I was reading the private debate between Owl and the other guy. You guys have a lot of interesting points but something I can't seem to wrap my head around is this:
Why do we need absolute certainty about the validity of our ideas to make educated guesses about which ones work and which ones don't?
Here's a poker metaphor that I like:
Anyone who plays poker seriously knows that the game is all about using incomplete information to form an educated guess about what cards our opponents hold. There are lots of different variables and conditions which change the way we use this information, and that is where the skill of the game comes in.
Say a poker opponent makes a big raise from an early seat, gets a couple of callers, continues to bet aggressively on all remaining betting rounds and finally completes his action that hand by going all in, risking all of his chips on that hand. Now suppose I have a mediocre hand and am attempting to discern whether or not I should call his bet.
But suppose for a moment that all ideas which do not have 100% certainty are inherently only 50% certain--no more or less likely than any other possible outcome.
Two competing voices in my head: One says, "He has a huge hand! Fold!" and the other says, "He's bluffing; you need to call."
Hmmmm, I'm not 100% certain of what cards this guy has. Every time I've seen him or anyone else make this particular sequence of plays under these conditions, he has almost always had a very strong hand.
Past evidence indicates that this sequence of conditions almost always leads to a certain outcome, but I can't be totally positive that his hand is strong, so I guess I will assume he's bluffing exactly 50% of the time, and therefore my decision is totally meaningless and fold/call are both equally likely to be the correct play.
If this line of thinking held up, it would be impossible for anyone to actually make money at poker in the long term, because all decisions about opponents' cards are inherently uncertain--but that doesn't prevent us from acting upon principles which past evidence indicates are most probably true! (Even if they're occasionally wrong.)
Now back to the poker hand. Of course, most of the time I would decide to go ahead and fold under these conditions. Now let's pretend that, after I and each other opponent folds, our strong bettor gleefully reveals his terrible hand and laughs at us as he rakes in the chips--he's made a successful bluff and fooled us all!
Hmm, that's interesting. I will have to take this new data into account the next time I am attempting to assess the probability of him having a strong hand--next time this situation comes up, I will be more likely to determine that he is bluffing. But what I will NOT do is make an assumption that, because my conclusion on this particular trial was incorrect, the method I used for reaching that conclusion was totally invalid. Yes, science is wrong sometimes. That doesn't make it useless for purposes of intuiting the probable outcomes of future trials.
But again, the entire system collapses if we don't allow that meaningful decisions can be made on incomplete information. I get rather bored with theists who would have us believe that all possibilities (like, say, whether or not the Christian God exists in the precise form that modern Christians imagine him) are equally valid until absolute proof is found--anyone who's played a game of poker knows that's bullshit.