Some Problems with the Anti-Luminosity-Argument

I argue that no successful version of Williamson's anti-luminosity-argument has yet been presented, even if Srinivasan's further elaboration and defence is taken into account. There is a version invoking a coarse-grained safety condition and one invoking a fine-grained safety condition. A crucial step in the former version implicitly relies on the false premise that sufficient similarity is transitive. I show that some natural attempts to resolve this issue fail. Similar problems arise for the fine-grained version. Moreover, I argue that Srinivasan's defence of the more contentious fine-grained safety condition is also unsuccessful, again for similar reasons.