The Metaphysical Muddles of Molinism – Some Thoughts on MacGregor (Part 1)

As part of my ongoing research for my book project on Open Theism (under contract with Cambridge Univ. Press), I’ve recently begun reading Kirk MacGregor’s latest book, Molinist Philosophical and Theological Ventures (2022). I’m looking at the book because he addresses some aspects of an anti-Molinist argument that I’ve been developing over the past few years. My charge is that Molinism entails fatalism and is thus internally inconsistent and so necessarily false. You can hear a simplified version of my argument here, between the 23- and 38-minute marks. I should make clear that MacGregor doesn’t respond directly to me, but rather to one of the premises I use in the argument which is also defended in a paper by philosopher Phil Swenson, whom MacGregor does respond to.

Having looked at the first three chapters of MacGregor’s book so far I am now more convinced than ever that Molinism is little more than a theologically driven exercise in egregiously bad metaphysics. What I mean is that the perceived theological need for Molinism is so strong for some that they will seemingly bite any theoretical bullet to make it work.

What follows is my take on MacGregor’s first three chapters. In a follow-up post I hope to examine chapters 4–6. The last two chapters (on sacred music and universal salvation, respectively) are not of central interest to me, so I shall set them aside.

Chapter 1: Molinism and the Bible: Part One (pp. 1–20)

In Chapters 1 and 2 MacGregor intends to show that Molinism is solidly grounded in the Bible. He begins Chapter 1 by discussing the “proper relationship between philosophy and theology”. In his view, theology should drive the bus because it is the “most advanced discipline of study” (p. 2) and “philosophical constructs cannot legitimately be superimposed on Scripture” (p. 1). MacGregor distinguishes between “genuine” philosophical constructs, which are grounded in Scripture, and “human-made” constructs, which aren’t (p. 6). In response, this part of the chapter strikes me as a kind of naïve Biblical positivism, or perhaps even a sort of fideistic presuppositionalism. Theology may have the most important subject matter (God), but I don’t know how anyone could think that it is the “most advanced” field of study. Mathematics and logic, for starters, are far more technically sophisticated. Heck, many theologians can barely agree on anything, much less the proper method for doing theology! And obviously one cannot approach Scripture or any text in a conceptual vacuum. It is important that we strive to be critically reflective about our philosophical “constructs”, and a big part of that is learning to step back from or bracket our prior commitments and allow ourselves to be seriously challenged by alternative viewpoints, including those found in Scripture. A much better approach is something like the Wesleyan quadrilateral, which suggests that we aim for a sort of reflective equilibrium between our major sources of information (reason, experience, divine revelation, etc.). This encourages a mutual dialog between philosophy, theology, science, and practice, where no one field completely dominates the others. All truth is God’s truth, after all.

Moving on, MacGregor next offers some Scriptural data on divine sovereignty (pp. 8–11) and human freedom (pp. 11–16). Regarding the first, MacGregor cites a handful of prooftexts to argue that “God possesses an absolute degree of control over the activities of created beings”, including things that are genuinely “random” (p. 10). Even taken collectively, I think MacGregor’s prooftexts are weak. Prov. 16:33 is perhaps the strongest one: “the lot is cast into the lap, but its every decision is from the Lord.” Right, but (a) many Biblical proverbs are just general rules of thumb, not absolute universal generalizations, and (b) this could just be talking about one particular lot, namely, the urim and thummim that were part of the high priestly breastplate and by means of which the high priest could inquire of the Lord. I’ll leave it to the reader to consider MacGregor’s other prooftexts: Prov. 21:1; Jer. 10:23; and Job 14:5. I find MacGregor’s discussion of human freedom more convincing as it relies on much longer and thus less contextually ambiguous passages. His main texts there are Deut. 30:11–19; 1 Cor. 10:13; Deut. 11:26–28; Ezek. 18:23,30–32.

MacGregor closes the chapter with “philosophical reflections” on divine sovereignty and human freedom (pp. 16–20). He starts out with a false dichotomy by assuming that God must either “strongly actualize” (i.e., causally determine) or “weakly actualize” (i.e., knowingly set in motion) everything that ever happens. This excludes the possibility, affirmed by open theists, that God might sovereignly decide not to decide everything that ever happens but rather to delegate some of those decisions to His creatures. MacGregor then starts talking about God’s intentions, and he insists that Molinism is in line with the ethical doctrine of double effect (DDE) that while God “foreknows” that bad things will happen, “God does not intend evil—period—much less that good may result from it” (pp. 17–18). I believe this is a major error on MacGregor’s part, one that disistorts Molinism and glosses over a genuine difficulty. On the Molinist system, God deliberately selects, and therefore intends, a complete world history, including all of the evils that ever occur. To be sure, He doesn’t intend these evils for their own sakes, but He does intend them as part of a “complete package” that He selects and weakly actualizes. He presumably wills the whole package including all of its evils because it is somehow necessary to achieve a greater good. My point is that Molinists can’t use DDE to exonerate God from complicity in bringing about evils the way MacGregor is trying to do. Molinists can’t say that the evils were merely foreseen but not intended, as DDE requires. Here’s how esteemed Molinist Alfred Freddoso puts it:

God, the divine artisan, freely and knowingly plan, orders, and provides for all the effects that constitute His artifact, the created universe with its entire history, and executes His chosen plan by playing an active causal role sufficient to ensure its exact realization. Since God is the perfect artisan, not even the most trivial details escape His providential decrees. Thus, whatever occurs is properly said to be specifically decreed by God. (p. 3 in this book; emphasis in original)

Frankly, MacGregor should just bite the bullet and accept the idea that, on Molinism, God intends evil, including moral evil, as Matthew Hart and Daniel Hill have recently argued. They are very clear that if God decrees something, He intends it (p. 4). And not incidentally, Hart and Hill reject DDE (pp. 104ff.)

Chapter 2: Molinism and the Bible: Part Two (pp. 21–37)

The first half of this chapter (pp. 21–29) focuses on Biblical texts that support (a) individual soteriological election and (b) God’s universal salvific will. I won’t say much about this section except that I agree with (b) but not with (a). I think the Bible teaches corporate, not individual, election for all who are, and remain, “in Christ”.

Well, I will say one more thing. On p. 23 MacGregor makes a very questionable inferential leap. Based on the assumption that God has prior relational foreknowledge of those individuals who are chosen for salvation, MacGregor infers (1) that this relational foreknowledge would have to be God’s acquaintance with their individual essences because He has it when the individuals don’t yet exist, and (2) that “full relational knowledge of someone’s individual essence would encompass knowing everything that individual essence, if instantiated, would freely do under any set of circumstances.” Now, I can see how (1) might follow given the assumption, but to my mind that’s a good reason to reject the assumption, for I see no reason to think that individual essences exist, especially not prior to the existence of the individual in question. As for (2), while it is a common Molinist idea, it doesn’t remotely follow from (1). Why should individual essences have to include all that? Why would it be essential to my being me that I would act in this or that way under this or that set of circumstances? Moreover, this makes no sense because a thing’s essence contains properties that it cannot ever lack as long as it exists. To lose or alter a thing’s essence would be for the thing to cease to exist and to be replaced by something else of a different kind. Likewise, individual essences, if they exist, contain properties that a particular individual could not ever lack while existing as that person. To lose or alter one’s individual essence would be for that person to cease to exist and to be replaced by someone or something else. So if everything one would freely do under every hypothetical circumstance were part of one’s individual essence, then the individual in question could not do otherwise on pain of ceasing to exist! Consequently, the individual could not be free in the libertarian sense that MacGregor and Molinists advocate.

The balance of the chapter (pp. 29–37) contains MacGregor’s “philosophical reflections” on the Scriptural data (as he reads it) and tries to show that Molinism straightforwardly follows. He begins (p. 29) by arguing that foreknowledge based on actual occurrences would be providentially useless. I agree. He concludes that foreknowledge must be based on prior hypothetical knowledge and cites Matt. 11:20–24 as backup (p. 30). (The passage doesn’t actually teach anything like Molinism because the hypotheticals there aren’t prevolitional in the sense of being ontologically prior to God’s creative decree.) MacGregor then introduces the Molinist distinction between possible worlds and feasible worlds (p. 31). Possible worlds are grounded in God’s natural knowledge, i.e., God’s knowledge of His own nature as the necessarily existent ground of all possibility. Feasible worlds are grounded in God’s middle knowledge, i.e., God’s knowledge of all contingent hypotheticals and, more specifically, of what any possible individual would do in any fully causally specified indeterministic scenario. As standard Molinist doctrine goes, that much is accurate.

One odd claim that MacGregor makes (on p. 32) is that God may not be able to actualize a feasible world in which everyone is saved because it may be that “some people are saved only if various lost people exist”. I don’t get this. If Molinism is correct, then why can’t God just actualize only those individual essences that would accept God’s salvation? If necessary, God could put each of them in their own virtual Matrix-like world to elicit precisely the kinds of free choices that would lead them to salvation. Furthermore, if God knows that Bob would be obstinate unto damnation no matter what circumstances he were placed in, then why can’t He actualize a different individual essence, say Steve’s, in Bob’s place, where Steve is such that he would freely receive salvation in those circumstances? If God’s got an infinite number of individual essences to work with—and why wouldn’t He in the Molinist system?—then it seems highly probable, if not inevitable, that God would have had at least one feasible world available to Him in which no one ever sinned and in which everyone is saved.

Finally, MacGregor presses an objection against “Calvinism” (i.e., theistic determinism) that, because of middle knowledge, God knows more on Molinism than He does on “Calvinism”: “the Molinist believes that on Calvinism there are logically knowable truths that God does not know, such that the Calvinist subscribes to a conception of a less than omniscient God” (p. 36). While I’m no fan of theistic determinism, this is nothing but a cheap shot on MacGregor’s part. He’s basically saying “My God is bigger than your God if my theory is correct and yours is wrong”. No self-respecting theistic determinist should be impressed by that because they won’t concede that MacGregor’s theory is correct. This is like a universal possibilist like Descartes saying to MacGregor, “My God is bigger than your God because my God knows that triangles could have had 4 sides and yours doesn’t.”

Chapter 3: An Intuitionist Defense of Divine Supercomprehension (pp. 38–56)

This chapter is a response to the well-known grounding objection (GO) against Molinism. I explain the GO in a previous post and why Molinism can’t meet it. Basically, the objection is that Molinism has no good explanation for how God could possibly come by His middle knowledge (MK) because there isn’t any plausible source for the information contained in His MK. It can’t come from God’s nature, because it’s stipulated to be “non-natural” or independent of God’s nature. It can’t come from God’s will, because it’s stipulated to be “pre-volitional”, or independent of God’s creative decree. It can’t come from creation, because it’s ontologically prior to whether there is any creation at all since God bases His creative decree on it. And it can’t come from any source independent of God’s nature and will that would render it metaphysically necessary, because MK is stipulated to be contingent. The GO is a big problem for Molinism, and to date Molinist responses to it have (in my judgment) been thoroughly underwhelming, if not pathetically bad.

MacGregor offers six “exit ramps” for the Molinist to escape from the GO but there’s actually only three or four exits because some of them are just alternative versions of each other. Here’s how I describe them:

• Reply #1: Who needs grounding?
• Reply #2: Grounding is trivial
• Reply #3: Supercomprehension or somehow-God-just-knows-these-things
• Reply #4: MacGregor’s new suggestion – God “intuits” His MK

Replies #3 and 4 are variations on the same theme, which is why MacGregor calls his position a defense of supercomprehension, so I’m going to consider them together. For short, I’ll call these three replies no grounds (#1), trivial grounds (#2), and mysterious grounds (#3 and 4), respectively.

Reply 1: No grounds are necessary

MacGregor’s first reply is to challenge the assumption that MK needs to be grounded at all. That MK needs to be grounded, he says, presupposes truthmaker theory (namely, that there is an asymmetric or anti-symmetric explanatory relation of ‘making true’ whereby truths depend on something in reality to make them true) and, moreover, it presupposes truthmaker maximalism (the idea that all truths have truthmakers). But, says MacGregor, truthmaker theory is a “minority position” (p. 38) and MK is plausibly an exception to truthmaker maximalism (p. 39).

This response is woefully inadequate. First, MacGregor is completely wrong to think that the GO “presupposes” truthmaker maximalism (p. 39). At most it presupposes that metaphysically contingent truths have truthmakers, or the even weaker principle that contingent truth supervenes on being (CTSB). Second, in the respects in which it’s relevant to the GO, I don’t think truthmaker theory, or some weaker supervenience principle, is a minority position at all. (Nor should it matter if it were. Most philosophers are atheist/agnostic, but that’s hardly a good reason to reject theism.) The idea that what’s (contingently) true must sync up with reality and that it’s reality that decides (explains) what’s true and what’s not is deeply plausible and is, in fact, little more than a corollary of the correspondence theory of truth. Third, even if there are some plausible exceptions to truthmaker principles—necessary truths and negative existentials are the most often proffered—MacGregor gives no reason to think that MK is a plausible exception. MK conditionals are contingent and make positive claims about what hypothetical individuals would do. Why shouldn’t we think that, if they are true, there needs to be something in reality that explains or grounds their truth? How is reality such that this MK would-conditional is true but it’s corresponding would not-conditional is false? Because the conditionals are contingent, the truth-values of the conditionals are not self-explanatory. All MacGregor does here is reference a paper by Bill Craig, but Craig’s response focuses mainly on Reply #2. So let’s turn to that.

Reply #2: Grounding is trivial.

MacGregor spends only a few sentences on this reply, but he clearly endorses it. He’s basically relying on the aforementioned paper by Bill Craig when he says “The facts or states of affairs revealed by the disquotation principle constitute the grounds for CCFs. Thus the ground of the CCF ‘If I had a stuffed animal in my office and it was a cat, I would name it Guido’ is the fact that if I had a stuffed animal in my office and it was a cat, I would name it Guido.” (p. 39) We can, Craig suggests, provide “grounds” for MK conditionals in the same way: just posit an abstract “fact” that restates the truth in question.

I find it very hard to take this reply seriously. Craig (and MacGregor) seem to think that the disquotation principle (it is true that p iff p) can provide us all sorts of “facts” on the cheap, and then those “facts” can supply us whatever truthmakers we need. Need a truthmaker for “The cat is on the mat”? Step 1: apply disquotation. Observe that it is true that “The cat is on the mat” iff the cat is on the mat. Step 2: call the right-hand side of the iff relation a “fact” and say that that’s what makes “The cat is on the mat” true.

This is seriously bad metaphysics, and a misunderstanding of the disquotation principle to boot. What makes (say) “the cat is on the mat true”, if it is true, is not any abstract “fact” but a real-live cat, a physical mat, and the two being spatially related such that the cat is “on” the mat. That’s what a substantive reply to a grounding question looks like. It tells us how concrete reality has to be structured in order for the claim to be true. It doesn’t offer merely abstract “facts” as truthmakers. Furthermore, disquotation is a strictly metalinguistic principle. It says that we can freely transition between an object language in which p and a metalanguage in which True(p). In other words, p and True(p) are just different ways of talking about the same thing. The principle thus doesn’t “reveal” (MacGregor) or “disclose” (Craig) any substantive “facts” distinct from the truths in question. If we want, we can use the disclosed “fact” as a label to name whatever it is in reality (if anything) that makes “The cat is on the mat” true, but this neither tells us that “The cat is on the mat” is true, nor does it tell us anything about what it is for “The cat is on the mat” to be true—that is, how reality must be configured for it to be true. Offering up such Craig-style “facts” is, therefore, just a trivial and vacuous response to the grounding objection. The challenge of the GO is to show how it is that MK conditionals can be true and knowable by God. All this response says is that there can be such truths if there are abstract “facts” that, as far as we can tell, are nothing more than alternative labels for the MK conditionals in question!

Reply #3 and 4: MK is true because … (insert appeal to mystery) … God!

This section occupies the bulk (all but one-and-a-half pages) of MacGregor’s chapter. It’s basically the section where he says, if you weren’t satisfied by Replies #1 and 2, then here’s a bunch a hand-waving “solutions” to the GO where we (or God) just make stuff up.

The first suggestion appeals to Francisco Suarez’s idea of supercomprehension. According to this idea, God is so well acquainted with the entire collection of individual essences that he can thereby come to know what each individual would freely do in any concrete situation it might find itself in. Obviously this will only work if all of the MK information is packed into those individual essences from the get-go. Otherwise no amount of advanced insight by God will reveal it. But then we run into the problem mentioned above in my discussion of Chapter 2 that if the essences contain this information, then no individual can freely act otherwise. So the information must both be in the individual essences (for God to discover it there) and not in the individual essences (for it not to circumscribe creaturely freedom). The simple appeal to supercomprehension is, therefore, a non-starter.

Ah, but what if we add a few more epicycles? Maybe that will fix the problem! So MacGregor turns to the work of John Laing who holds both that each individual creaturely essence “contains in it all the true counterfactuals regarding how the creature it represents would in fact act” and that “CCFs retain their contingent status because … they receive their truth values from the creaturely wills [??] which pre-exist in the divine mind as ideas” (p. 41). Got that? Me neither. To say that creaturely wills pre-exist as ideas in God’s mind is a clear category error. A will can’t be an idea. And if it exists as an idea in God’s mind, then how it is a creaturely will? As MacGregor relates, Laing further appeals to Suarez’s idea of habitudes or “unique properties of possible creatures whereby God knows their conditionally free decisions” (p. 41). But how is that supposed to help? It seems to be just another way of saying that the individual essences contain all the MK information, in which case, as we have seen, there is no way the individuals could be instantiated and act otherwise.

MacGregor disagrees, however: “[T]his objection fails because the essence, or idea, itself determines the accidental properties (habitudes) rather than the habitudes determining the essence. So long as the causal arrow [??] moves from the essence to the habitudes, … the notion of an essence can be expanded to include such accidental properties” (pp. 42–43). This quote left me stunned when I first read it. Pardon my French, but WTAF! This reply by MacGregor is horribly confused. First, if the essence determines the habitudes, then it effectively includes then in the essence, which brings us right back to the problem of no creaturely freedom. Second, how is there a causal arrow here? This is another category error. Third, to expand the notion of an essence to include accidental properties is just to expand it into something other than an essence. This is incoherent. It’s like expanding the notion of marriage to include singles. MacGregor’s answer then gets even worse. He says that if you’re not persuaded by what he’s just suggested then “one could argue that each individual essence owns but does not include [??] the CCFs appertaining to it” (p. 42). But how can an essence “own” anything? And how is that any different from including? I’m at a complete loss to make sense of this claim and MacGregor provides no explanation.

Finally, we get MacGregor’s own spin on the supercomprehension idea, but it’s really a different idea because according to it “CCFs are not grounded in properties of individual essences” (p. 42). MacGregor’s account runs on for about 15 pages. Here’s the quick run-down: (1) God infallibly believes all and only truths. (2) That some CCFs are true is guaranteed by conditional excluded middle (CEM). (3) God has the faculty of infallibly intuiting which of any contrasting pair of CCFs is true. (4) Through this faculty, God believes one of each pair of contrasting CCFs to be true. (5) God’s beliefs provide truthmakers for those CCFs.

In response, I’m on board with (1) as long as “believes” is taken to range only over what God believes with maximal credence.

(2) is false, however, and MacGregor’s defense of it is weak. The problem is that the connection between antecedent and consequent in a CCF has to be contingent because it’s a counterfactual of creaturely freedom. But then there have to be logically possible cases in which the antecedent is met and the consequent is also met, and also logically possible cases in which the antecedent is met and the consequent is not met. Fully specifying the antecedent, as Craig suggests, doesn’t help. In fact it only makes clear that nothing in the counterfactual itself can decide which of any contrasting pair of CCFs is true and which is false. It follows that, as far as the antecedent clause goes, the consequent might result and the consequent might not result. This falsifies CEM. MacGregor suggests, however, that CEM still works “for any would-counterfactual whose structure is sufficiently clear to guarantee either its truth or the truth of its [would-not] contrary” (p. 46). Well, okay, but that’s of no help to the Molinist as I’ve just explained. The only way a would counterfactual is going to guarantee either that the consequent would occur or would not occur is if the relationship between antecedent and consequent is necessary and not contingent. But then it can’t be a CCF.

(3) is highly implausible. Suppose for the sake of argument that CEM is necessarily true. It still doesn’t follow that either of any pair of contrasting CCFs is true, for it might be that bivalence fails for these propositions. Perhaps both are neither true nor false. This seems much more plausible to me than that one of each pair would be determinately true and the other false, for it would be wholly arbitrary if that were the case. And if it were wholly arbitrary, then I don’t see how God could have a faculty of infallibly intuiting their truth values, for given the nature of the case there is nothing that such a faculty could give God insight into. If, per (4), this faculty yields infallible beliefs on God’s part based on information (the CCFs in question) that is by definition compatible with either CCF in the pair being true, then God’s got to supply additional information, which would amount to, in effect, to God’s deciding which of each pair of CCFs is to be true. But then we’ve just got theistic determinism with extra baggage. This isn’t Molinism any more.

Finally, (5) gets things backwards. Truthmaking requires that the existence of the truthmaker necessitate the truth of the proposition in question and that the existence of the truthmaker be ontologically prior to the truth in question. Given (1), the existence of God’s believing p plausibly necessitates the truth of p, but it isn’t ontologically prior to p. That is, p isn’t true in virtue of God’s believing it to be true. Or, if it is, then that’s because God’s somehow making it so by Himself supplying the information contained in p. But then we’re back to grounding p in God’s nature or will, neither of which is an option when it comes to CCFs. And if God supplies extra information regarding p that the individual essences themselves don’t provide, then that brings us right back to theistic determinism.

In sum, here’s the dilemma for Replies #3 and 4. Either MK information is contained in creaturely essences or it isn’t. If it is, then these can’t be counterfactuals of creaturely freedom. If it isn’t, then either God can’t know these things because they are either false or neither true nor false, or God can know them, but only because He supplies the missing information, in which case they are, in effect, post-volitional and no longer instances of MK.

Concluding Thoughts

I give MacGregor’s first two chapters a passing grade. They fall far short of showing that the Bible teaches Molinism, but they do a reasonably good job of showing how the view can be Biblically motivated, at least if we allow ourselves a few inaccuracies and leaps in logic along the way. As for Chapter 3, it’s a train-wreck from beginning to end. It’s full of confused ideas and half-baked proposals that barely even address, much less rebut, the grounding objection.