(...)
Do you see the problem in the above scenario?
Yes. However, it’s fairly evident that the cost of a general policy against uttering potentially-damaging sentences for all cases is prohibitive; you pretty much by rule can’t exist if you’re absolutely not allowed to ever accidentally trigger a negative reaction. That’s hyperbole, though, but the key point is:
It’s all about quantities and how much whiter or blacker.
A super-general ultra-powerful general policy against possibly-patriarchal words or behaviors by adjusting for expected perceptions would not only carry a much higher cost-per-avoided-damage than other behaviors, but would also in most cases dramatically increase the risk of committing errors in judgment, on top of pretty much becoming your sole lifegoal and preventing you from ever doing any progress in any other topic of interest.
What’s more, the very policy is itself sexist, because it expects differences in minds and perceptions between genders. This expectation is not secret, so it would obviously have an effect on others’ behaviors.
Historically / experimentally, what have we seen happen when others know that X is expected of them?
They will either X, or -X in order to signal.
By placing Expectation X, you are focusing the behavior-space on X in particular, and discouraging escape from that zone. This is why I find the idea of behaving differently and having different expectations of different genders to be harmful. By placing my expectations and own behaviors in a wider X, and having the same X for all audiences, I believe myself to be actively contributing to the reduction of unfair discrimination.
Well, liberal feminists that I disagree with would point out that the sentence “You certainly seem to have sex with lots of people” provokes very different reactions when you say it to men than when you say it to women. If you disbelieve me, try it out on your Facebook friends.
I very much believe you. I have also seen experimental evidence of pretty much the same. I don’t really see how that is directly relevant, or what your underlying point is, though. A policy of identical behaviors and identical reaction expectations will, on pain of deliberately choosing to lower its expected total utility, take into account known evidence for realistically-expected behavior and expect its conjugate¹, so as to causally bring behaviors closer to the ideal expected identical behaviors.
Perhaps this is already what you are doing, perhaps this decision theory approach is wrong (I try to use TDT-like processes because, well, so far they work), or perhaps I’m confusing things or putting too many concepts together. The above does makes sense to me—the mathlike stuff works out and the anecdotal evidence supports—and is not a conclusion I arrived to trivially by Authority or some form of subculture programming, barring denial-of-denial problems.
Other than this, Bugmaster has made some interesting queries, which I’m also interested in.
(¹. I mean here in the algebra sense of “conjugates”, figuratively, as the behavior which is expected to multiply or add up with the current expected real behavior such that the resulting behavior after future corrections by the other party becomes 1, AKA gender-identical and otherwise calibrated as much as possible for utility-increasing future behavior. )
ETA: I’ve elaborated a bit more on general policies in more technical terms and with more standard LW jargon in this other comment.
Yes. However, it’s fairly evident that the cost of a general policy against uttering potentially-damaging sentences for all cases is prohibitive; you pretty much by rule can’t exist if you’re absolutely not allowed to ever accidentally trigger a negative reaction. That’s hyperbole, though, but the key point is:
It’s all about quantities and how much whiter or blacker.
A super-general ultra-powerful general policy against possibly-patriarchal words or behaviors by adjusting for expected perceptions would not only carry a much higher cost-per-avoided-damage than other behaviors, but would also in most cases dramatically increase the risk of committing errors in judgment, on top of pretty much becoming your sole lifegoal and preventing you from ever doing any progress in any other topic of interest.
What’s more, the very policy is itself sexist, because it expects differences in minds and perceptions between genders. This expectation is not secret, so it would obviously have an effect on others’ behaviors.
Historically / experimentally, what have we seen happen when others know that X is expected of them?
They will either X, or -X in order to signal.
By placing Expectation X, you are focusing the behavior-space on X in particular, and discouraging escape from that zone. This is why I find the idea of behaving differently and having different expectations of different genders to be harmful. By placing my expectations and own behaviors in a wider X, and having the same X for all audiences, I believe myself to be actively contributing to the reduction of unfair discrimination.
I very much believe you. I have also seen experimental evidence of pretty much the same. I don’t really see how that is directly relevant, or what your underlying point is, though. A policy of identical behaviors and identical reaction expectations will, on pain of deliberately choosing to lower its expected total utility, take into account known evidence for realistically-expected behavior and expect its conjugate¹, so as to causally bring behaviors closer to the ideal expected identical behaviors.
Perhaps this is already what you are doing, perhaps this decision theory approach is wrong (I try to use TDT-like processes because, well, so far they work), or perhaps I’m confusing things or putting too many concepts together. The above does makes sense to me—the mathlike stuff works out and the anecdotal evidence supports—and is not a conclusion I arrived to trivially by Authority or some form of subculture programming, barring denial-of-denial problems.
Other than this, Bugmaster has made some interesting queries, which I’m also interested in.
(¹. I mean here in the algebra sense of “conjugates”, figuratively, as the behavior which is expected to multiply or add up with the current expected real behavior such that the resulting behavior after future corrections by the other party becomes 1, AKA gender-identical and otherwise calibrated as much as possible for utility-increasing future behavior. )
ETA: I’ve elaborated a bit more on general policies in more technical terms and with more standard LW jargon in this other comment.