ABSTRACT. Zombies make trouble for physicalism. Intuitively, they seem conceivable, and many take this to support their metaphysical possibility -a result that, most agree, would refute physicalism. John Hawthorne (2002) [Philosophical Studies 109, 17-52] and David Braddon-Mitchell (2003) [The Journal of Philosophy 100, have developed a novel response to this argument: phenomenal concepts have a conditional structure -they refer to non-physical states if such states exist and otherwise to physical states -and this explains the zombie intuition. I argue that this strategy fails. The considerations Hawthorne and Braddon-Mitchell adduce in support of their analysis in fact do no such thing. Further, their main argument for the analysis is self-defeating: exactly similar reasoning would undermine the view it is meant to establish. Finally, on closer inspection the conditional analysis is incompatible with the zombie intuition. Thus, not only is the analysis incapable of explaining the intuition: the intuition's plausibility indicates that the analysis is incorrect. I also suggest that the allure of the conditional-analysis strategy may derive from a questionable view about what explaining the intuition would require.Zombies make trouble for physicalism. They are defined as creatures that lack consciousness but are physically and functionally identical to conscious human beings: there is nothing it is like to be a zombie.1 Intuitively, such creatures are conceivable, and their conceivability is often taken to support the claim that they are metaphysically possible. But if they are metaphysically possible, then -most agree -physicalism is false.