political sciences underwent fundamental changes; neo-classical economics, rational choice theory and experimental research design appeared to be victorious and sociologists are still searching for ways to respond. Students lost interest in radical theories. Many became conservatives, worried more about their vocational destinations and their pension funds. Sociology departments often are struggling to get enough majors to justify the size of the faculty and they often offer “sexy” – and often not very demanding – courses on issues vaguely related to what used to be the core concern of the discipline just to get enough enrollments. If one today compares sociology with economics and political science our discipline appears to be in a triple crisis: it lost its political appeal (and radical mission); it could not find, so far, an appropriate response to the methodological challenge from economics and rational choice; and the discipline appears to be in utter confusion over whether it has a common theoretical core (which are the “great books” every sociologist should be familiar with), and even debates whether such a core would be desirable.
The political crisis
During the mid/late 1960s and early/mid 1970s sociology attracted radically-minded young faculty and students. It was the “thing to do” if someone was interested in radical reform, or even revolution.
After I took up my chair as foundation professor of sociology at Flinders University in South Australia in 1976 the reform minded attorney general, Peter Duncan, asked me to visit him in his office. He reminded me of the “revolution” in Portugal. That was not carried out by the proletariat, but by the army. “Well, can we expect the army in Australia to be the revolutionary force? Forget about it,” he continued. “But I am in charge of civil servants in the state. I will send them to Flinders to major in sociology and your task is to raise their revolutionary consciousness.” And this happened. We got three times more applicants to Sociology 101 than we could seat in our largest classroom and indeed after the first year many of our students were radical reformers, and believed in a better society. Well, in 1977 the Labor Party lost the elections in South Australia, Peter Duncan started his law firm and that was the end of radical reform in South Australia (and the beginning of the end of my career as the “charismatic leader of sociology” at Flinders).
Not all sociology professors (even not the young ones) could go along with the radicalization of students. I was a postdoc in Berkeley in the spring of 1965. I took a course with Marty Lipset and my advisor was Nathan Glazer (what a lovely man), both former “lefties” turned rather conservative as a response to student radicalization. In a nutshell faculty by the late 1960s or early 1970s (especially the older ones) tended to be conservatives (or at least former lefties turned conservatives or “neo-liberals”) while our students were left-wing radicals. Much to the dismay of my senior colleagues at Flinders I went along with our students – which blocked my ascent to be a Dean or Vice-Chancellor (or even an effective department chair), and led me to escape to the US in 1981. Today the situation is the opposite: we still have quite a few 1960s radical faculty, but the students tend to be more conservative than the faculty. And if you are a conservative why on Earth do you want to major in sociology rather than in economics or rational choice political science? So suddenly our problem is not so much that we cannot find enough seats for those who want to major in sociology, but we may not find enough students to fill out lecture halls. This is what I call our “political crisis. This cuts both ways: we cannot attract students in sufficient number; and sociology is less and less likely to offer scenarios for radical social reform.
The methodological crisis
This also has something to do with the “methodological revolution” in social sciences. Ever since the terms social sciences was invented the disciplines labeled this way tried to justify why they qualify as “science.” For August Comte sociology was supposed to be the “science of society,” which studied social phenomena with the same rigor as “scientists” study nature. None of the social “sciences” could deliver on this promise and during the past few decades this tension became even more intense. After all, only those disciplines deserve to be called “science” which can establish “causal relationships” between “variables.” But can the study of social (and economic) phenomena make believable claims about causality?
Max Weber suspected we cannot, hence he opted for “interpretative social sciences.” But we could never get the question of causality out of our minds. In this respect I am also guilty as charged. Whenever I supervised an honors thesis or a Ph.D. dissertation I was never satisfied when my supervisee wanted to answer the question “how.” I was pushing him/her to explore the “why.” “How” is a lousy question, it can never be falsified so if you want to be scientific (and formulate your researched design in a way when you identify under what circumstances you accept “defeat,” or you acknowledge that your hypothesis was false – I learned this framing of the problem from E.O. Wright) you have to confront the question of “why.”
The trick however is that you can test causality only if you have a random assignment, when you conduct a properly designed “experiment.”
But social sciences – in particular sociology, but also to some extent political science and even economics – was better in random selection than random assignment. Our major method was survey research based on random samples. And we made quite astonishing success in random sampling. If the sampling was random enough we could predict for instance the outcomes of elections with tiny samples (of a few hundred) for populations of hundreds of millions. That was a formidable achievement and improved our knowledge about social, political and economic processes tremendously but did not get us an inch closer to test hypotheses about causality. You can test hypothesis about causality only if you can assign part of the population to an “experimental group,” which will be exposed to a certain stimulus (“treatment”) and leave the rest in a “control group,” isolated from such a stimulus.
Survey research – in contrast with experiments – is always suffering from the “selection problem.” You can never tell with any scientific rigor whether the outcome in population A/ is different from population B/ because A/ was already different in status quo ante (we do not know the condition of population A and B in status quo ante), or because it received a different “treatment.” A simple example: we know that people who are married live longer. But how can we tell whether they live longer since they were married, or people who were healthier, hence could expect to have longer lives were more likely to marry? Only if I could assign populations aged 14 into an experimental group who will get married and a control group who can never marry and revisit their heath condition X years later could I give a rigorous scientific answer to the question of causality (and such random assignment is of course impossible).
Social researchers tried to dig themselves out of this hole. One possible solution was to identify the “causal mechanism,” hence to write a narrative why x may actually cause y (married people drink less and eat more regularly, hence they live longer). This is a noble effort – I tried it a number of times in my own research – but not very persuasive for “normal scientists” (the sample selection problem is still with us). Survey researchers tried other technologies: panel studies, or life history interviews, both are good ideas, but do not resolve the fundamental problem (in panel studies you lose population over time, in life history studied you have a serious problem with “memory”; people tend to remember their lives rather selectively).
Economists and increasingly political scientists try to address this question with lab experiments. Lab experiment with a completely controlled environment is a great solution, but comes with an incredible costs: external validity. Lab experiments have a great random assignment, but fall always short of random selection so we do not have the faintest idea how we can generalize from the findings of lab experiments where the subjects usually are middle class college students… (Another “solution” is so called “field experiment” where random selection can be applied, but it is nowhere near random assignment).
Nevertheless, economics and political science are on the go, they offer a logically coherent (though as I will elaborate on this later on empirically problematic) solution to the causality problem while sociology is on the defensive with no good answers, hence in a methodological crisis.
The theoretical crisis
Sociology is not in much better shape theoretically and arguably was on a downward slope since the 1980s. I am certainly not nostalgic about the Merton-Parsons kind of unified theoretical orthodoxy. With the fall of structural-functionalism in my view a healthy theoretical dialogue replaced this, mainly dominated by Marx-Weber debate but leaving enough room for alternative challenges, to symbolic interactionism and ethnomethodology.
I have to confess even in the “good old days,” in the golden years of the 1960s or 1970s, sociology faculty often had a good fight over whose authors one should include in the required sociological theory course. Now there is even less agreement especially as sociology, desperate to retain a constituency, broadens up to interdisciplinary programs such as women’s studies, African-American studies, Asian-American studies, Chicano studies, cultural studies, and so on. These are all legitimate fields of instruction and scholarly inquiry, which should have their place in universities, but including them into sociology blurred the disciplinary boundaries of sociology.
The comparison with economics and to some extent with political science is instructive. Economists seem to have a great deal of consensus about what constitutes the theoretical base of their discipline. Almost all economists I know (and virtually all economics departments in the US, but increasingly even in Europe) seem to have a common understanding of why students have to take first Principles of Microeconomics and Principles of Macroeconomics to be able to deal with the rest of the more advanced courses. There is also little disagreement over what should be taught in these courses, the syllabi are so standardized that you can expect any economists with a Ph.D. to be able to teach any of those courses without much preparation.
Before I go too far praising economics (or rational choice political science) let me note the astonishing neglect of “classical” theorists in both disciplines. At Yale I taught a rather popular course mostly taken by seniors and often economics majors on “Varieties of capitalism,” which included readings from some of the classics, like Adam Smith, Karl Marx, and John Maynard Keynes. Much to my surprise most of the economics majors just about to graduate had never read any of these texts. Later I was told by economist friends that one can get a Ph.D. in economics in respectable departments without any significant exposure to the classics. So yes, there is an apparently coherent theory in economics but the coherence is probably so persuasive since students are not confronted by the big controversies of the classics. Those controversies may be old, but it’s too early to forget them – they may come back and haunt the discipline, as Keynes – and Marx – obviously did with the recent global financial crisis.
In contrast sociology departments either cannot agree what an introductory course should be and instead offer a range of electives with often strikingly different theories and epistemologies, or if they decide to require one intro course: something like a Sociology 101 fruit salad, a mixture of sexy topics and the usual boring telephone directory of “basic concepts.” Is economics doing it the right way, or is it sociology which solves the problem of “introduction” to the discipline more reasonably? I will come back to this question in the last section of this note, but it is rather clear economics does a better service to establish economics as a discipline than sociology does. I had a conversation with a chair of a sociology department, a person who would like to see sociology as “normal science” and who dreams about a Sociology 101 which would look very much like Principles of Microeconomics. Pluralism is great as long as those on different sides of the barricades can (and are interested to) talk to each other. But one can argue sociology is on the verge of chaos where the lines of communication tend to be breaking down.
Even more troubling: as we lose our agreement over the “classics” of our field we become less certain about the questions our discipline should pose. Sociologists once had a reasonable agreement over those problems they “owned,” such as inequalities (in power, income and life-chances, by class, race and gender), occupational and educational attainment, social mobility. Now we not only have difficulties identifying what our research questions are but, much to our embarrassment, economists (and political scientists) appropriate from us what used to be our turf. Is it not painful that the most important books on social inequality during the past decade were written by economists, like Thomas Piketty and Joseph Stieglitz? Are we left behind?
Way out of the crisis?
Let me conclude this rather pessimistic note by revisiting the virtues and strength of the sociological approach to social reality and warn my colleagues to be careful to imitate the new trends in economics and political science.
The strength of the sociological approach was reflexivity.
First of all there is a long tradition of sociology starting with Karl Marx (“The ideas of ruling class are in every epoch the ruling ideas”) followed by Karl Mannheim (“… opinions, statements, proposition and system of ideas are interpreted in the life situation of the ones who express them”), not least at all Alvin Gouldner (“The future of intellectuals and the raise of the new Class”) to wonder who the speaker is, what the (political) role of the sociologist is. As long as sociology finds the “voice of the voiceless” they will find their constituency. True, students shifted to conservatism (though after 2008-9 there has been some increase in the discontent with the inequalities of global capitalism — we are the 99 percent). As sociology returns to the concerns of the majority – class, racial and gender inequality, power, poverty, oppression, exploitation, prejudice – it will find its constituency and the good old days when students were sitting in the steps rather than leaving empty chairs in lecture halls may return. Michael Burawoy’s call for “public sociology” is a cautious call for this and in comparison with many sociology departments Berkeley is doing quite well, with full classrooms and high quality graduate students.
If sociology wants to retain its political mission it may not follow the trend set by economics. In Joseph Stiglitz’s words: “Economics had moved – more, than economists would like to think – from being a scientific discipline into becoming free market capitalism’s biggest cheerleaders … [And since e]conomics supposed to be a predictive science, yet many of the key predictions of neoclassical economics can only be rejected…. If the United States is going to succeed in reforming its economy, it may have to begin by reforming economics.” If Stiglitz is right sociology should not reform itself a la the model of neo-classical economics, but if our societies are to be reformed into better ones sociology also ought to be reformed and it should somehow return to its political mission, it should recapture from economics the investigation of big social issues, the critical vision that was so characteristic of the classical sociology of Marx and Weber. (I labeled this at one point – not with much success among my follow sociologist – “neo-classical sociology,” but what I mean was a return to the Marxian-Weberian critical macro-sociology in sharp contrast to the cheerleading of capitalism by neo-classical economics. Neo-classical sociology was supposed to reject the stereotype that economists are smart, sociologists are do-gooders, economists are dealing with the big issues, sociologists with the nitty-gritty).
Many of our colleagues try to resolve the methodological crisis of our discipline to turn sociology into a “normal science” much like economic of rational choice political science model behavior (increasingly relying on data from lab experiments), rather than describe reality with as much precision as possible. But as I pointed out lab experiments are wonderful to test causal hypothesis, but have a fatal problem with external validity and this may be the deepest reason why so many the “scientific predictions” of neo-classical economics proved to be false.
My dear colleague, Gilles Saint-Paul from the Paris School of Economics some time ago gave a faculty seminar at NYUAD and posed the question: is economics a science? He gave a persuasive answer: how could it be when it uses poor quality data and models that cannot be falsified? Gilles suggested: economics is a “cultural activity,” it frames the terms of the debate rather than offering falsifiable predictions.
As already confessed I also find the question “why” a more rewarding one than “how” and I have difficulties accepting as good social research that which is not falsifiable. But like Weber wrote about objectivity as “objectivity” I tend to write about social sciences as “sciences.” None of the social sciences are “sciences” (if science means a body of proposals where causal relationships can be tested). Social action is “voluntaristic” in the Hobbesian or Parsonion sense of the term, hence it always assumes an “agency” who can make choices (though under given circumstances). As Marx so astutely observed (writing the final world on the controversy between “free will” and “social determinism”): “Men make their own history but they do … under circumstances … given and transmitted from the past” People make choices and these choices are only in stochastic and not in deterministic relationship with their existence. Weber is right, we can interpret what people do, but we can never tell which of their action is “rational,” we cannot predict what rationally they can or will do. In this respect interpretative sociology is ahead of rational choice scientific economics (or political science) and sociologists would make a mistake if they were to imitate their more “scientific” economist or political scientist colleagues.
There is another respect in which sociologists have an advantage over other social “sciences”: What attracted me to sociology was the critical reflexivity about data sociologist tends to use (and this is often even truer for qualitative researchers). Once I was working with an outstanding scholar who carried out survey research all his life and he was well in his fifties when he confessed to me he never conducted a personal interview in his life. This blew my mind: how can one believe what people’s response to a silly survey question will be unless one asks those questions and sees the embarrassments of our respondents as they deal with the question we derived from some irrelevant social theory rather than from social life. Ethnographers educated by Howard Becker knew it better: one has to “immerse” in social conditions before one knows what the right questions are. I learned from Bob Emmerson, Jack Katz, Elijah Anderson, Mitch Duneier and some other ethnographers how much care one needs to have to record ethnographic observations (and interviews) before one can be confident one did indeed catch social reality. Well, there are survey researchers who do the same – my point is not to glorify ethnography over survey research, just to acknowledge generally the critical self-reflexivity in sociology about what data means, how they were collected.
My conclusion is sociology is better off if it accepts it as “science” rather than Science properly speaking. Yes, we should ask “why,” but remain skeptical about how good our answer to this question will be and economics and political science would be better off if they would learn some modesty in this respect from sociology.
So what is the bottom line of this short paper? Sociology is indeed in a triple crisis. It responds the wrong way to “scientific” challenge coming from neo-classical economics and rational choice political science. It either imitates them or moves into trendy interdisciplinary fields just to regain its lost constituency. My own recommendation is to return to the classical tradition of Marx and Weber, when sociology was able to confront to big issues, not to accept the role of do-gooders while economists are the “smart ones.” The response of sociology to neo-classical economics and rational choice political science may be the wrong one – either to be another “normal science” or to abandon rigor and become the PC narrative.
Why not return to the classical tradition of Marx and Weber when sociology asked the great questions and was in its reflexive, interpretative mode a serious challenge to economics (and the just-born political sciences). Why not a left-leaning, critical, neo-classical sociology?
 Karl Marx, The German Ideology, in Marx and Engels Collected Works, Volume 4. New York, International Publishers, 1976, p. 59.
 Karl Mannheim, Ideology and Utopia. New York, Harvest Book, 1936, p. 56.
 Alvin Gouldner, The Future of Intellectuals and the Rise of the New Class. New York: Seabury Press, 1979
 Joseph Stiglitz. Freefall. New York: WW. Norton, 2010, pp. 238 and 241.
 Marx and Engels Collected Works, Vol. 11  1978. New York: International Publishers, p. 103.