I anticipate expressing free-floating beliefs would get me negative karma on Less Wrong.
More seriously:
I do not anticipate free-floating beliefs being useful in the same sense that maps of reality are useful. A map can turn out to be accurate or inaccurate, and insofar as it is accurate it can help me navigate and manipulate reality. My belief that “a proper belief should not be free-floating” prohibits free-floating beliefs from doing any of that.
Or one might as well see it as not a belief, but as a definition. There’s BeliefType1 which is grounded in reality, and BeliefType2 which is not, and we happen to call BeliefType1 a “proper belief”. (Of course we still do it for a reason, because we care about our sheep, or rather, we care about our beliefs being true and thus useful.)
The ability to anticipate experiences is one of our maximands because we have goals that are optimally achieved with this ability. To believe that beliefs should allow us to anticipate experiences is grounded in the desire to achieve our goals.
One way of answering might be to say that there is no separate “belief” that beliefs should be grounded. But i’m not sure.
All I know is that the question annoys me, but I can’t quite put my finger on it. It reminds me of questions like (1) the accusation that you can’t justify the use of logic logically, or (2) the accusation that tolerance is actually intolerant—because it’s intolerant of intolerance. There might be a level distinction that needs to be made here, as in (2) - and maybe in (1) though I think that’s different.
(1) has come out of my mouth on a few occasions, albeit not in those exact words. It’s normally after a few beers and I feel like playing the extreme skeptic a la David Hume, just to annoy everyone. I think the best way around it is to resort to the empirical argument and say that, in our experience, it is always right: Essentially the same thing Yudkowsky does with PA arithmetic here. Trying to find an argument against it which is truly “rationalist” in the continental sense has been a dead end in my experience.
(2) sort of depends on the pragmatics and what “tolerance” actually means to the persons involved in a given context. If you define tolerance as simply being tolerant of other viewpoints, then you can still be tolerant of the intolerant viewpoints. However, if you define it as freedom from bigotry, then that could indeed be called “intolerant” by the standards of the first definition.
In what way is the belief that beliefs should be grounded not a free-floating belief itself?
I anticipate expressing free-floating beliefs would get me negative karma on Less Wrong.
More seriously:
I do not anticipate free-floating beliefs being useful in the same sense that maps of reality are useful. A map can turn out to be accurate or inaccurate, and insofar as it is accurate it can help me navigate and manipulate reality. My belief that “a proper belief should not be free-floating” prohibits free-floating beliefs from doing any of that.
Or one might as well see it as not a belief, but as a definition. There’s BeliefType1 which is grounded in reality, and BeliefType2 which is not, and we happen to call BeliefType1 a “proper belief”. (Of course we still do it for a reason, because we care about our sheep, or rather, we care about our beliefs being true and thus useful.)
Not sure which approach makes more sense.
The ability to anticipate experiences is one of our maximands because we have goals that are optimally achieved with this ability. To believe that beliefs should allow us to anticipate experiences is grounded in the desire to achieve our goals.
One way of answering might be to say that there is no separate “belief” that beliefs should be grounded. But i’m not sure.
All I know is that the question annoys me, but I can’t quite put my finger on it. It reminds me of questions like (1) the accusation that you can’t justify the use of logic logically, or (2) the accusation that tolerance is actually intolerant—because it’s intolerant of intolerance. There might be a level distinction that needs to be made here, as in (2) - and maybe in (1) though I think that’s different.
(1) has come out of my mouth on a few occasions, albeit not in those exact words. It’s normally after a few beers and I feel like playing the extreme skeptic a la David Hume, just to annoy everyone. I think the best way around it is to resort to the empirical argument and say that, in our experience, it is always right: Essentially the same thing Yudkowsky does with PA arithmetic here. Trying to find an argument against it which is truly “rationalist” in the continental sense has been a dead end in my experience.
(2) sort of depends on the pragmatics and what “tolerance” actually means to the persons involved in a given context. If you define tolerance as simply being tolerant of other viewpoints, then you can still be tolerant of the intolerant viewpoints. However, if you define it as freedom from bigotry, then that could indeed be called “intolerant” by the standards of the first definition.
I hope I’m making sense here.