On second thought, that's not right. But you probably understood what I mean. If you happen to make an a conjecture about something like Kolmogorov complexity or the halting problem, and it just happens to be undecidable, it's still either true or false.