The Principle of Sufficient Reason can be construed as the principle which states that for any fact having the property of being true or false (contingently), there is some corresponding reason which explains why the fact should have that property. A Deflationary Theory of semantics commits one to saying that propositions represent modes of speaking, but that propositions do not belong to the set of all contingent beings. In other words, propositions do not really exist themselves, as predicate-bearing substances which have the predicate ‘being true’ or ‘being false’. Instead propositions merely belong to a manner of speaking in which we represent the world to ourselves.
Now, if the PSR is truly a Canon of rational thought, and if for it we have no defeater, and if the deflationary theory of semantics is accurate, then perhaps we can say that the PSR applies properly only to substances themselves, and not, therefore, to propositions. This would mean that anything which is such that it may not have been itself requires explanation, but that propositions do not themselves always require sufficient explanation since propositions themselves are not anything.
The obvious advantage of this view is that one can still rationally obviate God’s existence, which is the only candidate for ‘ultimate’ explanation to which the principle intuitively directs us, and yet also accept the truth value of counterfactual conditional propositions about the exercise of libertarian free will. Perhaps a disadvantage would be that it is hard to see how one could speak about ‘events’ as substances, but that would only make a Kalam cosmological argument more difficult to construct, and in any case construing the Principle of Sufficient Reason as applying only to substances, and sets of substances, may be more in line with our rational intuitions.
Thus, the PSR would be construed as ‘any thing which exists such that it is not logically necessary that it exists, has some explanation of why it exists in terms of a grand story to which nothing more could be added‘. The obvious problem is that many philosophers have immense trouble applying the concept of necessity to beings rather than to propositions. Thus, a logician might say that “not every statement is both true and false” is a necessary proposition, but find it unintelligible to say that a necessary being exists. Necessity can apply to a being in the following sense: if the modal status of some being is necessary (that is to say, if the being in question exists in all logically possible worlds), then it seems that that being is a necessary being. Thus x is a necessary being iff x is a being and x exists in all logically possible worlds. For one to say that this is unintelligible they must argue that it is not logically possible that a being exist in all logically possible worlds, since that would imply that the idea of that being involves existence (or ‘necessary’ existence). The objection would then be the old Kantian objection that existence cannot be a first order predicate. We can deny this by either saying that it can be a first order predicate, or by saying that even if it isn’t a first order predicate, ‘necessary existence’ is a first order predicate. So long as even one of those avenues remains dialectically open to us we can run the arguments for a necessary being through.
Once saying that some being is necessary in the modal sense becomes intelligible it may follow logically that such a being exists (for instance by a modal ontological argument such as the one presented by Plantinga), but even if it didn’t the PSR as construed above would be sufficient to run a cosmological argument from contingency strong enough to prove with deductive closure that a necessary being does exist.
I don’t know that I accept this construal of the Principle of Sufficient Reason, but if I did it would certainly allow me to be a Molinist (which would be neat). The above is just my outline of the only way I think I could see myself marrying my intuition about sufficient reason with believing in counterfactual conditionals of the Molinist type. Just deny that necessity applies to propositions as though propositions really existed and had properties, and ground necessity in either descriptions of the real world which are logically necessary, or else ground them in the nature of that which is real more directly by finding in substances, rather than in their relations with other substances, the ground of modal necessity.