“A theory is a program of perception.” (Pierre Bourdieu)
I’m holding a talk on how to write theory for students working on their research projects this afternoon. My aim will be to show them how to use theory to bring together the results of their literature review. In fact, I intend, albeit with some trepidation, to pass on Ezra Zuckerman’s advice: “Never write literature reviews.” Instead, they should use the literature to frame an interesting problem; or, as I like to put it, they should write a theory section. Now, some of their teachers will disagree with me about this, and where there’s a disagreement between a writing coach and a thesis supervisor, the supervisor is, of course, right. In those cases, I will suggest that students write the literature review as a narrative that leads up to the theory, which will be a representation of the current “state of the art” that the student wants to practice. Everyone agrees that the literature should inform your research, and we see this formation of your perspective very clearly, very explicitly, in the theory section. Theories shape our expectations of our object; they condition our observations of practice; and they frame our models of reality. Lets look at each of these functions in turn with our reader in mind.
Theories are systems of expectation. Just telling your reader about your object (the social practice your project is about) doesn’t suggest anything very specific about what your analysis will show. Your reader needs to know how you will be approaching the material. So, by invoking theories of organization, or culture, or sensemaking, or institution, or finance, or governance, or discourse, or … you activate particular expectations in the appropriate reader. A reader that doesn’t recognize your theory will not really know what to expect, but, to be frank, they aren’t part of your intended audience. Address yourself to a reader that knows enough about your theories to form relevant general expectations of what your analysis might show. And then write your theory section with the aim of making those expectations more specific and vivid. Theories guide our thinking; they have consequences for what we think will happen in the analysis. They should lead us to reasonable hypotheses — first, what Ezra calls a “compelling null” and, then, after a bit more of what Karl Weick calls “disciplined imagination”, some interesting testable hypotheses to investigate. We might also say that theories are conceptual frameworks.
Concepts are categories of observation. For Kant, they really did discipline our imagination, ensuring that things appear to us as knowable objects in experience. Indeed, one important function of a good concept is to make some things immediately recognizable to us. Other things we have to think a little more about (and we use our concepts to do this thinking.) That is, concepts lead us more or less directly from a brute experience to the conclusion that something is true in the world. Our theoretical concept of, say, “identity” can lead us from statements that were made in an interview to claims about who the interviewee thinks she is. The concept includes a sense of the indicators that lets us apply it to experience, to our data. This makes them important analytical tools but, in the theory section, we just remind the reader of the possibilities implicit in the concepts we intend to use. We make clear what sorts of data would make the sorts of claims that use these concepts true. We activate the reader’s imagination and direct it towards particular features of empirical reality. Bringing together several concepts, we can go on to build a model.
If concepts are the working parts of theories, models are collections of concepts, along with a set of assumptions that mediate between the theory as a whole and the specific practice you are studying. They are theoretical constructs. They do not model the objects as such (the things in themselves) but the relationships that the theory establishes among them. That’s why we sometimes say that models are never “true”. They are merely instruments for interpreting and predicting features of the practices you are studying. They will match your observations only approximately, but these approximations will help you and your reader make sense of what you see. It may be useful to think of your model as the “output device” of your theory section. It applies your concepts with enough specificity to let you generate hypotheses on the basis of some well-chosen assumptions. The literature review (if you choose to write one) leads to your theoretical framework (of concepts) which then lets you build a model. The model, in turn, frames your hypotheses, which really are the actual “output” of theory, making the expectations of reader concrete, and setting up your analysis.
The most important thing when writing about a theory, is to imagine a reader who is already familiar with it. This makes writing about theory very different from writing your background or analysis sections, where you should not assume your reader already understands what you will say. At the end of he day, the reason that you should carry out a literature review is, not to learn a new theory (though you may of course need to do this too), but to get to know your reader, who has presumably read the same books and articles that you have. Obviously, that’s not really going to be the case — everyone has a slightly different reading background — but you can presume that your reader is as “aware” of the literature as you are. Having spent some time in the library, you know what’s out there, and you can identify the major figures. So can your reader, who will feel at home among the texts you cite. It is not presumptuous to assume that your peers have read and understood the same things that you have read and understand. That’s what makes you peers. You should write with the familiarity that comes with this reasonable presumption.
Thank you. Very well argued.