Molinism and Circularity
May 19, 2015 — 10:40

Author: Kenny Pearce  Category: Divine Providence Molinism  Tags: , , , , , ,   Comments: 9

Yesterday, I discussed Thomas Flint’s response to the grounding objection in chapter 5 of Divine Providence: The Molinist Account. Today, I want to discuss his response to Robert Adams in chapter 7.

Adams’ objection turns on a notion of explanatory priority which, Flint complains, is not adequately defined. Flint argues that there is an equivocation in the argument, and that Adams relies on a transitivity assumption which is not plausible when applied across the different sorts of priority involved. I think, however, that Flint is mistaken on both counts: first, the notion in question is not equivocal. Rather, it is a genus containing several species. Second, transitivity is not actually required. What’s required is just an anti-circularity principle. The anti-circularity principle is abundantly well-justified across the entire genus.

The notion of priority here corresponds to the notion of objective explanation. That is, A is prior to B iff B because A. That’s simple enough. Of course, there are many different uses of ‘because’ and I’m inclined to agree that the anti-circularity principle won’t apply to all of them. That’s why we require that the because or priority here track objective explanation, i.e., that A really be a reason why B is true, and not merely a fact that helps make B intelligible to some particular mind. It is extremely plausible to suppose that there can be no cycles in chains of objective explanation.

The types of priority/explanation at issue include these:

  1. The priority of reasons (and, more generally, considerations) to actions (whether divine or creaturely).
  2. The priority of God’s creative act to all creaturely activity.
  3. The priority of causes to effects.
  4. The priority of free choices to free actions.

Now, it is, as I said, extremely plausible that an anti-circularity constraint applies here. For instance, it is incoherent to suppose that I should choose to act in a certain way because I am going to act in that way. Similarly, if my action causes it to be the case that P, then P can’t be among the reasons for my action, since (barring overdetermination, etc.) P won’t be true unless I take the action. (Of course, I might take the action because taking the action will cause it to be the case that P. That’s different.)

Now, let C be a proposition describing a total circumstance and let A be a proposition stating that a creature takes some free action in that circumstance. The Molinist is clearly committed to:

(1) C []-> A is prior to God’s decision to weakly actualize C.
(2) God’s decision to weakly actualize C is prior to the agent’s having the reasons, considerations, etc., which lead her to choose A.
(3) The agent’s reasons, considerations, etc., are prior to her choice that A.
(4) The agent’s choice that A is prior to A.

By the anti-circularity constraint, this implies that neither the agent’s choice that A, nor A itself, is prior to C []-> A.

But then why is C []-> A true? If the Molinist says, for no reason at all, she runs into the randomness objection. The anti-circularity constraint prevents the Molinist from saying it’s because of the agent’s choice or the agent’s action. The Molinist obviously can’t say it’s due to God. If it’s due to the agent’s essence, nature, character, etc., then we’re presupposing a compatibilist theory of freedom and don’t need to bother with all the complexities of Molinism. There’s a serious problem here, and Flint hasn’t defused it.

(Cross-posted at

  • Heath White

    If the Molinist says, for no reason at all, she runs into the randomness objection.

    Kenny, could you spell this out?

    May 21, 2015 — 9:30
  • If, prior to my being in C, C []-> A is true for no reason (where C is the total circumstance and A is my action), then there is no reason why I do A in C. If we cannot give an explanation of my doing A in terms of my causal activity and/or my reasons, motives, character, etc., then how can my alleged action A be distinguished from a spasm? A purely random event, far from being a free action, is not an action at all. But if the CCFs are true for no reason and my action depends on them, then we are in exactly this situation.

    Note that I’m not assuming compatibilism here. There are lots of libertarian strategies for saying that my action can have an explanation in which I am appropriately involved. One such strategy is the theory of Agent Causation (causes explain their effects). But, according to the argument, none of these things that might figure into the libertarian explanation can possibly be prior to C[]->A, since all of these things (my causal activity, motives, beliefs, character etc.) depend on God’s decision to actualize C, and that decision was based on God’s knowledge of C[]->A.

    May 21, 2015 — 9:43
  • Michael Almeida

    . . .if my action causes it to be the case that P, then P can’t be among the reasons for my action

    Hi Kenny, I think that’s an interesting thing to say. But what about cases like this: My studying for the test causes it to be the case that I will thereby pass the test. But if you ask me the reason I’m studying, I can say, because I will thereby pass the test. It’s false that I will thereby pass the test, if I do not study. But it seems like a perfectly good reason for studying.

    May 25, 2015 — 13:55
  • Hi Mike,

    This is an interesting case. Although I agree that we can say this, I think we have to regard it as elliptical for something else. I might believe that I will pass the test by studying, but if that belief is not somehow implicitly conditional then I don’t see how I can hold that belief prior to deciding to study (if I genuinely decide). So I think it is more correct to say either

    (a) I am studying in order to pass the test


    (b) I am studying because (I believe that) if I study I will thereby pass the test.

    It’s plausible to suppose that your sentence (“I am studying because I will thereby pass the test”) is elliptical for (b). Thus I’d want to assimilate it to the case of taking the action because I believe that taking it will cause the proposition to be true. Of course, there are all sorts of interesting issues here with foreknowledge and time travel cases, but even in those tricky cases, it’s hard for me to see how a belief about my own future actions could figure as a reason (in the subjective, deliberative sense) for those very actions.

    May 25, 2015 — 14:09
  • Michael Almeida


    Just a quick comment on the randomness problem. Assume (as seems likely) that our world is indeterministic. Suppose I ask at t0 whether the ball would fall were it dropped at t1. It’s an indeterministic world, so it might not drop, but surely it is true that were it dropped at t1, it would fall. So, why is the counterfactual D ☐⟶ F true? The most you can give is a probabilistic explanation, I think. But that seems fine. What I don’t see is why we can’t give a similar explanation for true CCF’s. Maybe they are not true in the case of agents whose lives involve few or no self-forming actions. But it is hard to see why they are otherwise not true.

    May 25, 2015 — 14:29
  • I agree that any adequate version of libertarianism will include some account of character and character-formation. Furthermore, a version of libertarianism which includes an adequate account of character will probably not be vulnerable to the randomness objection. Finally, any adequate account of character will allow a lot of CCFs to be true.

    Having conceded all of that, three problems remain:

    (1) I’ve never seen a libertarian account of character that seems adequate to me. Generally speaking, it seems compatibilists can do much better on this score. On the other hand, this is not my specialty and there’s lots out there I haven’t read. Within my actual primary specialization (early modern) the closest thing I know is Reid, but Reid goes primitivist and refuses to philosophize further just at the point where things get really tricky. (This is typical of Reid; he sometimes frustrates me.)

    (2) I don’t see how this strategy can get as many CCFs as the Molinist wants. I don’t see, for instance, how we could get anything like conditional excluded middle for finite creatures like us.

    (3) If the CCFs are made true by my character, and my character is created by my self-forming actions, and my self-forming actions are free choices I make because I am in certain circumstances, and God puts me in those circumstances because of the true CCFs about me, then we’ve got exactly the kind of circle we’re worried about. If, on the other hand, God ‘sees’ my character ‘in the realm of the possibles’ and that character makes the CCFs true (prior to my choices), then we’re back at (Leibnizian) compatibilism. Or so it seems to me.

    May 25, 2015 — 19:56
  • Josh

    Does the article, “Explaining Counterfactuals,” by Pruss and I, help here?

    May 29, 2015 — 23:16
  • Josh – Yes, that looks like it might! I somehow missed that article; thanks for pointing it out. I’ll have to read it.

    May 30, 2015 — 9:25
  • Josh – Just got around to reading the paper. That’s a very nice move. I need to think about it more. But this might convince me that Molinism doesn’t create new worries about libertarianism and explanation (compared to whatever worries libertarianism already had). Hmm…

    June 3, 2015 — 9:55
  • Leave a Reply to Kenny Pearce Cancel reply

    Your email address will not be published. Required fields are marked *