4 ms·

Doesn't anyone have beliefs that are not modeled best by booleans? I intuitively like to use trivalent logic "true/false/not known" but the Bayesian reals from

Doesn't anyone have beliefs that are not modeled best by booleans? I intuitively like to use trivalent logic "true/false/not known" but the Bayesian reals from 0 to 1 is also pretty useful. Does one not continually adjust the weights assigned to all manner of propositions? For that matter, doesn't one adjust the meaning assigned to vague propositions as the specifics of the case become clearer? ("I thought DRY meant have your code look like gzip'ed output, but now I see it means after spotting a pattern, take a moment to think about it and see if there's not some nice coherent abstraction you are brushing against and could profitably use?")

mistermann 3y ago
> I intuitively like to use trivalent logic "true/false/not known" but the Bayesian reals from 0 to 1 is also pretty useful. In my experience less than 10% of people think in non-binary, non-heuristic-based forms on these and similar topics, and even Rationalists fare not much better.
- lanstin 3y ago
  I think it is a more productive tweak to thinking to understand beliefs as not being fully true or fully false than to believe that one can be purely rational in one’s actions or beliefs.
  - mistermann 3y ago
    I mostly agree, although the meme-based beliefs about the simplistic notion that it is ~impossible to be "purely rational in one’s actions or beliefs" is highly inaccurate imho. I don't disagree that people can't currently do it, but then consider how much we've improved when it comes to things like racism. Maybe if we actually tried to be rational (rather than patting ourselves on the back as is a popular practice here and in Rationalist and Scientific communities) we could make similar progress.
    - lanstin 3y ago
      I find it a category error to ask reason for a why do one thing over another. Our animal nature gives us motivation which we use reason to help. Not that it isn’t usually a benefit to fulfilling our motivations to have our reason finely honed.
      - mistermann 3y ago
        > I find it a category error to ask reason for a why do one thing over another. Our animal nature gives us motivation which we use reason to help. Of course, but your animal nature causes you to guess, no? Try overriding that instinct and see if you can come up with something else...

mistermann 3y ago

> I intuitively like to use trivalent logic "true/false/not known" but the Bayesian reals from 0 to 1 is also pretty useful. In my experience less than 10% of people think in non-binary, non-heuristic-based forms on these and similar topics, and even Rationalists fare not much better.

lanstin 3y ago
I think it is a more productive tweak to thinking to understand beliefs as not being fully true or fully false than to believe that one can be purely rational in one’s actions or beliefs.
- mistermann 3y ago
  I mostly agree, although the meme-based beliefs about the simplistic notion that it is ~impossible to be "purely rational in one’s actions or beliefs" is highly inaccurate imho. I don't disagree that people can't currently do it, but then consider how much we've improved when it comes to things like racism. Maybe if we actually tried to be rational (rather than patting ourselves on the back as is a popular practice here and in Rationalist and Scientific communities) we could make similar progress.
  - lanstin 3y ago
    I find it a category error to ask reason for a why do one thing over another. Our animal nature gives us motivation which we use reason to help. Not that it isn’t usually a benefit to fulfilling our motivations to have our reason finely honed.
    - mistermann 3y ago
      > I find it a category error to ask reason for a why do one thing over another. Our animal nature gives us motivation which we use reason to help. Of course, but your animal nature causes you to guess, no? Try overriding that instinct and see if you can come up with something else...