@prothero,
As I see it, the situation is as follows:
1. We only know something if our belief in it is both
justified and
true.
2. We can never be certain that any empirical belief is true; hence we can never be certain that we know something.
3. For a realist, there are truths independent of our minds. If any of these truths actually accord with our beliefs, then the beliefs in question (if justified) constitute
knowledge. This is so by definition: truth + justification = knowledge. Certainty, or lack of it, is a separate issue. (Similarly, I may not be certain that X is a male sibling, but if he
is, then he is definitely a brother, because male + sibling = brother.)
4. So we can never have any
guarantee of knowledge, but we can still (as a matter of fact, from an outside perspective) actually
have knowledge.
5. Arguably, although we cannot be
certain, we can in some cases 'know that we know'
beyond reasonable doubt.