I agree that with current cultural training data, training data collected from a lifetime of human experiences in a world where phrasing and behaviors enact racialization based on things like people’s bodily form, this method of verbal capability judgement is going to tend to have a correlation with people’s racialized attributes, because the cultural aesthetic you’re trying to match on by detecting verbal intelligence will present differently in people from different dialects and backgrounds.
different groups have different word bindings for the word racism based on whether it is an agentic action (“I’m not racist, dude, come on!”) or being synchronized with the correlated patterns of behavior that result in people being excluded (ie, the “I’m not racist, come on!” is probably being racist by this standard). IMO, because racism is such a high magnitude word for some people, there are going to be those who have the inclination to downvote because of your use of it.
but I agree that there’s a correlation that is worth highlighting: people’s verbal and aesthetic presentation does not represent intelligence the same way across different cultural aesthetic backgrounds, and so attempting to guess whether someone is competent from their presentation will tend to have a correlation with racialized groups—ie, people who others treat as being of a category because of their simple physical characteristics rather than their complex and agentic ones.
There’s, of course, some issue as well with actual intelligence probably being at least partially causally downstream from having a poor household in early childhood, and poor households being correlated with racialization—people tend to have less aggregate capability, including iq, when their nutritional, social, and chemical environment was bad growing up (eg poorer households are more likely to have undetected mold and chemical hazards hidden in the floor and walls). But I don’t think that was the key thing you were getting at.
I agree that with current cultural training data, training data collected from a lifetime of human experiences in a world where phrasing and behaviors enact racialization based on things like people’s bodily form, this method of verbal capability judgement is going to tend to have a correlation with people’s racialized attributes, because the cultural aesthetic you’re trying to match on by detecting verbal intelligence will present differently in people from different dialects and backgrounds.
different groups have different word bindings for the word racism based on whether it is an agentic action (“I’m not racist, dude, come on!”) or being synchronized with the correlated patterns of behavior that result in people being excluded (ie, the “I’m not racist, come on!” is probably being racist by this standard). IMO, because racism is such a high magnitude word for some people, there are going to be those who have the inclination to downvote because of your use of it.
but I agree that there’s a correlation that is worth highlighting: people’s verbal and aesthetic presentation does not represent intelligence the same way across different cultural aesthetic backgrounds, and so attempting to guess whether someone is competent from their presentation will tend to have a correlation with racialized groups—ie, people who others treat as being of a category because of their simple physical characteristics rather than their complex and agentic ones.
There’s, of course, some issue as well with actual intelligence probably being at least partially causally downstream from having a poor household in early childhood, and poor households being correlated with racialization—people tend to have less aggregate capability, including iq, when their nutritional, social, and chemical environment was bad growing up (eg poorer households are more likely to have undetected mold and chemical hazards hidden in the floor and walls). But I don’t think that was the key thing you were getting at.