Four Debates for the Price of One: Structure vs. Agency

I’ve talked before at some length about my issues with the terms structure and agency. I think they often shed more heat than light, and often represent philosophical positions about how to interpret the world more than empirical alternatives. Recently, I had the opportunity to read Sewell’s classic piece, A Theory of Structure: Duality, Agency and Transformation. Reading through his discussions of the structure vs. agency debate clarified my own thinking a great deal, and his own version (a slightly critical combination of Giddens and Bourdieu) is one of the least bothersome I’ve yet seen*. For example, Sewell does a nice job of laying out one debate often subsumed under the heading of structure vs. agency, but really orthogonal to the debate many sociologists want to have: the material vs. the cultural. In thinking through that issue, I came up with a list of 4 debates that sometimes fall under the heading of structure vs. agency:

  • The material or economic vs. the cultural or ideational.
  • Determinism vs. Free-will.
  • Stability vs. change.
  • Intentionality or purposivity vs. unanticipated consequence or accident.

    The first and the last are a bit stranger than the middle two. The material vs. the cultural is often, as Sewell notes, an issue about what counts as a structure. For Anthropologists, culture is the structure of most interest. For some sociologists culture opposes the material as an alternative to deterministic theories like some kinds of Marxism.

    Intentionality vs. accident is, on the other side, about what counts as agency. Does agency refer simply to the capacity of individual action to alter future circumstances (i.e. structures)? If so, then intentionality is irrelevant. In Foucault’s excellent phrase, “People know what they do; they frequently know why they do what they do; but what they don’t know is what what they do does.” (quoted in Dreyfus and Rabinow). Sewell, on the other hand, wants to restrict agency to the capacity for intentional action to intentionally change structures (as I read him – in particular, he’s interested in how actors take schemas and apply them to new situations or apply them in new ways to old ones, thus generating new structures).

    The middle two are more self-explanatory, I think. Determinism vs. free-will goes back at least as far as Durkheim (in Sociology, that is). It’s pretty clear in the early parts of The Rules of the Sociological Method where, if I recall correctly, Durkheim directly opposes social facts (which exist outside of individuals and constrain their actions) to free will, thus setting this variant of the structure vs. agency debate at the center of the discipline.

    Stability vs. change can be seen nicely in Bourdieu’s work, and Sewell’s criticism of it. Bourdieu creates an elaborate and useful way of thinking about stability – posit a field with a certain logic, a set of positions, a set of actors who possess certain kinds of resources, and the capacity to use those resources more or less effectively for the given field, and you can see what kinds of outcomes are stable. His framework is a little trickier when it comes to understanding change – as Sewell notes, change comes from exogenous shocks to the system rather than being produced endogenously by agents. The habitus is a reproductive concept – it’s about how the system stays the same rather than how it changes. Or, as a I said a few days ago, Bourdieu turns agents into structures (indeed, structured structures that are also structuring structures…). Sewell wants to add back in some ability for individual (purposive) action to alter structures without any exogenous shocks.

    So, in sum, another reason I hate the whole structure vs. agency debate is that it’s not one debate. It’s several debates, nested into one big dichotomy. But perhaps if we parse out what those underlying debates are we can get somewhere a bit more productive.

    Thoughts? Am I missing any other debates that fly under the banner of structure vs. agency?

    * I know, damning with faint praise. But I do mean it – most of the time these debates give me headaches, at best.


    1. Ben

       /  August 4, 2009

      Seems we are working on similar things. Why I mentioned the Emirbayer & Goodwin the other day is that I found it useful an interesting crit of network theory for ignoring culture and agency.

      I too liked Sewell’s paper and found it much less bothersome than most. No ‘structured structuring structures’ there. And no page-long sentences either, which is a relief.

      Michelle Lamont has some interesting comments on Bourdieu somewhere or other (can’t remember where at the moment)which I found enlightening, and of course Alexander does that too.

      Re your comments determinism vs free will, the paper by Sharon Hays (1994) on ‘the sticky problem of culture’ in the structure/agency debate is very interesting on the ideological component of that point.

      I’m currently reading Jens Beckert who has another take on Giddens in his book, ‘Beyond the Market’. He draws on pragmatism (via Hans Joas’ Creativity of Action) to examine innovation in the economy. That might be interesting for you. Sorry no time to go beyond giving the refs today.

    2. liminal

       /  August 4, 2009

      I highly recommend reading Sherry Ortner’s 2006 Anthropology and Social Theory: Culture, Power, and the Acting Subject. Duke University Press.

      I think you would find it very interesting. 🙂

    3. I just noticed your blog when I found your recent performativity post on google, great blog you have!

      A debate that is missing is the one between micro- and macro-agency, usually in the form of individual vs. collective. As I noticed you are in a way favouring the Callon version (that I really find sharp, too), I have to point you to the Leviathan-Paper (with Latour, 1981) which tries to map both sides of that debate onto on single version of agency. Also I would like to note that Niklas Luhmann found a nice explanation of what the term agency is helpful for: as the social is an actorless operation referencing only communicative events, agency is used as a scheme of attributing the social to identifiable entities: agents.

    4. @Jan – thanks for the suggestions! I had heard about the Leviathan paper ages ago but forgot about it, I’ll go check it out. Same for Luhmann.

      Also, re: agency – I think the connection between “agency” and “blame” or “responsibility” is deep and should be made more explicit. For me, it implies that before you can answer “did X have agency?” you have to know for what purpose you are trying to answer the question.

    5. Michael Bishop

       /  September 14, 2009

      I think it is sometimes useful to refer to these dichotomies, and even to claim that someone overemphasizes one of them. But backing up that claim requires determining where they make different predictions, and then doing the necessary empirical work.