Is "a priori knowledge" really knowledge? If I know the language, I "know" ipso facto everything that is necessarily true when stated in that language, some self-referential Godel stuff aside, anyway. Thus, isn't each bit of a priori knowledge just one more example in the dictionary? ("Colored" means having color. Green is a color. So, for example, a thing that is green is colored.") We could then wander off into closing the gap between "is" and "has" and the past participle, but it's all just language.

But, of course, I don't know everything that is true in the language. That's what proofs are for. But that raises the question of whether a really good geometer just "speaks Euclid" better than the average fellow. It's been a loooooong time since I read Quine, but I associate with him an example of expanding "Time is money" to "If time is money, then time is money.," and then adding "If" at one end and "then time is money" at the other to create statements that are necessarily true but which, at some point stop being comprehensible to the reader, at which point one might say the reader does not "know" the fact represented by the statement. We can imagine a square and know that it is a rectangle, but we cannot imagine the infinitely nested "if, then" expression. Can we say that we "know" that it's true, or, if we do, are we using the word in precisely the same way when we do?