If causal closure of the physical holds, then there are uncaused physical states of affairs. Thus, at least one of the following two theses is true:
- Some physical states of affairs have a non-physical cause.
- Some physical states of affairs have no cause.
Imagine two worlds. In w1, there are non-physical causes for all the physically uncaused physical states of affairs. In w2, there are no causes for any of the physically uncaused physical states of affairs. Moreover, the physical parts of w1 and w2 are exactly alike, and match our observations. Is there very good reason to prefer w2 to w1 as a hypothesis accounting for our observations? Well, maybe sometimes: it depends on how weird the non-physical stuff in w1 is. But in general, no. Roughly, what w2 gains in parsimony it may lose in explanatory value.
So we do not have very good reason to believe (2) and deny (1). And (1) has a serious advantage over (2): unlike (2), (1) is compatible with our PSR-ish intuitions.