Here's my argument:
1. In our experience, every moment of spacetime is caused by a previous moment of spacetime.
2. There are two concievable options: either (a) it is possible that moments of spacetime can be uncaused/caused by something nonspatiotemporal or (b) it is impossible for moments of spacetime to be caused by anything except previous spacetime.
3. On the hypothesis that (b) is true, it is 100% likely that we would observe (1). On the hypothesis that (a) is true, there must be less than a 100% chance that we would observe (1) because on the (a) hypothesis, spacetime can be caused by nonspatiotemporal things or uncaused altogether.
4. We should prefer the hypothesis that does the best job of predicting the data, all else held equal.
Conclusion: From 3 and 4, spatiotemporal moments are probably always caused by previous spatiotemporal moments. The only way this can happen is if the universe is eternal, or if something like Quentin Smith's scenario for a self-caused universe is correct.