Computationalism, the thesis that mental states are computational states realized by physical systems such as (but not limited to) central nervous systems, drives the research in the cognitive neurosciences and artificial intelligence. But Hilary Putnam and John Searle, have independently argued for the triviality of computationalism with what I dub the Putnam-Seorle Thesis (PST), the claim that up to complexity limitations, every ordinary physical system implements every computation. The consequence is that unless the computationalist is prepared to accept a radical form of pan-psychism (violating obvious pre-theoretical constraints), computationalism is doomed.

This dissertation examines the ways computationalism might avoid this unwelcome consequence of mental realizations run rampant. I find that computationalism does not need to be abandoned for that end. Instead, we must re-evaluate the conception of program implementation. Given a robust notion of implementation, computationalism avoids the alleged consequence of being too liberal.

After presenting the case for PST, I focus on the work of Ronald Chrisley, David Chalmers, Jack Copeland, Matthias Scheutz, and Paul Teller. Much of the effort here removes limitations found in the initial case for PST. The rationale is that instead of dismissing PST for prima facie reasons, a more profitable approach is to make the case for PST as strong as possible, thus subjecting computationalism to the most rigorous version of their objection. To this end the case for PST is strengthened so as to meet warranted constraints regarding causality, disposition, and complexity.

I find that the case for PST ultimately backfires, entailing not only the trivialization of computationalism but the notion of program implementation for any and all systems. Rejecting computationalism will avoid only some of the absurdities engendered. I argue that rejecting the conception of program implementation under which the case for PST is made is the more reasonable course, and I propose that a more robust conception of program implementation will align computation theory with empirical work. This argument provides a schematic for other problems of a similar nature having to do with isomorphic relations or other relations that are alleged to hold with respect to some description or interpretation.