Ideal reasoners don’t believe in zombies

The negative zombie argument states that p&~q is ideally negatively conceivable and, therefore, possible, what would entail that physicalism is false (Chalmers, 2002, 2010}. In the argument, p is the conjunction of the fundamental physical truths and laws and $q$ is a phenomenal truth. A sentence phi is ideally negatively conceivable iff phi cannot be ruled out a priori on ideal rational reflection. In this paper, I argue that if its premises are true, the negative zombie argument is neither conclusive (valid) nor a priori. First, I argue that the argument is sound iff there exists a finite ideal reasoner R for a logic x with the relevant properties which believes <>(p&~q) on the basis of not believing p->q on a priori basis. A finite reasoner is a reasoner with finite memory and finite computational power.  I argue that if x has the relevant properties and R is finite, then x must be nonmonotonic and R may only approach ideallity at the limit of a reasoning sequence. This would render the argument nonconclusive. Finally, I argue that, for some q, R does not believe <>(p&~q) on the basis of not believing p->q on a priori basis. For example, for q=`someone is conscious'. I conclude that the negative zombie argument (and, maybe, all zombie arguments) is neither conclusive nor a priori (the choice of q relies on empirical information).