If theory is not understood also as a practice, it will often make the mistake of confusing itself with knowledge, and will then cease to be theory. In other words, theory, unless it is understood as a set of ways of talking that prepares for purely theoretical engagement, becomes philosophy--and poor philosophy at that. Unless theory is understood sometimes as the effort needed in order to get to theory proper, then when we theorize, we are not doing theory--we're doing, instead, something like metaphysics.
This is the main reason why Paul de Man says that the resistance to theory is itself theory (in "The Resistance to Theory"). Indeed, his formula here also means that if you are opposing theory, the only way you can oppose it is by theorizing--but to lay too much stress on this misses his point. The point is that the purely theoretical, no longer muddied up by the effort to try and actually do theory, is not theory. The effort to become theoretical, or the effort needed to theorize, which itself comes from outside theory proper and from the domain of reading (in de Man's understanding of that word)--this effort that is an effect of some resistance to pure theory, is indeed theory.
One can see how right de Man is in looking at crappy theoretical work. People just start pronouncing upon things, working out their structure in accordance to certain conceptions justified by their favorite theoreticians. Theory has stopped working here, precisely because there is no problem of how, on a practical level, to go about theorizing. In other words, the real force of de Man's statement comes from seeing it as exclusionary, as a limit upon what can legitimately be called theory: theory that does not resist theorizing (or encounter the problem of how in practice to go about its business) is not theory. The resistance to theory is theory: "the language it speaks is self-resistance" ("The Resistance to Theory").