Two Types of Empiricism
Type I empiricism says that all knowledge comes through the five senses. Type II empiricism says that all knowledge derives from experience. Neither entails the other. The senses could be the sole source of knowledge without being conduits of experience: the process might be entirely physical-causal, or proceed by way of blindsight and analogues thereof. And experience could be the source of all knowledge without the five human senses being involved: there might be no body at all, or different senses, or experiences of a type other than experience of sense. Type I empiricism focuses on specific human organs; type II empiricism focuses on the concept of experience as such. It is type II empiricism that we are dealing with when we say things like, “All ideas derive from impressions” or “Beliefs must have an experiential foundation”. Type I is usually a steppingstone to type II: for it is quickly pointed out that we need to include an inner sense in order to make room for self-knowledge (“ideas of reflection”). A natural response to type I empiricism is to ask what motivates such a view—why should we believe it? It sounds like a stipulation we can happily ignore: what is it about knowledge that requires it? A rationalist will simply scoff at it. But type II empiricism has a clear rationale: knowledge requires reasons, and experience looks like the only possible source of reasons (ultimately speaking). Knowledge can’t just come from nowhere; it has to be based on something: but what else could that be except experience—impressions, presentations, sensations, events of seeming? Things strike us a certain way and we form beliefs based on this striking: that’s how knowledge works, isn’t it? Otherwise it is all groundless piffle, mere posturing. So all knowledge must be based on experience—or else not be knowledge at all. This reflection puts two familiar types of epistemology in peril: religious and rationalist (and course these were in the empiricist’s sights). We are instructed to believe religious teachings because they appear in the Bible or are intoned by clergymen in fancy robes, but none of this affords the kind of individual subjective impression of truth that knowledge requires; it is just so much taking on trust (“faith”, “revelation”, “tradition”). But the empiricist is having none of it: he wants an actual human experience that provides a real reason for the belief in question. Similarly, the rationalist tells us that certain ideas are found woven into the innate fabric of the mind: but then, there are no experiential reasons for such items of alleged knowledge. We just have them, but we can’t say why. We don’t have them because things seem a certain way to us; we have them simply because we were born having them. Their rationality thus comes into question. The empiricist is thus unhappy with the “brute knowledge” assumption of the rationalist, i.e. knowledge without experiential reasons. But what can the empiricist say about such knowledge?
Ideally, he could say that mathematical knowledge has its own proprietary mode of seeming that provides the necessary experiential grounding—“I have the distinct and vivid impression that Pythagoras’s theorem is true” etc.—but the trouble is that such experiences don’t seem to exist. The only alternative, then, is to declare that the knowledge in question doesn’t really exist (it’s all tautologies or human conventions). The problem here is actually quite severe, because the type II empiricist has hold of a solid point, but this type of knowledge poses a serious threat to it. For we can’t just concede empiricism defeated by counterexample and move on to another theory: empiricism has to be true (in the sense intended), and yet it appears not to be true for this type of knowledge (also logic and ethics). This may prompt us to search for some other notion of experience capable of meeting the case—some sense in which we do have experiences of mathematical facts. Here is where philosophy reaches that state in which we just want to scream (“It was all going so smoothly and then this!”). A priori knowledge has always been difficult, but this makes it irritatingly difficult. Empiricism looked so good, so manifestly sensible, and yet it can’t apparently be made to fit the case of one large category of knowledge—but not in such a way that we can simply move to a superior theory that dispenses with its central tenet. And surely we don’t want to say that a prioriknowledge is mysteriously experience-based! What we need, evidently, is the idea of a priori experience, but that idea looks hard to make sense of. If we had it, though, we could say that all knowledge rests on experience, thus vindicating type II empiricism.
 Notice the form of the classic formulation: “A posteriori knowledge is knowledge based on experience; a prioriknowledge is knowledge not based on experience”. So what is it based on? We get the rather unhelpful, “Knowledge based on reason”. We can see how experience can provide a basis, but how does reason do anything comparable? What does it serve up that can play the reason-providing role of experience without being experience? The answer is obscure at best–hence the appeal of a generalized empiricism. Thus we have the problems besetting Western epistemology from Plato on. Epistemology shouldn’t have been this difficult!