I see Foucault as more a philosophically informed and oriented historian than as a philosopher in any traditional sense. He typically writes what he calls “histories of the present”, meaning that he starts from what he sees as an ethically intolerable practice of contemporary life (e.g., the treatment of the mad or the system of imprisoning criminals) that, despite its obvious flaws, we tend to see as necessary given certain general views our society holds (e.g., that madness is a medical condition, that prison is the only humane form of punishment). His histories are genealogies showing that the view allegedly justifying the practice is a contingent feature of our society that does not impose a genuinely normative limit on what we think and do. (A genealogy is a diachronic causal story, usually also accompanied by synchronic archeological analyses of the conceptual structure at various key temporal points.) Foucault’s histories are philosophical in that they require critical discussion of philosophical views, but he does not put forward his own philosophical views in any traditional sense. At most, he sometimes constructs an ad hoc theoretical apparatus (e.g., a “theory” of power) designed to expose the limitations of a view he is criticising. But once the critical points are made, he is happy to abandon the apparatus, which functions like a scaffolding that’s removed when the work is done. Toward the end of his life, Foucault did move toward a conception — akin to that of the ancients — of philosophy as a way of life. But he does not seek a body of theoretical truth.
…
As to historical accuracy, Foucault does paint with a broad brush, often with a limited palette of data. On the other hand, his goal is not, as it is for many historians, a meticulous delineation of “exactly how it was”. Rather, he’s after a general interpretative framework that will shed critical light on current practices. Such large-scale work must still answer to the facts, but it is less vulnerable to simple counterexamples. So, for example, when Foucault claims that confinement of the mad in asylums expresses the Classical Age’s distinctive conception of madness, we can’t refute him by simply pointing out that there were cases of confinement well before the Classical Age.
Foucault’s relativism and skepticism is always very specifically local. His critiques are aimed at certain psychological and social scientific disciplines and associated practices. He does not try to undermine knowledge or reason in general (e.g., in math, physics, chemistry) and even allows that disciplines thoroughly implicated in the network of social power (e.g., economics) still produce bodies of objective knowledge. As to obscurity, Foucault has his faults, particularly in some literary essays and when (as in parts of The Order of Things) he’s channeling Heidegger. But he’s overall much more accessible than, say, Deleuze orDerrida, especially from Discipline and Punish on. His last two books, on ancient sex, are quite lucid—he was ill and may have realised he didn’t have time to be obscure.
– Gary Gutting interviewed by Richard Marshall at 3:AM Magazine