“Rational” is a systematic process for arriving at true beliefs (or high-scoring probability distributions), so if you want true beliefs, you’ll think in the ways you think are “rational”. But even in the very best case, you’ll hit on 10% probabilities one time out of ten.
I didn’t see anything wrong with your original comment, though; it’s possible that Ciphergoth is trying to correct a mistake that isn’t there.
Well, if you got a very improbable result from a body of data; I could see this happening. For example, if most of a group given a medication improved significantly over the control group, but the sample size wasn’t large enough and the improvement was actually coincidence, then it would be rational to believe that it’s an effective medication… but it wouldn’t be true.
Then again, we should only have as much confidence in our proposition as there is evidence for it, so we’d include a whatever-percent possibility of coincidence. I didn’t see anything wrong with your original comment, either.
I’ve since learned that some people use the word “rationality” to mean “skills we use to win arguments and convince people to take our point of view to be true”, as opposed to the definition which I’ve come to expect on this site (currently, on an overly poetic whim, I’d summarize it as “a meta-recursively applied, optimized, truth-finding and decision making process”—actual definition here).
Now I am really confused. How can a belief be rational, and not true?
“Rational” is a systematic process for arriving at true beliefs (or high-scoring probability distributions), so if you want true beliefs, you’ll think in the ways you think are “rational”. But even in the very best case, you’ll hit on 10% probabilities one time out of ten.
I didn’t see anything wrong with your original comment, though; it’s possible that Ciphergoth is trying to correct a mistake that isn’t there.
Well, if you got a very improbable result from a body of data; I could see this happening. For example, if most of a group given a medication improved significantly over the control group, but the sample size wasn’t large enough and the improvement was actually coincidence, then it would be rational to believe that it’s an effective medication… but it wouldn’t be true.
Then again, we should only have as much confidence in our proposition as there is evidence for it, so we’d include a whatever-percent possibility of coincidence. I didn’t see anything wrong with your original comment, either.
I’ve since learned that some people use the word “rationality” to mean “skills we use to win arguments and convince people to take our point of view to be true”, as opposed to the definition which I’ve come to expect on this site (currently, on an overly poetic whim, I’d summarize it as “a meta-recursively applied, optimized, truth-finding and decision making process”—actual definition here).