Hard but so valuable to define hard criteria for empirical theories of consciousness.

I congratulate Doerig, Schurger, and Herzog for their stimulating contribution for an empirical approach to theories of consciousness. I must also admit that the excellent way Global Neuronal Workspace Theory (GNWT) I've contributed to since 2001 passes the test they designed may contribute to my enthusiasm (see Table 1 of their article). In this Comment, I focus on two points: highlighting a potential epistemological weakness of their approach, and proposing how GNWT can solve the 'small network argument.'