Middle Knowledge and Eternal Omniscience

by Max Andrews

The following is a guest blog post by Ryan Hedrich. Ryan is an undergraduate Public Policy major at Georgia Tech with an interest in epistemology and theology, subjects he writes about at his blog unapologetica.blogspot.com.

__________

“If God did not determine [an act], then there must be in the universe a determining force independent of God.”[1]

This argument, stated simply, has been and continues to be the crux upon which the legitimacy of Molinism is hinged. Can the biblical doctrine of an eternally omniscient God be harmonized with a theory that human choices [in particular] are self-determined?

The Molinist claims that God possesses middle knowledge, “…non-determinative knowledge of the realm of creaturely possibility, a foreknowledge of events which depend not on his decree but on the liberty or free choice of the creature.”[2] It is referred to as “middle” knowledge because it is said to logically reside in between – so to speak – God’s knowledge of both possibilities and logical necessities and God’s knowledge of what will be due to His free instantiation of a particular possible world.

In other words, although God’s knowledge is eternal, His “natural” knowledge of that which must and could be can be said to be a precondition for or logically prior to His middle knowledge of what one would freely choose – in the libertarian sense[3] – given individuating conditions of a possible world. In turn, God’s middle knowledge functions as the precondition for or logically prior to God’s “free” knowledge of that which will occur based on His unconditional decree by which He effects the conditions for a particular possible world.

The important idea is that divine middle knowledge as believed by Molinists is with respect to self-determined human choices which would be exercised given the conditions of a particular possible world (including this one). The important question, again, is whether or not an adequate explanation can be provided as to how God could from eternity know what humans would choose in a given possible world by means other than His own determination.

The “grounding objection” to Molinism has consistently been cited as the greatest obstacle to its acceptance. The issue is this: God doesn’t determine what humans freely choose; what possible conditions God could instantiate would not themselves determine what humans would freely choose, though they may limit what could be chosen. It is clear that if one possesses a libertarian free will, nothing extrinsic to him would cause what he would choose in a given possible world.

Hence, on the assumption that humans possess libertarian free will, the question is begged as to how a contingent object of divine middle knowledge – i.e. that person X would choose Y given particular antecedent conditions Z – is certain. An answer was provided by the father of Molinism, who wrote that middle knowledge is:

…the knowledge through which God, before he decides to create a being endowed with free choice, foresees what that being would do on the hypothesis that it should be placed in a particular order of things – this knowledge depends on the fact that the being in its freedom do this or that, and not the other way around.[4]

Human choices are self-determined. Thus, free choices to which God’s middle knowledge corresponds are the grounds for God’s middle knowledge itself. But this has an interesting implication:

…for Molina’s concept to function, the conditions standing prior to the contingent event must be understood as not merely possible, but as having some sort of actuality or quasi-actuality apart from the divine willing – inasmuch as the point is… that God knows what will occur contingently upon certain conditions lying outside of his will: these conditions are not mere possibility nor divinely will actuality, but foreknown conditions, foreknown as actual apart from the decree, at least for the sake of stating the contingency.[5]

That is, if a person’s choices are the result of libertarian free will, God knows what X would choose “on the hypothesis” of Z only if God quasi-instantiates Z such that X is in a position to, after deliberating possible alternatives, choose Y. [That God would quasi-instantiate Z rather than actually instantiate Z follows when it is remembered that God’s middle knowledge is under consideration, not His free knowledge]. Because Molinists believe in a doctrine of libertarian free will, man’s choice is naturally to be considered the means by which it becomes evident what X would choose from the sphere of what it would be possible for X to choose.

Perhaps the reader may think the grounding objection has now been defeated, since an account has been offered as to how God can know the truth value of a counter-factual of creaturely freedom. Parenthetically, it might even be relevant to the possibility of a multiverse, which is (from what I have read) a hot topic on this blog.

However, the explanation comes at an unaffordable price: “…things contingent, till they are determined to come to passe, or not to come to passe, are not knowable that they shall come to passe, nor are knowable that they shall not come to passe.”[6] It is trivial to observe that God’s middle knowledge is not natural knowledge, and one cannot argue that what a man would freely choose in a given possible world could be necessitated by factors external to the exercise of his own will. But on Molinism,

…there are two phases, as it were, of the divine knowing of an event prior to his willing it – namely, that God first knows an entire possible world in an indeterminate way, as containing (possibly!) both an event and its contrary (scientia necessaria), and then knows by scientia media the outcome of the contingency or free choice were he to actualize that world, with the result that God in (or, indeed, temporally subsequent to) his actualization can also introduce other factors into that world order that are consequent on his knowing of the particular outcome.[7]

The problem, then, is that prior to the point in quasi-instantiated Z at which it can be discerned that X would actually choose Y – viz. when X chooses Y – it can legitimately be claimed X could have chosen not-Y, for that is what libertarian free will entails. Arguing that God’s knowledge is predicated upon the outcome of what a person would choose in quasi-instantiated Z is just a roundabout way of saying that the purpose of the quasi-instantiation of Z is that God can observe and thereby learn that X would, in fact, determine to choose Y. This is, of course, incompatible with the biblical doctrine of an eternally omniscient God.

On the other hand, to insist God knew X would choose Y at the point it was possible that X could have chosen not-Y is intuitively untenable and can be quickly demonstrated. “There must be a causal determination that moves any future contingent from the realm of mere possibility into the realm of actuality…”[8] On Molinism, this causal determination is the free choice of X. What possibilities X could choose are eliminated only when X chooses Y given [quasi-]instantiated Z. Only at that point is it certain X would choose Y in Z.

This is why Molina recognized that the way in which God possesses determinate middle knowledge would be by “foresee[ing] what that being would do on the hypothesis that it should be placed in a particular order of things – this knowledge depends on the fact that the being in its freedom do this or that, and not the other way around.” These points are wholly contrary to the suggestion that God could know X would choose Y given Z apart from foreseeing the self-determination of such, as that would mean God’s knowledge is not dependent on anything external to Himself.

To conclude, there can be no middle knowledge, no human indeterminacy, and no eternal priority of contingencies to any facet of God’s knowledge.


[1] Gordon Clark, Predestination, pg. 39.

[2] Richard Muller, Post-Reformation Reformed Dogmatics Volume III, pg. 417.

[3] cf. Luis de Molina, Concordia, Disputation 2.

[4] Luis de Molina, Concordia, Disputation 52.10.

[5] Richard Muller, Post-Reformation Reformed Dogmatics Volume III, pg. 421.

[6] William Twisse, A Discovery of D. Jackson’s Vanity, pg. 338.

[7] Richard Muller, Post-Reformation Reformed Dogmatics Volume III, pg. 430.

[8] Richard Muller, Post-Reformation Reformed Dogmatics Volume III, pg. 424.


6 Responses to “Middle Knowledge and Eternal Omniscience”

  1. “Arguing that God’s knowledge is predicated upon the outcome of what a person would choose in quasi-instantiated Z is just a roundabout way of saying that the purpose of the quasi-instantiation of Z is that God can observe and thereby learn that X would, in fact, determine to choose Y. This is, of course, incompatible with the biblical doctrine of an eternally omniscient God.”

    The above quotation–at any rate, the conclusion of the above quotation–is false. The upshot of the Molinist view is that God’s beliefs depend on the way the world is and not the other way around. But this is just a trivial truth about knowledge at all. Since knowledge is, among other things, a kind of true belief, and since truth depends on the way the world is, it follows of necessity that a person’s knowing something P (whether that person be God or someone else) depends on the way the world is. How else could anyone be said to *know* anything?

    And it doesn’t follow from God’s beliefs depending on the way the world is–as they all would, since all of his beliefs constitute knowledge–that God *learned* anything. Learning requires, among other things, an acquisition of *new* beliefs. But none of God’s beliefs are *new* in any sense of the word ‘new’. Why should we think any of them are?

    And if you think that God ‘learns’ because Z is only quasi-instantiated and so there’s nothing true about Z until it’s fully instantiated, then I’d say you’re basing your argument on the false belief that all truths require truthmakers; that is, all truths require some existing thing (say, a state of affairs) that *makes* some truth true. But that theory (called: Truthmaker theory), while fairly popular, is demonstrably false. For example, God knows that the proposition ‘that there are no hobbits is true’. But what is it that *makes* ‘that there are no hobbits’ true? As far as I can tell there’s nothing that makes it true that there are no hobbits. ‘That there are no hobbits is true’ because there are no hobbits. So, truthmaker–and the grounding objection which is predicated on the truth of Truthmaker theory–is false.

    There’s no reason to give up on God’s having middleknowledge. There’s no reason because God’s beliefs depend on the way the world is. But that’s just a trivial truth; so, no problem.

  2. Hi Roger,

    //The upshot of the Molinist view is that God’s beliefs depend on the way the world is and not the other way around.//

    Well, from what I understand, that isn’t the case. As I noted in the post, God’s knowledge of what “is” would refer to His free knowledge, not His middle knowledge. That is why I said “That God would quasi-instantiate Z rather than actually instantiate Z follows when it is remembered that God’s middle knowledge is under consideration, not His free knowledge.”

    At any rate, you are correct that God’s knowledge is dependent on extrinsic contingencies. However, I do not see any part of your reply which interacts with my last three paragraphs which were written specifically to explain why God’s middle knowledge must be predicated (grounded) on the [quasi?]observation of X making free choice Y in quasi-instantiated Z.

    //How else could anyone be said to *know* anything?//

    The first sentence of the post alludes to such an alternative possibility; namely, God determined what will be. His knowledge is not externally dependent. He knows what He will ultimately effect, whether directly or by means of secondary causes.

    //But that theory (called: Truthmaker theory), while fairly popular, is demonstrably false. For example, God knows that the proposition ‘that there are no hobbits is true’.//

    What does “there are no hobbits” mean? It seems to me there are hobbits… in Tolkien’s book, at least. [I am not trying to be cheeky! The point has serious metaphysical implications.]

    Additionally, regardless of your beliefs on metaphysics, if “all things that ‘exist’ (‘are’) are not-hobbits,” that would ground your claim. If no thing that “is” can be subsumed under the class of “hobbits,” then there are indeed no hobbits.

    Thanks for the reply!

  3. “Well, from what I understand, that isn’t the case. As I noted in the post, God’s knowledge of what “is” would refer to His free knowledge, not His middle knowledge. That is why I said “That God would quasi-instantiate Z rather than actually instantiate Z follows when it is remembered that God’s middle knowledge is under consideration, not His free knowledge.”

    You misunderstand. It’s true, in one sense, that God’s knowledge of what ‘is’ depends on his free knowledge, but that’s only if by ‘is’ you mean ‘actual’. God’s middle-knowledge has to do with what’s not actual; in fact, it has to do with what’s counterfactual. But there is still a sense of ‘is’ that’s true about counterfactuals; namely, what *would’ve* been the case had the actual world gone differently than it, in fact, did.

    Here, I’ll deal with the first of your final three paragraphs:

    “The problem, then, is that prior to the point in quasi-instantiated Z at which it can be discerned that X would actually choose Y – viz. when X chooses Y – it can legitimately be claimed X could have chosen not-Y, for that is what libertarian free will entails. Arguing that God’s knowledge is predicated upon the outcome of what a person would choose in quasi-instantiated Z is just a roundabout way of saying that the purpose of the quasi-instantiation of Z is that God can observe and thereby learn that X would, in fact, determine to choose Y. This is, of course, incompatible with the biblical doctrine of an eternally omniscient God.”

    I’m afraid I don’t quite understand just what you mean by ‘prior to the point in quasi-instantiated Z at which it can be discerned that X would *actually* choose Y…” Aren’t we dealing with counterfactuals, here? And not what’s actual? Unless, do you just mean that X would, if Z *had been* actual, choose Y? But if so, then I don’t understand what you mean by ‘viz. when X chooses Y’. X doesn’t choose Y; that’s why such a scenario is counterfactual. So, there’s no *when* to it. It may be true that X *would* if he was in Z, choose Y; but, in the actual world, he doesn’t (perhaps X isn’t in Z at all, or some other reason besides).

    Perhaps those criticisms are beside the point. You say that “it can legitimately be claimed X could have chosen not-Y, for that is what libertarian free will entails.” First point: this is true only if the libertarian position is committed to the principle of alternative possibilities. It’s not clear, however, that it is. But suppose it is. What follows? Only this fact: if X had been in Z, he could have chosen to Y or refrain from Ying. How is that “the problem” again?

    Perhaps you mean “the problem” is something like this. Suppose there is some time t, in Z, when X chooses to Y. Before t, (at t0, say), X could have Yed or refrained from Ying. And, before t, it’s unclear (because, perhaps, X’s choice is undetermined?) what X will do. But if this is the problem, then I suspect it’s a problem because it confuses what X *will* do with what X *would* do. What X *will* do is a fact about the actual world. What X *would* do is not; it’s a counterfactual claim. So, this can’t be the problem; and I’m still lost with respect to what “the problem” might be said to be.

    In reply to my question “How else could anyone be said to *know* anything?”, you responded as follows:

    “The first sentence of the post alludes to such an alternative possibility; namely, God determined what will be. His knowledge is not externally dependent. He knows what He will ultimately effect, whether directly or by means of secondary causes.”

    But this response won’t do. Suppose it’s true that God determined what will be. What follows about how God would *know* something about that world? Certainly not that his knowledge isn’t externally dependent, as you claim. Knowledge–all knowledge–depends on the way the world is; it depends, among other things, on a belief’s being true. But truth depends on the way the world is, not the other way around. Even if it’s true that God decreed 13.7 billion years ago that I write the sentence “that I write the sentence” just now, it doesn’t follow that his true belief 13.7 billion years ago *that Roger will write the sentence “that I write the sentence” just now* is *why* it’s true that I write the sentence “that I write the sentence” just now. It’s true that I write the sentence “that I write the sentence” just now iff I write the sentence “that I write the sentence” just now. Sure enough, God may have *caused* me to write that particular sentence; but he didn’t cause that proposition to be true. My writing the particular sentence causes the proposition to be true. And this is just a trivial truth about how truth works. So, since all of God’s beliefs are beliefs that constitute knowledge, it follows that all his beliefs depend on the way the world is, even if he determined what the world will be like. So, God’s knowledge *is* externally dependent. And necessarily so, since that’s how knowledge works. Knowledge is an externally dependent phenomenon.

    Lastly, what I mean by the proposition ‘that there are no hobbits’ is precisely that: there are no hobbits. Sure enough, there are fictional characters; but those fictional characters are fictional because they represent possibilia, and not actual entities. There is the actual entity of the fictional character (where ‘the fictional character’ denotes some non-actual possibilia in the mind of, say, a book reader or book writer, as the case may be); but there is no actual fictional-character. There are no actual hobbits, in other words. So it’s just a truth about the actual world ‘that there are no hobbits.’

    Now, you wonder if all ‘non-hobbits’ might make it true that there are no hobbits. But the problem with a notion like that is that ‘that there are no hobbits’ isn’t *about* non-hobbits. If it were, the proposition ‘that there are no hobbits’ would be logically equivalent to ‘that there are non-hobbits’. But then, that’s a necessary truth. So, it’d follow, if you were right, that ‘that there are no hobbits’ is a necessary truth. But that’s obviously false. Why? Well, suppose some possible world where middle earth exists and hobbits drink ale whilst telling stories about a certain ring all while relaxing on the Shire. If that sort of world is possible–and it is–then it’s possible that ‘that hobbits do exist’ is true. And that’s just to say that ‘that there are no hobbits’ is not a necessary truth. So, no. The existence of non-hobbits does not make it true ‘that there are no hobbits’. I suspect there isn’t anything in existence that makes it true ‘that there are no hobbits.’ And that’s because not all truths need truthmakers (which is just to say that Truthmaker theory is false). If Truthmaker theory is false, then the grounding objection fails. QED :-P

  4. //I’m afraid I don’t quite understand just what you mean by ‘prior to the point in quasi-instantiated Z at which it can be discerned that X would *actually* choose Y…” Aren’t we dealing with counterfactuals, here? And not what’s actual?//

    Sure, that’s why I said “quasi-instantiated.” But would you not agree that what you would have chosen would not have been externally necessitated? You would have possibilities. But only Y would have occurred. Why I instead of any of the other possibilities (since you don’t seem to agree with my depiction)?

    //…if X had been in Z, he could have chosen to Y or refrain from Ying. How is that “the problem” again?//

    The problem is footnote 8 and what follows from it, which is what my above question is trying to get at.

    //Perhaps you mean “the problem” is something like this. Suppose there is some time t, in Z, when X chooses to Y. Before t, (at t0, say), X could have Yed or refrained from Ying. And, before t, it’s unclear (because, perhaps, X’s choice is undetermined?) what X will do. But if this is the problem, then I suspect it’s a problem because it confuses what X *will* do with what X *would* do. What X *will* do is a fact about the actual world. What X *would* do is not; it’s a counterfactual claim.//

    But it’s a counter-factual claim which has to do with our choice, unless you disagree (and given your position on truth bearers/makers, that could be the hangup).

    //God may have *caused* me to write that particular sentence; but he didn’t cause that proposition to be true. My writing the particular sentence causes the proposition to be true.//

    Since my writing the sentence in turn depends on God’s eternal decree that you write the sentence, ultimately, I do not see why God’s knowledge has to be dependent upon the world. It will necessarily be the case that you write the sentence. That you will write the sentence does not hinge at all upon the fact you write it.

    This may simply be a question of tensed statements. We are talking about what “is” and “will be,” but if we simply change the statement without altering the meaning – say, “it is true that at time t you will write x” – God could know that on the basis of his decree, since His decree is the sufficient condition for the act to occur. At best, you are using “depends” in a sense equivocal to what I have been using it as.

    //”Lastly, what I mean by the proposition ‘that there are no hobbits’ is precisely that: there are no hobbits. Sure enough, there are fictional characters; but those fictional characters are fictional because they represent possibilia, and not actual entities. There is the actual entity of the fictional character (where ‘the fictional character’ denotes some non-actual possibilia in the mind of, say, a book reader or book writer, as the case may be); but there is no actual fictional-character. There are no actual hobbits, in other words. So it’s just a truth about the actual world ‘that there are no hobbits.’//

    I hope you don’t take this as an evasion, but what does it mean to be actual rather than possibilia? Some realists hold that predicating the word “exist” or “real” to a subject is meaningless because there is no subject to which it cannot be predicated. You are saying “actual” cannot be predicated of hobbits, but I still don’t understand why.

    //Now, you wonder if all ‘non-hobbits’ might make it true that there are no hobbits. But the problem with a notion like that is that ‘that there are no hobbits’ isn’t *about* non-hobbits. If it were, the proposition ‘that there are no hobbits’ would be logically equivalent to ‘that there are non-hobbits’.//

    Excluded middle means there are either hobbits or ~hobbits. If we looked at everything that is “actual” and determined that none of them were hobbits, then hobbits would not be actual. Or, more to the point, if no thing that God decreed be “actual” can be properly identified as a hobbit, there indeed would be no hobbits.

    //suppose some possible world where middle earth exists and hobbits drink ale whilst telling stories about a certain ring all while relaxing on the Shire. If that sort of world is possible–and it is–then it’s possible that ‘that hobbits do exist’ is true.//

    Right. But what I’ve said has nothing to do with necessary truths. It has to do with what God has freely decreed. It has to do with what is “actual,” since that is how you answered my question regarding the meaning of the original proposition, ‘that there are no hobbits is true.’

  5. My first question should read: “Why [Y] instead of any of the other possibilities (since you don’t seem to agree with my depiction)?” Sorry.

  6. Trackbacks

Leave a Reply