Abstract

A number of philosophers hold that some types of mental states are composed of two or more mental states. It is commonly thought, for instance, that hoping involves the desire for some outcome to occur and the belief that such an outcome is possible (but has yet to occur). Although the existence of combinatory states (CS’s) is widely accepted, one issue that has not been thoroughly discussed is how we know we token a given combinatory state. This paper aims to fill this lacuna. I do so by first discussing one way of knowing our CS’s—namely, by knowing we token the relevant constituting states, and then inferring that we have the relevant CS from such a knowledge-base. I argue that while anti-skeptics of self-knowledge should embrace the view that we can know our CS’s in this manner, this way of knowing we possess such states is quite demanding. Given the latter, I proceed to examine whether there are alternative ways we can know our CS’s. I defend the view that given the tenability of particular accounts of self-knowledge for non-CS's, we can avoid the view that we only know our CS’s by in part knowing the constituents of such states.