## Friday, July 24, 2015

### Incomplete Reductio on Free Will

Let worlds be partial. A complete world is a set of all possibly true and possibly false propositions, together with their truth values. A partial world does not specify truth values for some of the propositions. In a partial world W, p is not either true or false, but either both possible true and possibly false, or either true or false. For every partial world, every possible (i.e., logically coherent) completion of it, whether complete or partial, is also a possible world. Thus, if in W = {p is F, q is T, r is both possibly T and possibly F} then {p is F, q is T, and r is T}  is a possible world, and {p is F, q is T, and r is F} are both also possible worlds. Any proposition which is specified in the same way in every possible completion is considered specified in that way in the original partial world.
Assume that there exists a possible (partial) world where the proposition ‘x is saved’ is unspecified for all x where x is a human, and that there are no propositions of the form “if p, then a is saved” (where a is some human) where p is unspecified (it cannot be true without specifying the truth of a is saved), and where there is no corresponding proposition “if not-p, then a is saved” (where a is some human). In that case, in some possible world, the proposition ‘x is saved’ would be false for all x where x is a human.
The original possible world, call it W, corresponds to a world where individuals have the capacity in themselves to determine, to some extent, their eternal destiny. The argument is not dependent on the actual unspecifiedness of truth values (and thus, granting God’s complete foreknowledge does not get one out of the conclusion). The unspecifiedness of the truth values of all ‘x is saved’ where x is human is intended to model LFW. So long as my salvation or damnation are not dependent on external forces, or at least so long as my damnation cannot be precluded by external forces, the argument should follow. Even assuming a Calmenian approach, where eternal security is granted with irresistible grace, the conclusion will follow, since each human has some choice over whether to enter into that state where they are stuck saved. The only way that the conclusion may be resisted is if propositions in the original partial world require that some people be saved, in which case the salvation of some will be based on external forces.

There are a variety of ways to modify the original world so as to more closely model views of our actual initial position. One may specify that certain special individuals are predestined, but so long as there are not too many of them, the conclusion retains its force. One may specify that the damnation of some is specified in the original world on account of a lack of opportunity, but this does not change the conclusion. One may specify that there are propositions in W of the form “if p, then a is saved” (where a is some human) where p is unspecified, and where there is a corresponding proposition “if not-p, then a is saved” (where a is some human), but in this case the specification of p will either be up to God, in which case he will be choosing sets of humans, rather than individual humans, and a kind of determination is included (I have suggested this as a kind of deterministic Arminianism before), or p is up to humans (or chance). In the case where it is up to humans (because humans have LFW) or chance, human freedom over one’s own salvation is still not retained, although neither is particular election—which strikes me as a worst of both worlds option, although intriguing. There is a plausible exception to the loss of human freedom where p is determined by humans in very restricted cases where p is a moral decision by some human, such as seeking God, but it is difficult to imagine many cases where it would intuitively be of the sort desired here, where either p or not-p will result in someone’s salvation. One further modification is to specify that the world will not end until some minimum number of humans have been saved. This last modification results in some world never ending, and having a number of humans approaching infinity, while others will be front loaded with damned people. It does not, however, evade the conclusion.

Given the argument, then, we effectively have
Premise: humans have LFW
Conclusion: There is some possible world where no one is saved.

The above modifications either limit LFW or change no one to almost no one. In the latter case, the modification does not seem to be of the sort which would alter our consideration of the argument. The argument is intended as a reduction on LFW, and so the first kind of modification is the intended result of the argument. The remaining option is to acknowledge the argument, claim that both the premise and conclusion are true, albeit not in our world. In this last case God could not guarantee that he would save a people for himself. He could know that, as it happens, he would save a people for himself, however, which may be enough. There would be no guarantee that, if God created a world, any of the humans would be saved as opposed to damned. He could create the world knowing how it would turn out, however, but could not constrain the worlds at his disposal to create to include worlds with humans who would be saved as a people for himself. We would expect, in that case, that, had there been no worlds where anyone was saved, God would not have created. We might likewise suppose that of the worlds available to God, on account of LFW, this was the one with, say, the least evil and the most saved people.

We also result in an order of creation as follows:
1.       God exists, and imagines all possible worlds.
2.       God’s foreknowledge of these worlds includes how people would behave in each of these beginning worlds.
3.       God determines how he would interact with each of the possible worlds (and foreknows the humans’ free responses to any of his possible interactions with them, to which he determines his interactions, etc., knowing how each possible action on his part would affect the worlds).
4.       God selects a (best, according to some measure) world, together with the foreknown free actions of humans and of himself.
5.       God creates the world, lets it run, and interacts with it and the humans in it, as foreknown.

Effectively, this appears to be a variety of Molinism. Placing divine choice where the above order places it, that is, above/outside of time, avoids the issue of God being determined by having foreknowledge (which I have argued for before). If we eliminate divine foreknowledge, however, there is no guarantee that the world God creates ends up being one in which anyone at all is saved.