I linked to the page with this fake koan from computer science that I like, but I might as well steal it and post it on my own:
In the days when Sussman was a novice, Minsky once came to him as he sat hacking at the PDP-6.
“What are you doing?”, asked Minsky.
“I am training a randomly wired neural net to play Tic-Tac-Toe,” Sussman replied.
“Why is the net wired randomly?”, asked Minsky.
“I do not want it to have any preconceptions of how to play”, Sussman said.
Minsky shut his eyes.
“Why do you close your eyes?”, Sussman asked his teacher.
“So that the room will be empty.”
At that moment, Sussman was enlightened.
This is really the core of any epistemology. We have some preconception of how the world works–NO ONE is immune from it, and that’s a good thing. Often, good preconceptions allow us to see things that would not be obvious from data itself, but bad ones either keep us from seeing the patterns or see things that aren’t there. That means that we need to, first, train ourselves with the good preconceptions (what people call “domain knowledge”) and equally train ourselves to update them as necessary when the data insists on telling us otherwise. There is an inherent tension between the two, and successful pursuit of knowledge (and indeed, the whole point of “science”) is to strike an appropriate balance between the two. I am not going to disguise my distaste for the current obsession with “data” that seems to insist on forcing uninformed priors on us, but sometimes, “theory, like fog on eyeglass, obscures facts” too. (I stole this quote from Stephen Jay Gould, but he says it is from a Charlie Chan movie that I never watched).