ITTs and steelmanning feel like they serve different (though overlapping) purposes to me. For example, if I am talking with people who are not X (libertarians, socialists, transhumanists, car-owners...), we can try to steelman an argument in favor of X together. But we can’t do an ITT of X, since that would require us to talk to someone who is X.
Yes, though I assume the best test for whether you really steelman someone would be if you can take a break and ask her whether your representation fits.
I don’t think an ITT is a good test of a steelman. Often you’re coming from a different frame from someone else, and strong arguments to you will be framed completely differently from strong arguments for someone else.
I would like to see that people who write articles about what the supposed actions or motivations of other people—or government agencies, firms, or whatever—are to actually try to present their actions and motivations in a way that at least assumes that they are not completely dumb or evil or pathetic. It seems to be fashionable that when people do not see the sense behind actions, they do not try hard but jump to the conclusion that it must be due to some despicable, stupid, or at least equilibrium-inefficient behavior (e.g. some claims about “signalling”; no proper analysis whether the claim makes sense in a given situation required). This may feel very insightful; after all, the writer seemingly has a deeper insight into social structures than the social agents. But supposed insights that feel too good can be dangerous. And that a model is plausible does not mean that it applies to every situation.
I believe that steelmanning has mostly been deprecated and replaced with ideological turing tests.
ITTs and steelmanning feel like they serve different (though overlapping) purposes to me. For example, if I am talking with people who are not X (libertarians, socialists, transhumanists, car-owners...), we can try to steelman an argument in favor of X together. But we can’t do an ITT of X, since that would require us to talk to someone who is X.
Yes, though I assume the best test for whether you really steelman someone would be if you can take a break and ask her whether your representation fits.
I don’t think an ITT is a good test of a steelman. Often you’re coming from a different frame from someone else, and strong arguments to you will be framed completely differently from strong arguments for someone else.
Yes maybe an ITT tests a fleshman instead of a steelman or a strawman...
What I mean is:
I would like to see that people who write articles about what the supposed actions or motivations of other people—or government agencies, firms, or whatever—are to actually try to present their actions and motivations in a way that at least assumes that they are not completely dumb or evil or pathetic. It seems to be fashionable that when people do not see the sense behind actions, they do not try hard but jump to the conclusion that it must be due to some despicable, stupid, or at least equilibrium-inefficient behavior (e.g. some claims about “signalling”; no proper analysis whether the claim makes sense in a given situation required). This may feel very insightful; after all, the writer seemingly has a deeper insight into social structures than the social agents. But supposed insights that feel too good can be dangerous. And that a model is plausible does not mean that it applies to every situation.