OK, I think I agree with you: We don't know if ASD or any of our current diagnoses are a particularly good model, but it is possible to cluster people and some people are farther from whatever group of clusters collectively contains the majority or strong plurality.