It was already discussed earlier: men and women are distributions. Two distributions can be distinct even if there are overlaps between the distributions. For instance, men are much more likely to have beards than women are. This is a difference between men and women, and is not invalidated by the existence of women with beards.