> Lots of people seem to be incapable of thinking in term of distributions, which makes it really hard to have any discussion. Most people seem very capable of

> Lots of people seem to be incapable of thinking in term of distributions, which makes it really hard to have any discussion. Most people seem very capable of thinking statistically, but they don't communicate in distributions. For two people communicating in good faith, the two statements (1) "Boys are taller than girls" and (2) "Anne is taller than Sean" don't prompt talk about how both can't be true at the same time, sexist. People communicate statistical differences as "absolutes", but when communicating in good faith, everybody knows what everybody means. Look at stereotype research. Most stereotypes are statistically true. People communicate them "as if" absolute, but nobody's head explodes when individual examples don't "conform" to the stereotype. That's because people can think in distributions. That's not surprising, it's how to survive in a stochastic environment. This gap between thought and speech only becomes problematic ("makes it really hard to have any discussion") because we can rhetorically exploit the fact that most people can't explicitly communicate about distributions. Just tell them they're bigoted, and they don't know how to refute you. Yay, you win.