This morning, I will be embarking on graduate contemporary theory for the eighth time in my career. Every year, it has evolved – sometimes quite significantly – making me the guy who won’t commit to a recurring syllabus and, thereby, taking advantage of the amazing teaching load UBC offers research-oriented faculty. This year, it has taken one final step forward, but that is not the specific point of this specific entry (though, I do want to revisit what happened last term with my undergrad theory course and discuss the way I’ve designed this current course in future entries…entries which I hope come at a more consistent and quicker pace). For now, I want to build a little on a abbreviated twitter thread I left yesterday. In short, the title of this essay is the subject of the post.
What is The Point of Theory?
In re-reading Abend’s (2008) piece on the meanings of theory and Turner’s (1985) piece defending positivism, alongside the first couple of chapters of Reynold’s *Primer on Theory Construction*/Homans’ *The Nature of Social Science* – all assigned for day one, the question once again was as salient as it can be. I was once again excited to read Abend’s arguments, which, in essence, put the burden, first, on the discipline to work out, politically, what exactly we mean by theory; and, second, on individual sociologists – in their reviews of papers, for instance – to not just say something like: “it does not sufficiently advance theory,” but rather be self-reflexive and specify which of his seven commonly attributed meanings of theory the reviewer implies. Both of these points seem wise, and in fact useful. That I have never seen a paper of mine reviewed this way or reviewed a paper this way, a decade plus later, I would say Abend’s practices have not been adopted.
What I didn’t like about Abend’s argument, however, kept haunting me. Especially as I read Turner’s argument for science and positivism (which, is a loaded word with myriad meanings and really not the point of this essay, so will be left alone for now). On the one hand, Abend’s first, second, and third types of theory share several key features the others do not; mainly, a commitment to engagement between theory and empirical reality. The others are exercises belonging to the humanities. To be sure, my argument ultimately is not that these do not belong in the discipline – surely they do, and some of my most favorite work comes from scholars mining the depths of a specific theorist – but that Abend does not go far enough in distinguishing what is theory – by the actual definition of the term – from that which is something else. Of course, that is not his intention or rhetorical strategy. Yet, I was left unfulfilled as I have always been in reading his otherwise excellent paper.
If the point of sociological theory is not to engage with the empirical world through some type of methodological strategy, then it is probably best not to call it theory. Why? Well, for one reason, there is a scientific community that is significantly larger than sociology that has adopted this meaning; and as we know, what is believed to be real becomes real in its consequences. Intersubjectivity is a foundational element of any community, and if the scientific community, writ large, defines theory in terms of its application to empirical facts and regularities, then who are we to take a term and use it however we want. If it was subversive, then I might understand. But, I think Abend is right: people have usurped the term in order to fight other political battles in the discipline, and this fact does not make it right or better or appropriate.
So, in a few hours, my students will be hearing that they have four tasks this term: learning substantive theoretical frameworks/theories (something I am simply no longer able to focus my primary energies on in course); how to read theory and/or theoretical-elements of research articles; how to extract, write, and draw formal social scientific theory; and how to publish the types of theory offered by Abend, Turner, and Alexander. My first thoughts about this compromise is that it is a lot. But, it really isn’t as much as one might surmise. I have taught theory in several ways for over a decade plus, and I have realized students will glean what they glean, regardless of your efforts. It is, like stats and many other classes at the grad level, there for the student’s taking and up to the student. So, substantively, I give them as much range and depth as I can and they choose their level of weekly engagement. In terms of the reading of theory, every week is split into the first/second half of the course, with the latter reserved for posting PDFs of articles with detailed notes about how people present theory in different ways when writing papers. This also contributes to the substantive goal and to the writing/publishing goals. Its the latter two that are potentially in conflict.
Abend cautions us against pushing a scientistic epistemological and ontological solution to the so-called semantic problem. He worries practically and politically about the outcome, for good reason. But, I am less worried about these problems. After all, as Turner has said: if we aren’t doing science, then what are we doing that is unique? Critical literary analysis? Covered in another department whose training is primarily focused on this? Philosophy of science. Covered. Ideographic historical case-studies? Covered. We study societies, social organization, social behavior and attitudes and feelings, and the like. And, our contribution to social problems comes most cogently from our best methodological practices. Turner’s extremism, unfortunately, obscures the better parts of his argument. Arguments I know well having been his student and also having countless conversations with him. He is much less ideologically-rigid than his polemics presume.
First, Turner is agnostic methodologically. For many reasons the word positivism has become synonymous with quantitative orthodoxy, but neither Comte nor Turner think this way. Positivism for both simply states: we should be working towards identify the key properties of the social world, the law-like relationships regardless of time and space, and proximate rather than ultimate causality. These are lofty goals, and we can debate if there are laws or not (there are, incidentally). But, neither staked out a side as to how we achieved these goals. Turner is somewhat un-empirical, but he prefers historical and ethnographic data to stats (he is quite critical of complex modeling strategies that pretend to be theory). Comte argued that naturalistic observation was as important as anything else.
Second, while Turner says there are better strategies for building cumulative theory, he recognizes that any theory committed to empirical analysis is, by definition, scientific. It just falls short in the ultimate goal of cumulative knowledge. That said, like many general theorists of his generation, middle range theories, historical explanations, analytic schema, etc. all serve as inspirations for the the sociological imagination in building systems of causal laws. [Side note: I am not even sure Turner is committed to laws, as laws are empirical regularities and less abstract than the systems of interrelated propositional statements he prefers (see 2010, 2010, 2011). His broader argument is we know a lot more than we act like we do, and we could teach our students a common theoretical language before they move on to their own specific interests and without robbing them of the sociological imagination].
Third, Turner accepts that there are other goals of science that can serve as criteria for evaluating the value of theory. In Reynolds’ (1971) archetypal text on the construction of sociological theories, several goals are listed besides cumulative knowledge, and overlap/extend Abend’s types 1, 2, and 3 theories. The first is description/classification. Theories can be descriptive! This is actually nice. Note description for description sake is the weakest criteria of theory-building, because, presumably, one’s taxonomic efforts should also contribute to explanation and, potentially, prediction; the next two goals.
In terms of explanation, Reynolds, like Turner, is agnostic. He clearly prefers those explanations that are independent of time/space, but also realizes historical explanations of specific cases are useful. This heterodoxic stance also pries open space for qualitative research committed to scientific rigor. In my own experiences, qual is essential to revealing mechanisms, processes, and, of course, meanings that should – in a perfect world – inform future deductive research in terms of the questions asked and the instruments developed to answer them. This is not always the case, but I would argue this is more a function of a false divide between quant and qual and, worse, the idea that theory shouldn’t be rooted in scientisim. Again, other types of activities not scientistic are not less than, subordinate to, or non-sociological; they just aren’t theory. If we push back and see explanation as a central goal to theory, then description strives to reach explanation. Quant folks maybe read qual folks more, or collaborate, to improve surveys and analytic strategies; and new qual folks emerge to deal with the ever-present gaps science, as an epistemology, purposefully creates.
The bogeyman, here, is prediction. I won’t dive too deeply here. But, I do think sociologists can predict some things. And, by prediction, I am going by other science’s standards – e.g., biologists can predict the general time a leaf will fall and specify why their prediction is as such, but they cannot give you a day or time. We know the conditions under which ethnic or class conflict should arise. Like the leaf falling on Tuesday, conflict may not happen next week or month; it may simmer. And like a chance fire burning the tree to the ground before the leaf falls, other intervening variables may change the trajectory of the conflict. This is not precise, to be sure, and thus sociology does not necessarily lend itself to an applied physics, but we know a lot about issues like mental health stigma, poverty, formal organizations, and so on. We can actually predict, within reason, more than we presume. More broadly, this is but one of four goals I’ve delineated, and as such, not a make-or-break criteria for sociology’s status as a science.
The final two goals Reynolds discusses are “understanding” and control. He dismisses the latter, arguing that scientists have a tough time controlling lots and lots of things (e.g., earthquakes), so most of what we study is also probably quite difficult to control. But, understanding, for Reynolds, is related to the construction of paradigmatic (for lack of a better word) systems of causal relationships. It is more a framework that a set of scholars work within that informs their decision-making. Evolution in biology, for instance, is a perfect example: it guides most of the assumptions, questions, and explanations biologists deal with. Again, one criteria among others.
To these six goals (cumulative knowledge; description; explanation; prediction; paradigmatic; control), I would add understanding as a part of the theory-building process, though perhaps not really a criteria for evaluating theory (I continue to think and evolve on this). I think interpretivist sociology, when done well and with serious rigor, can become explanatory over time. Scholars committed to a particular milieu, process, set of actors, etc. can build, over several projects, clearer explanations. With abductive strategies, this logic is already built in to the approach (Tavory and Timmermans 2012). However, I see a place for understanding as a goal of science too. Especially when the scholar abandons the outdated pretense of the naive observer. In a highly diverse world, where little pockets of the world are distant cognitively and geographically, illuminating the discrete attributes and meanings may, in fact, lead to powerful social science. I know our work on suicide has benefited from trying to understand a community’s meanings surrounding suicide, and we believe with more research, much of what we found can be made more abstract and generalized. Not so much the details on the ground, unique to Poplar Grove, but the processes and what not. In this sense, understanding can begin the process of developing practical tools for dealing with social problems.
One could argue these points are implicit, baked into the sociological enterprise. As with everything I muse about re theory, it all returns to how we teach it; and, by way of, how we train students to use and be theoretically-minded. The way we teach it is a mess. It lacks coherence and consistency. If we all began with the guiding principle of scientific criteria, we would no longer be able to teach the canon without re-engaging with what it is we are even teaching. If we began with science, then contemp theory courses would cease to be eclectic, arbitrary exercises in pet theorists, ideological axes, and the like.
This underlying epistemological stance will, hopefully, guide the course I am teaching this term. The goal is to provide practical tools for those who gravitate towards writing more theoretically pieces, but really for all students to be able to more clearly articulate what their theories are, how they inform their research, and what their contribution is. I have always made clear that my position is not the only position, and that they need to decide for themselves what kind of sociologist they want to be. Moreover, I do not discriminate against those pursuing Abend’s other types of theory. But, to accomplish the four goals of the class, and make well-rounded, competent, prepared sociologists, it seems scientism is the best thread to tie it all together.