For one of the first times in the history of science, we have a large class of unanswered scientific questions for which we have hypothesized answers that are probably correct, but simply don't know which of these answers (if any) are the correct ones for particular problems.
There are thousands of academic papers (easily mass produced with the rubric found in today's edition of PhD comics), for example, that reconcile gravity and the other fundamental forces of physics, and explain phenomena called dark matter and dark energy. Some combination of them is probably either correct, or very nearly correct. But, we lack the means to discern which of those theories is the correct one.
We have basically two hypothesises for late human evolution ("Out of Africa" and "Multi-Regional") with fairly subtle variations within each. But, an ultimate answer may never be known with complete certainty.
Of course, unlike nice neat riddles that are somewhat similar in form, like Sudoku puzzles, we can't be sure for any particular question that the right answer is out there somewhere.
One supposes that this state of affairs has something to do with the fact that a large share of the all the scientists in all of human history are alive today, while scientific experiements seem to face a law of diminishing returns. As science progresses, we need more expensive experiements to give us smaller chunks of new data (measured in explanatory power, rather than terabytes, for exampe). The last part per million of accuracy (which may in turn invalidate whole classes of theories), can cost more to establish than all the experiements that brought science to that point.