Premises:
- If there are any Molinist counterfactuals, there are ungrounded true contingent propositions.
- Propositions reporting divine beliefs are grounded.
- If p is a contingent truth (i.e., true proposition), then either God's belief is explained constitutively or causally by p, or p is explained constitutively or causally, or there is some third truth that explains both p and God's belief constitutively or causally.
- An ungrounded truth cannot be explained causally.
- An ungrounded truth cannot explain causally.
- When a truth p explains q constitutively, something that grounds p grounds q.
- God believes every truth.
Premise (3) is a way of working out the idea that God's beliefs are knowledge and cannot be merely contingently related to what makes them true.
Premise (3) is a way of working out the idea that God's beliefs are knowledge and cannot be merely contingently related to what makes them true.
But wait. I have knowledge that p too, where p is a contingent proposition that say "It's raining now", and my belief that p is contingently related to what makes it true. So it's hard to see why the fact that God's beliefs are knowledge entails that they cannot be contingently related to what makes them true.
If there are any Molinist counterfactuals, there are ungrounded true contingent propositions.
Sorry, should have added this. I think we've had a discussion on this before, because I recall worrying about why (1) would be true. What is the ground for counterfactual claims, what makes them true? What makes them true is a fact that obtains in the closest worlds in which their antecedent is true. So a modal fact grounds the modal truth in CCF's. How is a response along these lines implausible?
1. I should have said "merely coincidentally related".
2. So, let's suppose that there is a modal fact, and facts are not just true propositions (or else you don't have grounding). There is still a puzzle about how the modal fact is explanatorily related to God's act of believing. It's surely not causally related to it (either by the fact causing the belief or the belief causing the fact or some third thing causing both). About the only option I can see is that the fact is (partly) constitutive of God's act of believing. But if so, then it is unclear how the God's act of believing can causally explain why, say, God strongly actualizes one thing rather than another. For then the modal fact would be entering into a causal explanation. But it doesn't seem to be the sort of entity that can enter into causal explanations.
But it doesn't seem to be the sort of entity that can enter into causal explanations.
Right, but this is not a problem unique to God, or to CCF's. We have modal knowledge too and, I agree, the explanation cannot be causal. We have mathematical knowledge and that too is not causal. So we have a rather pervasive problem in explaining how we know many of the things we do that does not seem a special problem for God or especially related to knowing CCF's.
In premise 3, I restrict the principle to contingent propositions.
In the case of necessary propositions, maybe David Lewis is right that it's not a big deal, because you can't merely coincidentally get them right. I expect he's wrong. But there are nice solutions, like saying that necessary truths are grounded in the nature of God, and then they're directly available to God, who can in turn make them available to us.
"5. An ungrounded truth cannot explain causally."
Why take 5. to be true? Suppose it is just the case, ungroundedly, that if Peter were in C he would libertarianly do A (p). And suppose God has a faculty whereby he apprehends all truths. Wouldn't p's truth (contingent, ungrounded), along with God's faculty (necessary), causally explain his belief that p?
Alex, I'm not sure an omniscient being needs to know any contingent truths. It certainly looks like God knows everything there is to know iff. he knows which true propositions hold in each possible world. But God can know which propositions hold in each possible world and know only necessary truths. Suppose worlds are denumerable, w0, w1, . . .wn and God knows for each world wn the propositions p0, p1, . . .,pn that hold at wn. So God knows p0 holds (or is true) at w0, p1 holds at w0...and so on. If w1 = @, then God knows all of the actually true propositions iff. he knows p1 is true in w1, p2 is true in w1, etc. If God knows all of those propositions, then he knows every true proposition at every world. But he knows only necessary truths. So he knows every true propostion at every world but he knows no contingent truths. I've set aside Grim and Russell problems for the moment.
Mike:
God only knows all the actually true propositions if he knows that w1 is actual.
Dan:
Causal explanation is explanation by means of causation. While explanation, including causal explanation, is a relation between propositions, causation is a relation between concreta (I think it's a relation between a substance and an event, but many others think it's a relation between events). Causal explanation is related to causation thusly: p causally explains q iff the ground of p causes (setting aside questions of what sorts of causation are needed for it to be explanatory) the ground of q. But then a groundless proposition cannot be a relatum of causal explanation.
God only knows all the actually true propositions if he knows that w1 is actual
He does know that. He knows that the proposition expressed by 'w1 is actual' is true at w1. And that is a necessary truth.
Mike:
Surely one can know that 'w1 is actual' is true at w1 without knowing that w1 is actual, even if in fact w1 is actual. After all, everybody who thinks about it for a moment knows that 'w1 is actual' is true at w1, while it may take serious investigation to know that w1 is actual.
Maybe, though, you are assuming Lewisian modal realism. If so, then I concede that in that setting, God only needs to know necessary truths. But that setting does not, in fact, obtain.
Maybe, though, you are assuming Lewisian modal realism. If so, then I concede that in that setting, God only needs to know necessary truths. But that setting does not, in fact, obtain.
I wonder why you are willing to concede in the Lewsian case. The very same argument you just offered holds for Lewis as we well. Setting that aside, it is also a necessary truth that 'it is true at w1 that w1 is an actual, concrete world'. But that is not true at any uninstantiated world. So, God knows all of the truths that obtain at w1 and that w1 is an actual, concrete world.