Friday, June 5, 2015

A Modal Argument for Molinism

For your consideration, I have constructed what I am calling a “modal argument for Molinism,” relying on free will. These concepts have cropped up and even been developed other places, so it’s not particularly original. However, it might be helpful in framing the debate between Molinists and non-Molinists.

1.     If libertarian free will (LFW) is possible, then there are truths about how possible libertarian free agents use their LFW.
2.     LFW is possible.
3.     Therefore, there are truths about how possible libertarian free agents use their LFW.
4.     It is possible that all counterfactuals of creaturely freedom (CCFs) have truth-values.
5.     If (2) & (4), then there are at least some libertarianly true CCFs concerning creaturely essences who never exist.
6.     Therefore, there are at least some libertarianly true CCFs concerning creaturely essences who never exist.
7.     If (6) and OMNI, then a middle knowledge account of God’s omniscience is correct.
8.     OMNI.
9.     Therefore, a middle knowledge account of God’s omniscience is correct.

Let “OMNI” stand for the thesis that for any proposition p, God knows p and does not believe not-p. (1) seems to be definitional. After all, it means that there is at least one possible world where libertarianly free actions are exercised. (2) may need some defense; admittedly, this won’t do a bit of good against the objector who thinks that, as a matter of fact, LFW is incoherent. But it certainly seems to us to be the case that we possibly have LFW, and, typically, we don’t modally perceive something so strongly that isn’t even possible. That is to say, our modal intuitions, on the face of it, count for at least something, and in the absence of other, stronger intuitions or evidence, we are justified in holding it. (3) is an entailed conclusion.

Similarly to (2), (4) can be denied by those who think it’s impossible that there are any true CCFs describing any libertarianly free agents. This might be open theists, who claim that such truths would render LFW actions impossible, or those who think there are no CCFs at all. However, people should accept (4) independently of (2). That is to say, one can accept (4) whether or not he believes in LFW, or even its possibility. (5) is an analytical truth: if it’s possible that all CCFs have truth-values, and LFW is possibly true, then there is a possible world such that creatures with LFW exist and would use it to perform particular actions and make particular choices. Thus, there are truths about how possible creaturely essences would act with LFW who never exist. (6) is thus an entailed conclusion. (7) is also analytic: granting OMNI, and (6), it just seems almost definitional. God would know about these libertarianly free CCFs in possible worlds, and they are not true due to his decree.[1] Thus, middle knowledge is correct!

The upshot of this entire argument is not to convince the anti-Molinist to become a Molinist. Rather, it’s to frame the discussion in terms of the following: if libertarian freedom and CCFs are even possible, then they are known to God prior to the creative decree—even if, in fact, God does not actualize a world with libertarianly free creatures—or even if God does not actualize a world at all![2]

[1] I do recognize classical Thomists would be scandalized at this, and thus they have an out by denying (7). However, I find LFW on Thomism to be sketchy, at best. But to each his own.

[2] Technically, this is not possible. However, all I mean by this is that God doesn’t create a world; the world he actualizes is just a world in which the Trinity exists alone.


  1. Thanks for your thoughts on this.

    However, I believe there are a number of problems with the above argument.

    First off, according to your definition of "OMNI," premise (8) is obviously false. You claim that "OMNI" stands for the thesis that "for any proposition p, God knows p and does not believe not-p."

    That can't be right. God would believe all sorts of false things if that definition were right since, obviously, there are all sorts of false propositions (e.g., "2 + 2 = 3," "The Earth is ten miles from the sun," etc.); since you claim that, for ANY proposition p, God knows p, he would have to somehow *know* false propositions. That won't work.

    What I think you mean to say is that for any TRUE proposition p, God knows that p (and does not believe that not-p). But after revising "OMNI" in this way and, thus, allowing for (8) to be true, (1) becomes question-begging for the majority of Open Theists out there (and even many Calvinists, e.g., Jonathan Edwards). Some OT's will allow there to be truth-VALUES concerning future indeterministically free actions (i.e., sometimes called "future contingents"), but it doesn't follow that such truth-values are "true" since, depending on whether "will" and "will not" propositions are contraries or contradictories, they *could* all be "false" at a time (the principle of bivalence merely stipulates that for any proposition p, p is either true OR false. So, for propositions that are contraries, such as, say, "All animals are dogs" and "No animals are dogs," the fact that these propositions have a truth-value does NOT mean that one of them is TRUE...the propositions just stated, for instance, are both false.). Some Open Theists, then (such as myself), will go on to argue that "will" and "will not" propositions are actually contrary propositions rather than contradictories (I do this in my thesis, and could direct you to some good literature on this particular subject if you'd like).

    Other Open Theists, however, will reject (1) even if you revise it to read "truth-values" rather than "truths." That's because these Open Theists will argue (along with Aristotle) that future contingents--which is what future libertarianly free actions are--are neither true nor false, i.e., they deny bivalence applies to such a category of proposition. Future contingents, on this view, don't have a truth-value; they are somewhere "in-between" true and false (this requires multi-valued logics, and a number of these logics have been formally developed quite apart form any concerns in philosophical theology).

    So, whichever way you slice it, premise (1) is question-begging. On top of this, moreover, (5) does not follow from (2) and (4). In fact, NO libertarian theist who isn't already a Molinist will think that (5) follows from (2) and (4). See, for example, classical Arminian philosopher David Hunt's response to Craig's essay in Divine Foreknowledge: Four Views (2001, pp. 151-154).

    There is more to go into but, suffice it to say, either premise (8) is false or, if true, then (1) is question begging. If (1) is revised to read "truth-values" intend of "truths," then I would accept it (though some other OT's wouldn't), but (5) still wouldn't follow from any of the premises before it. Therefore, it is difficult to see why any libertarian who isn't already a Molinist ought to regard it as sound.

    1. Hi Eli, thanks for the comment! You're right with respect to OMNI; elsewhere I have used precisely this definition ; I just typed it too quickly and didn't go back and check. My bad!

      (1) should only be rejected if one thinks the antecedent is true and consequent is false; this I can understand the open theist doing, but I'm not begging the question by asserting it. I'm certainly not assuming (9), nor am I assuming (3) in asserting (1).

      However, the open theist objection to (1) is noted. It seems the first account won't work, at least as far as I can tell, because we're not dealing with "will" and "will not" propositions, as much as we are "will" and "not the case that...will." The principle of bivalence is that p is either true or false. Let "I will go to the store" represent p. In that case, for the principle of bivalence, it is either the case that I will go to the store or it is not the case that I will go to the store; either p or not-p. Of course, you're right that this will logically result in will or will-not statements. However, we can see, using the animal example, the two samples are not symmetrical. If they were, we would go from "all animals are dogs" to "not all animals are dogs;" clearly, if bivalence holds, we have either p or not-p; but it's just not the case that they're both false. It looks like you're making the negation come within p itself, which is not aligning with the principle of bivalence (there is a difference, as we can see, between "not all animals are dogs" and "all animals are not dogs" and "all animals are dogs"; even though "not all animals are dogs" and "all animals are not dogs" are sufficient, if true, to render "all animals are dogs" as false, they are not equivalent to each other and not what we mean when we talk about libertarian propositions).

      I don't see any reason to think that the consequent of (5) doesn't follow from (2) or (4), and if it ends up entailing Molinism, then this was rather the point! So suppose there is some possible world such that libertarian actions take place (that is, grant 2). Suppose further that there are then truths about how agents would and would not act for all counterfactuals of creaturely freedom. This will include the world in which LFW is possible. This means there are at least some libertarianly true CCFs concerning creaturely essences who never exist.

      Of course, a good way to look at (1) is to consider the question of God's creative action. Suppose he never created, and so actualized a world where only the Trinity exists, for example (or no LFW agents other than God, anyway). Surely that world is not the only possible world, nor is it an impossible world. Thus, there is at least one other possible world, namely this one, in which God acts libertarianly to bring about us. God is a possible libertarian agent (surely it's not impossible for God to have such freedom), and thus there are truths about how possible libertarian agents use their LFW. Since we've affirmed both antecedent and consequent, we can grant (1). But if (1) is true, and (2), and (3) follows, it's difficult to see what the argument is for why we should restrict such a domain to God alone. I'm not begging the question; it just seems to me, modally, that such is possible. I'm not saying, for example, "well open theism is false, so that's why I assert (5), or (1), or whatnot." Instead, I am saying, "it seems to me (1) is true, and that gets us through a chain that gets Molinism as true, and therefore open theism is false." So be it, but I'm not question-begging!

      Now if what you mean is that those who don't share such a modal intuition won't be convinced, I can readily grant that (I gave up long ago of trying to convince others of much)! I hope you have a great week, and look forward to more of your thoughts!

  2. Hey Randy,

    Thanks for the response.

    Regarding your first paragraph concerning the difference between "it will be the case that p," "it will not be the case that p," and "it is not the case that it will be the case that p," there are strong arguments to support the idea that the latter two are logically DISTINCT propositions even though we often treat them as synonymous in everyday language (it's the difference between the scope of negation, and this, as it turns out, is going to be very relevant to this debate). The negation of "will" is NOT "will not" but, rather, "not[will]." Future indicatives like will/will not form a contrary pair. All of your concerns in the first paragraph I've addressed in-depth in my Master's thesis. I can gladly email you a copy if you'd like (and it would save me a lot of time tying here :) ).

    As far as (5) goes, it does not follow from (2) and (4) because, on the standard semantics for counterfactuals, "would" and "would not" counterfactuals (which is what CCFs are supposed to be) are not contradictories but contraries. A "would" counterfactual is contradicted by a "might not" counterfactual, and a "would not" counterfactual is contradicted by a "might" counterfactual (see p. 53 of Moreland & Craig's book called Philosophical Foundations for a Christian Worldview). Thus, even if (4) were true and all CCFs had a truth-value, it wouldn't follow that there are at least some TRUE "would" and "would not" CCFs (the only CCFs that matter to the Molinist) since, again, being contraries rather than contradictories, they could all be false. I spell all this out in my thesis and engage heavily with the relevant literature on the Stalnaker and Lewis semantics for counterfactuals that much of this debate turns on. But you can very quickly see how someone who grants you (1)-(4) and STILL deny (5) by looking at David Hunt (whom I mentioned above). Hunt agrees with (1)-(4) but denies that God has the knowledge claimed in (5) if that knowledge is supposed to be *middle* knowledge. For Hunt, and other non-Molinist libertarians, this would have to be acquired in God's free knowledge for reasons Hunt gives in the reference I already cited above.


  3. One last thing about question-begging.

    If you're not trying to convince an Open Theist with this argument that a middle-knowledge view of God's omniscience is correct, then I agree that the Open Theist ought to simply regard (1) as false, rather than question-begging.

    However, after giving it some consideration, there *is* some question-begging going on in this argument that any astute simple-foreknowledge Arminian will likely point out. When you claim, for instance, that "there are at least some libertarianly true CCFs concerning creaturely essences who never exist" in (5) and (6), the suppressed premise I see here is that you are assuming that these CCFs are true *logically prior* to the divine decree. But that is precisely what is at issue. As Craig himself points out,

    "Christian theologians have traditionally not disputed that God has knowledge of true counterfactuals and, hence, of the conditional future contingent events they describe. What theologians did dispute, however, was, so to speak, *when* God has such hypothetical knowledge...[e]verybody agreed that logically prior to God's decree to create a world, God had knowledge of all necessary truths, including all the possible worlds he might create. This was called God's 'natural knowledge.' It gives him knowledge of what 'could' be. Moreover, everyone agreed that logically subsequent to his decree to create a particular world, God knows all the contingent truths about the actual world, including its past, present, and future. This was called God's 'free knowledge.' It involves knowledge of what 'will' be. The disputed question was where one should place God's hypothetical knowledge of what 'would' be. Is it logically prior or posterior to the divine decree?" (2011, Four Views on Divine Providence, pp. 80-81).

    Your argument is meant to show that God knows these CCFs to be true via his middle knowledge. But in order to do this, you are ASSUMING in (5) and (6) that the CCFs you refer to there are a part of God's middle-knowledge (after all, that's why you think you can derive the consequent of (7) from (6) and OMNI)! But that's begging the question at hand, which is WHEN does God know true CCFs.

    So, from a simple-foreknowledge Arminian's perspective, you've begged the question about when true CCFs are known by God.

    1. Thanks for these thoughts, Eli! I'd definitely appreciate a chance to read your thesis! Where did you go? Did you end up pursuing a PhD? Or was it a Euro/UK PhD where "thesis" just is doctoral?

      I think we're in total agreement about contradictories, if I understand you correctly, with respect to your first paragraph in the other response. I also appreciate your reminder about the counterfactual square of opposition; that's quite right also.

      I do think that instead of begging the question, I'm simply showing what I take to be an entailment. For example, someone could affirm (5) and not be a Molinist: they'd just deny (2) or (4). With respect to the Arminian, while I think that if they accept it, it will entail middle knowledge, I think this is half the fun!

      But I think you're right: the worry is, as you say, that this won't convince anyone of anything who's not already a Molinist.

    2. This comment has been removed by the author.

    3. Hey Randy,

      I tried typing my email address above but it wouldn't publish...mmm, I'll just send you my thesis via message on Facebook, I guess :)

      As far as graduate work goes, I'm a PhD student (Philosophy) at the University of Arkansas.

      Prior to that, I earned an MA in Philosophy here at Arkansas (2014) and an MA in Philosophy of Religion from Denver Seminary (2011).



Please remember to see the comment guidelines if you are unfamiliar with them. God bless and thanks for dropping by!