Jan has given an excellent start to think about STS and methods. According to Jan, we are in a world of “messiness” “If we look at the conceptual apparatus”, but not so much “if we look at the standard set of methods (especially of qualitative research) still in use.” I wholeheartedly agree with this analysis, and I think it points to what is wrong with the idea of mess, and how mess relates to the world and methods in the first place.

The thinking assumes that in “classical social science” sociologists believed that it is the role of the social researcher to create methods and theories that show the hidden order of the world. First of all, I think that a lot of social science never believed in this logic (most vehemently, Harold Garfinkel, but also Georges Devereux), a long time before ANT and STS came along. Second, – here is my reflexivity boomerang – even a paper like John Law’s cleans up the mess, by following precisely the logic of ordered articles: introduction, thesis, discussion, conclusion. The “need” for order, is not only one of theories of order.  It comes from how writing as practice unfolds (one word after another, quite unlike the world) and how scientific writing is standardized. This at least in part has good reasons, as John Law’s lucid article shows. But even if the diagnosis were right, and we disregard the reflexivity boomerang, the treatment is too timid.

From “the world is a mess” does not follow that our methods and descriptions should be a mess. This would simply leave us with a descriptivist duplication of the world, akin to Borges’ famous map that is a copy of the territory. The underlying problem here is that the treatment is a post-structuralist reconceptualization of methods. This is fine with me, as far as this implies to stop using methods as hammers in search of nails, or as identity (as in: I am an ethnographer, I do biographical interviews etc.). But the treatment stops with theoretical thinking about methods, leaving the practice of methods intact. John Law, in sync with most of STS, still does some form of ethnography. Post methods then, is before methods. Or, as I put it in a forthcoming article: Post-method is still based on a very particular kind of doing methods, namely textual loose translations. These are methods, such as ethnography that do one large jump from the world to a text. I prefer widening our set of methods with more and other methods instead: non-textual tight and loose translations.

I would like to suggest to explore such new methods that re-order and probe and challenge the mess. These are methods that do not translate the world into a text, but rather create new worlds. It is very much like what natural scientists do: to translate the world into something different, which then becomes an actant in itself with unforeseen repercussions for the world and the social researcher themselves. This is something very different from both (post-methods and post-structural) descriptivism and doing critical research. It is different from descriptivism because it accepts that social science needs a strong take on methods. It needs to create methods, as forms of intervention and analysis that slice the world in ways that the scientists, and not the world, decide on. It accepts all the things that ethnographers and large parts of STS abhor: creating actual laboratories, doing experiments, tinker with machines, using automated recording procedures, standardizing protocols, using and even designing all kinds of media and materials and even using force to make research participants do things they would otherwise not do, make subjects object to these procedures.

But it is also different from “critical research” in the sense that such methods do not aim towards an outcome that the researchers pre-determine. Such a world is neither a world of mess, nor a world of “post-method”. Together with my colleagues of shared inc., we call it incubations. You can call it what you like, but I suggest that you at least try it.