Driving both ways: Wilson & Sober's conflicting criteria for the identification of groups as vehicles of selection

We argue, first, that the notion of group selection discussed in Wilson & Sober's target article is much stronger than that in Sober (1984). Next, we show that the authors are tacitly using two independent criteria in applying this notion, a “differential fitness” criterion and a “shared fate” criterion. These criteria are flawed and cannot be applied jointly to the cases Wilson & Sober describe without creating inconsistencies.