On the Force of "Possibly" in Plantinga's Free Will Defense
(TWD) Possibly, every creaturely essence suffers from transworld depravity.
According to Plantinga, if a creature suffers from transworld depravity, then *every* God-accessible world (i.e., every world that God can create) is one at which the creature goes wrong at least once.
So if some free creature FC is transworld-depraved, then we have:
1) Necessarily, if God actualizes FC, then FC goes wrong at least once.
And if every creature is transworld-depraved, then we have:
2) Necessarily, for any x, if x is a free creature, then if God actualizes x, then x goes wrong at least once.
If so, then if Plantinga is using "possibly" in (TWD) in the metaphysical sense (as in (1)), then (TWD) amounts to:
3) Possibly, it's necessary that for any x, if x is a free creature, then if God actualizes x, then x goes wrong at least once.
But Plantinga accepts S5 modal logic. If so, then he accepts the following axiom of S5 modal logic:
(AS5) If it's possible that P is necessary, then P is necessary.
But if so, then by (3) and (AS5), (TWD) reduces back to (2):
2) Necessarily, for any x, if x is a free creature, then if God actualizes x, then x goes wrong at least once.
But this can’t be what Plantinga meant to assert, can it? For now we don’t just have a defense – we have a theodicy. For we have an account that’s not just possibly true, but necessarily true. And you can’t have a stronger theodicy than one that’s necessarily true.
The problem, though, is that it’s extremely implausible to think that (2) is true: is there some shortage of souls, so that there is no possible creaturely essence that has at least one God-accessible world at which it never sins? Plantinga grants that there are possible worlds at which free creatures never sin; it’s just that none of them are worlds that God can actualize. Is this really plausible?
I think that this problem (in addition to some things that Plantinga says) leads many to say that Plantinga's "possibly" shouldn't be construed as *metaphysical* possibility (i.e., that there is, as a matter of fact, at least one possible world at which it's true), but rather as *epistemic* possibiliity (i.e., *we can't rule it out*, given all our evidence, that it's metaphysically possible).
Now the relevant notion of epistemic possibility can be construed in at least two ways:
(Strong EP) We're not quite justified in thinking that P really is metaphysically possible; however, we're not justified in thinking that P is metaphysically impossible, either -- given our evidence, it could go either way.
(Weak EP) We're not justified in thinking that P is possible; however, although it's implausible to think that P is possible, we can't *conclusively* rule it out that P is possible.
Of course, the theist hopes that (TWD) is at least strongly epistemically possible; if it's merely weakly epistemically possible, one wonders how interesting the Free Will defense really is: "Sure, it's pretty far-fetched to think that every essence suffers from transworld depravity, but it hasn't been *conclusively* ruled out as imposssible -- hooray!")
The problem is that the same objections arise all over again for the strong epistemic possibility construal: it seems *implausible* that it's metaphysically possible. It seems that there are infinitely many free creaturely essences that God could actualize; are we to think that *every one of them* is such that *all* of the worlds in which they always freely do right are inaccessible to God? And as I’ve mentioned before, it looks to be a part of conservative Christian theology that angels exist, are free, and that some never sin. But if so, then it’s not necessarily true (because it's not *actually* true!) that all free creatures are transworld depraved. Thus, not even theologically conservative Christians believe it’s epistemically possible – let alone metaphysically possible. Even if the Old and New Testaments don't force belief in a doctrine of sinless angels, it needs to be pointed out (again) that Christians who endorse Plantinga's Free Will Defense *have no choice* but to reject such an idea.
What about weak epistemic possibiity: is it true that we can't *conclusively* rule it out that every creaturely essence would freely go wrong in all God-accessible worlds? Well, maybe for non-theists, some non-Christian theists, and some moderate and liberal Christians. But again, it doesn't seem to be even *weakly* epistemically possible for theologically conservative Christians (recall the problem of angels who always freely do right).
What, then, does Plantinga's Free Will Defense really show? In light of the previous discussion, just this: for people who aren't theologically conseverative Christians, it's not conclusively ruled out as impossible that the Free Will Defense saves theism from the logical problem of evil; but for the theologically conservative Christians, it is.