Until recently, there were two basic camps on "how we know stuff." The Rationalists and the Empiricists.
The Rationalists thought that knowledge was came from mental abstractions, not from reality. For example, Plato believed that we could all access the "ideal" world of "forms" by just introspecting. The medieval scholastic philosophers tended to be very rationalistic - they made arguments about abstractions not connected to real experience, like how many angels could fit on the head of a pin. Decartes was a rationalist - "I think, therefore I am" is rooted in the idea that we look inward to perceive reality.
The Empiricists, which are somewhat more modern, were a reaction to Rationalism, and rejected all of that. They said that all we can really "know" or "trust" is sensory experience. We can't build up complex mental models, and we can't have principles - we just have to be pragmatic all the time and do what seems best, because we can't really "know."
Kant tried to improve on both of these approaches by saying that we can't really know anything about "true" reality, because reality has to be filtered through the senses. There is only subjective "reality."
Ayn Rand, of whom I am a big fan, went in the other direction from Kant, but also rejected Rationalism and Empiricism. She said that you can build complex models of reality, but they have to be based on actual observations of reality (i.e., sense data). To do this, she proposed a new theory of concept formation (i.e., how to form abstractions based on observation of reality). Her approach is very Aristotelian, which is interesting because Aristotle was the main guy who disagreed with Plato at the very beginning. I'd highly recommend "Introduction to Objectivist Epistemology" if you're interested in this approach (although you may be better served by starting at a less advanced level... still, the book is short and very accessible, albeit kind of a mind trip).
