Like many metaphysicians, I think the world is structured. Some truths hold true in virtue of others; some things exist in virtue of other things; some truths are made true by things. I think that there’s only one relation here, and it is the in virtue of relation, that holds between true propositions. For A to exist in virtue of B (i.e. for A to be ontologically dependent on B) is for the proposition [A exists] to be true in virtue of the proposition [B exists]; for the proposition P to be made true by A is for P to be true in virtue of the proposition [A exists].
Sometimes I hear the objection that this assumes that propositions are themselves fundamental constituents of reality. This objection is misplaced, for the view does not assume that. I can’t really see why one would think it did, but I’ve heard it enough times that I think it’s worth spelling out why I don’t think it does. If I’m just confused, I’d like to hear why!
Suppose you have an in virtue of chain that terminates in the proposition P. All that is entailed by this is that what P says to be the case is fundamentally the case; but that P exists may well be true in virtue of something else, and so P may itself be a derivative entity, despite its content being a fundamental truth.
Here is a toy example, just to illustrate the consistency. Suppose for every proposition, p, that p exists is true in virtue of the fact that it is possible for someone to entertain the content of p. So P might be true in virtue of Q, which is itself fundamental. But the proposition [Q exists] needn’t be fundamental. On the toy proposal, [Q exists] is true in virtue of [Possibly, someone entertains the content of Q]. Of course, now I’ve invoked another proposition, call it R; so if it is to be a derivative entity I need to invoke a new instance of the in virtue of relation. [R exists] in virtue of [Possibly, someone entertains the content of R]. And now we have another new proposition, so need a new instance of the relation; and so on, and so on. We generate an infinite sequence of in virtue of relations. But this is not, I think, a vicious regress. The success of an instance of the in virtue of relation never depends on the success of the instance of the relation it ‘generates’. P obtains in virtue of Q, and that generates a new instance: [Q exists] in virtue of R. But the success of ‘P obtains in virtue of Q’ doesn’t depend on the success of ‘[Q exists] in virtue of R’, for it doesn’t matter to P’s being grounded in Q whether or not Q is fundamental. That Q is not a fundamental existent is nice, but it’s irrelevant to Q’s ability to be the relata of the in virtue of relation. So the fact that there is an infinite sequence of in virtue of instances is, I think, unworrying.
Now, I don’t particularly recommend that account of what grounds the facts concerning the existence of propositions, but it’s clearly just a placeholder for a better account. So I think taking propositions to be the relata of the in virtue of relation simply has no consequences for whether or not propositions are fundamental constituents of the world.