There are things that are (almost) definitely knowable, and things that are (almost) definitely unknowable, and things that might be knowable.
Another category is things that are in a certain sense knowable in theory, but definitely unknowable in practice. For example:
- the heavy-tailedness of a probability distribution, given the constraint that you don't have much data to work with (some claim that the financial markets are this way)
- decisions that depend on scientific results that will not be available (with a high degree of certainty) until long after the decision must be taken (eg some decisions about what to eat; nutrition science is still evolving)
- NP-complete computations (assuming P != NP)
Related: 'unknown unknowns' (as opposed to known unknowns)