blog.kennypearce.net

Omnipotence and the 'Delimiter of Possibilities' View

Aquinas notes that some analyses of omnipotence have a serious problem: they reduce the apparently substantive claim "God is omnipotent" to the trivial claim that God "can do all that He is able to do." Now, perhaps it is true that to be omnipotent is to be able to do everything God is able to do (or at least that omnipotence entails this), but this is hardly an illuminating analysis.

In several places in his Anselmian Explorations, Thomas Morris defends the view that the Anselmian God is the 'delimiter of possibilities.' This view has been endorsed by other Anselmians, and I am inclined to it myself. What Morris means by it is that many apparently conceivable worlds are in fact impossible precisely because it is impossible that God should permit them. God exists necessarily, and no world can be actual except by God's permission. Hence if God's character (or whatever) prevents him from permitting a state of affairs, then that state of affairs is not genuinely possible.

When this view is combined with a result theory of omnipotence, Aquinas's worry recurs. The simplest result theory is the Leibniz-Ross theory (the 'Ross' is James F. Ross, who advocated this view in his Philosophical Theology (1969) and "Creation" (1980)). According to this theory, God's omnipotence consists in the fact that he gets to choose, among all the possible worlds, which one is to be actual. According to the combination of the delimiter of possibilities view with the Leibniz-Ross theory of omnipotence, God is omnipotent insofar as he can actualize any possible world, and a world is possible just in case God can actualize it.

Now, the Leibniz-Ross theory is widely rejected for an entirely different reason: Plantinga famously pointed out (in The Nature of Necessity and elsewhere) that it has the consequence that, necessarily, no creature has libertarian freedom. Adopting a more complicated result theory of omnipotence, one that allows for creaturely libertarian freedom, will make it more complicated to state the exact form of the circularity, but the problem will, I think, arise as long as one holds that God's omnipotence consists in his ability to actualize some specified subset of the possible worlds, for this requires the class of possible of possible worlds to be determined in advance of the determination of what God can actualize. Furthermore, it does not seem to matter whether one uses strong actualization or (with Flint and Freddoso) weak actualization in the analysis.

The Pearce-Pruss theory of omnipotence neatly avoids this worry. In our view, God's omnipotence consists in the fact that, necessarily, God brings about whatever he wills, and God's willing is perfectly free. We argue that God's perfect freedom of will is consistent with its being impossible that he should will certain propositions. This provides an account of omnipotence which does not require that the class of possible worlds already be determined before the scope of God's power is determined.

(Cross-posted at The Prosblogion.)