If I understand your question, you may be referring to what is termed "apodictic" knowledge by
Husserl and others. This is knowledge which, according to phenomenologists (and others) is
immediate and absolutely certain. Examples that they give are things like: you see the red color on a
house and it is certain that you are experiencing red. Not unreasonable on the face of it, but the
problems come up when you start thinking about what it is, really, that you know apart from anything
else. That is, you can question your memoryof seeing red 30 seconds ago, perhaps you're
misremembering that it was red, or that you saw it at all, or even that it was called "red". And that
brings us to the next problem, viz., that whatever you are experiencing, even in the moment, in order
to be identified at all, must call on memory. To know that the color red is not the smell of roses means
we remember, at least, that we have had smells and roughly what they "feel" like. But can we rely on
that? Why? Dennett, in "Quining Qualia" has an essay bringing up problems with this, and Levin, in
Reason and Evidence in Husserl's Phenomenology(a difficult, but very good book) does even better.
The problem with Dennett, which he glosses over, is that his critique is not applied widely enough.
When it is, his own thesis (which basically says that consensual empirical evidence is not subject to
the same doubts) is also undermined.
Anyway, it's an interesting position, that we have such certain, immediate knowledge, and it seems
that it should be true, but it's extremely hard to pin down just where that happens.