Conceivability Arguments for Dualism, Part II: Stoljar's Critique of David Chalmers' Argument

Notes on Daniel Stoljar's “The Conceivability Argument and Two Conceptions of the Physical” (Phil. Perspectives 15, 2001)

In this paper, Stoljar attacks Chalmers’ conceivability argument for property dualism. Recall that the conceivability argument runs as follows:

1. It is conceivable that I have a zombie-twin, i.e., someone physically identical to me and yet who lacks phenomenal consciousness.
2. If it is conceivable that I have a zombie-twin, then it is possible that I have a zombie-twin.
3. If it is possible that I have a zombie-twin, then physicalism is false.
4. Therefore, physicalism is false.

In addition to his t-physical/o-physical property distinction, Stoljar brings in Van Cleve’s weak/strong conceivability distinction. Weak conceivability is the inability to conceive that P is impossible, and strong conceivability is the ability to conceive that P is possible. Among other problems, the former allows that Goldbach’s Conjecture both is and isn’t conceivable, in which case it is not a reliable guide to possibility. Goldbach’s Conjecture isn’t strongly conceivable, however, and therefore (in conjunction with other virtues) strong conceivability provides prima facie justification for modal propositions.

Stoljar brings these two distinctions to bear upon Chalmers’ conceivability argument by using them to construct a dilemma against it: Start off with the assumption that “conceivable” is intended to be read as “strongly conceivable” throughout Chalmers’ argument. Then:

1. Either “the physical facts” are to be read throughout the argument as “the t-physical facts” or as “the o-physical facts”.
2. If they are to be read as “the t-physical facts”, then although the argument is valid, we have no reason to think it is sound. For the extra, o-physical facts, when combined with the t-physical facts, may then metaphysically necessitate that the twin is not a zombie, in which case (3) would be false.
3. If they are to be read as “the o-physical facts”, then the argument is invalid. For then ‘conceivable’ in (1) can only be truly read as ‘weakly conceivable’ (in which case there is an equivocation in the use of “conceivable” between (1) and (2)).
4. Therefore, either the argument is invalid, or we have no reason to think it is sound.

In short, Chalmers’ argument fails by artificially limiting the consciousness-relevant properties to t-physical properties.

*Stoljar’s own position is “o-physicalism” – the view that physical objects have o-physical properties in addition to their t-physical properties – which is, roughly, a modern version of Russell’s neutral monism.