Foucault and Design Patterns

Every so often, I crack a book by Foucault: not to advance my fortunes as an amateur social scientist, but to try to bridge some discursive gaps between my idiosyncratic social theories and the views of “the regular guys”. I suppose you might think that there is surely some easier and more effective way to “get across”, like reading the Bible or taking up yoga or campaigning for Mitt Romney. But Foucault books are readily available, readable enough, and actually not without their parallels in the more “pragmatic” realms of human activity. For example, the epistemological statuses of Foucault’s theories are almost as questionable as the epistemological statuses of the disciplines and subdisciplines he classifies, to the point where he himself described his work as an “antiscience”; if it seems that this makes his writing totally useless when confronted with the rigor of “harder” disciplines, I’d like to advert to a precinct of one of those disciplines, “design patterns” in computer science, as an example of how a discourse orthogonal to “scientificity” serves a useful cognitive purpose. I hope you’ll find, as I have, that the parallels to Foucault are significant and striking enough to justify this rather outre comparison.

Design patterns are an idea originating in the study of architecture. The architectural theorist Christopher Alexander thought to systematically work through the successful elements of various architectural designs and develop a “pattern language”, a set of meta-blueprints that could be tweaked such that new architectural designs might build on old strengths. Although Alexander himself later decided the architecture designed using this method was rather sterile, design patterns were “in the air” long enough to percolate down to computer researchers trying to decide how best to make use of the new “object-oriented” computer languages, which were originally designed to model real-world situations using an “object” to represent each element of that situation. It was decided that objects offered fertile ground for implementing something very much like Alexander’s pattern language: a successful object would share very many features with other objects designed for similar tasks, so why not cook up patterns for such scenarios?

The computer world is, of course, full of fads that become buzzwords for ambitious managers: however, design patterns were one trend which had legs, since they operate at the interface between human ingenuity and necessary structure and help to make programs easy to develop and intelligible. In fact, quite a lot of other phenomena operate at a similar level: the interaction between the increase of scientific knowledge and “technological” exigency creates orders which are neither “nomological” instances of exceptionless laws nor matters of arbitrary choice — and this is precisely where Foucault situates his works. I don’t think it’s beyond reason to say that he provides us with “design patterns” for organizing knowledge; in fact, I think it captures very well his insistence that the fundamental structuring of disciplines is not a matter of veridicality per se, but instead a place where the complex forces of history channel and multiply their energies through the interaction of knowledge and Foucault’s version of power. (And who knows, it might even be some comfort to the Panoptically-monitored programmer to conceive of the great intellectual considering relatively abstract factories and their effects on flyweights).