carado.moe

tammy's blog about
AI alignment, utopia, anthropics, and more;

also on twitter, lesswrong, rss, and at

refusing to answer ≠ giving a negative answer

a lot of people like to, when faced with a hard-to-determine question such as "is a chair happy" or "does a dog have buddha-nature", with the negative.

i think this is very fundamentally erroneous. the correct answer is that the question itself is erroneous; to answer it would be to erroneously accept the question, like in the classic case of "have you stopped beating your wife?".

the answer to "is a chair happy?" is not "no"; that would imply that the question of whether it does or not makes sense, which it doesn't. the answer is "the question is erroneous", or "mu".

this comes up pretty often in philosophical discussions i have with friends. one reason why this is important is that i like statements to be meaninful, and it's very useful for stating that something does not have a characteristics, to mean something different than not stating anything. this also kind of ties into bayesianism: knowledge of a negative is not absence of knowledge, or updating your belief towards negation wouldn't do anything.