At least with science, adaptive knowledge is self-consistent within the context of evolution and science. Tautological and limited? Perhaps, but it is at least an internally consistent truth system. Within the system, adaptive knowledge will tend to be closer to the truth than non-adaptive knowledge.
Philosophical naturalists assume, of course, that there actually is an absolute noumenon or truth, and that science is the path to it - which IMO, is difficult, if not impossible, to defend.
They also assume that our knowledge is always becoming progressively more accurate, approaching the noumenon in a linear or asymptotic way - but there is no way of knowing that. We could simply be approaching a local maxima in our tiny local truth system, without realizing that greater and possibly unrelated maximas exist elsewhere in what we call reality.
~Tom S. (seventh comment here)
Based on my readings in Pragmatism, we accept that science produces knowledge because doing so seems to "work." (see this xkcd) However, suppose it didn't work. We would then try something else, and that itself would be invoking the scientific method. I definitely need to take a philosophy of science class. At the very least, I'll know what things to be confused about then.