Random and sub-conscious

Sometimes a random chance conversation can play around in your head till you are sub-consciously wondering about the crux of it, until some clarity emerges from a muddle of thoughts.
I had such a conversation recently, on asking a question. It freaked me out a little, because in a way I have chosen to do this asking questions business for a living. Here goes.

Research in most fields is fuelled by existing paradigms, to a large extent. There are large bodies of work that support the ‘famous’ theory. But once such theories are dethroned (for want of a better word), there is an equally impressive body of work against it. Stuff happens in a cyclical fashion. Just as the wheel of time goes around, so does the wheel of theories, it seems.
What freaked me was: as a young novice wanting to start somewhere, in what context do I place existing theories? Using them to frame the question you want to answer has an obvious bias. Take sexual selection for instance. Asking questions assuming that there is sexual selection in the system studied may color perception; you may end up reading too much into the behavior of that animal.

So how about throwing the theory out of the window and looking at data for its own sake? But I am not ok with that as well. I like a well thought out question, which not only has an expected outcome but also explicit alternate scenarios thought out a priori. A central unifying theme with little radiating questions.
I guess the way out is to base questions on existing knowledge and literature on the field of your choice, and then trying to make predictions for your study system. To go out and make these bold hypotheses based on literature reviews and reasoning that sounds logical to a few minds at least! Does not matter if you don’t end up seeing what you expected. Actually, it is even more interesting when expected patterns don’t pop up!