We need a better standard of evidence for harmful language than someone's claim at offense. I have no doubt that you could find an actual victim/target/member

We need a better standard of evidence for harmful language than someone's claim at offense. I have no doubt that you could find an actual victim/target/member of a marginalized group to claim to be bothered by any of these phrases in the first person. But I'm also certain that members of such groups enjoy telling other people what to do just as much (on average) as all of humankind at large, which is to say a lot, and so a better standard of evidence is needed. I've struggled to envision what such evidence would look like, but I think it will have to be someone who is better at devising experiments than me.