A thought about the way that Michel Foucault writes:

Derrida says that he has to explain the "trace" the way he does because of the difficulties of the concept, its special character. This is a kind of claim familiar from Heidegger, and perhaps it goes back to Hegel. Last night, the continental in question was Foucault.

Denying that one of your ideas can be submitted to a redaction -- boiled down to an essence -- and instead hinting at the notion, restating it, revisiting it, like Derrida does with the idea of the "trace" and Foucault does with "power" . . . this is, among other things, an excellent way to deny people the ability to paraphrase your ideas in their own vocabulary. They have to use yours. In a sense, you haven't provided an "idea", and in fact you probably resist that kind of packaging. You've provided a set of statements which are cagey about their propensity for adding up, fusing together, into something other than the sum of their parts.

If twenty different places within your magnum opus have to be synthesized, compared, and corroborated, argued over, just for someone to get a bead on what one of your main ideas is, then you're making the work of hermeneutics a primary activity of your readership. That poring over, that comparing: it makes interpretation paramount, reinterpretation, argumentation. You're forcing a kind of Talmudic scriptural relationship between your text and its audience (and in Foucault's case, you're succeeding). How fares the chance for someone to dialogue with your text? It's reduced. People can dialogue with each other: your text is above the fray, unplumbable, enigmatic, handed down from the mountain. It only states; it doesn't know. And so it can never be wrong.