Does Calvinism Have Room For Middle Knowledge? A Conversation

By Paul Helm And Terrance L. Tiessen

[Paul Helm is a Teaching Fellow at Regent College, Vancouver. He was Professor of the History and Philosophy of Religion, King’s College, London, from 1993 to 2000. Terrance L. Tiessen is Professor Emeritus of Systematic Theology and Ethics at Providence Theological Seminary, Otterburne, Man., Canada.]

I. Helm: “No”

Terrance L. Tiessen has recently offered a detailed advocacy of why Calvinists should believe in divine middle knowledge while at the same time rejecting Molinism.[1] He believes that such a position has various advantages. In this rejoinder I first examine the cogency of the position that he defends, and then take a look at the supposed benefits of it, and finally consider an objection.

1. Background

It is a great merit of Professor Tiessen’s presentation that as part of it he engages with the history of Reformed theology, notably with the Reformed Orthodox such as Francis Turretin. So it is appropriate that we should briefly remind ourselves of aspects of the position that they adopted, particularly the distinction that they drew (drawing in turn on medieval discussions) between God’s natural knowledge and God’s free knowledge. Turretin, for example, puts the distinction in the following way:

It [viz., God’s knowledge] is commonly distinguished by theologians into the knowledge of single intelligence (or natural and indefinite) and the knowledge of vision (or free and definite). The former is the knowledge of things merely possible and is therefore called indefinite because nothing on either hand is determined concerning them by God. The latter is the knowledge of future things and is called definite because future things are determined by the sure will of God. Hence they mutually differ: (1) in object because the natural knowledge is occupied with possible things, but the free about future things; (2) in foundation because the natural is founded on the omnipotence of God, but the free depends upon his will and decree by which things pass from a state of possibility to a state of futurition; (3) in order because the natural precedes the decree, but the free follows it because it beholds things future; now they are not future except by the decree.[2]

The object of God’s natural knowledge

is both himself (who most perfectly knows himself in himself ) and all things extrinsic to him whether possible or future (i.e., as to their various orders and states; as to quantity—great and small; as to quality—good and bad; as to predication— universals and singulars; as to time—past, present and future; as to state—necessary and free or contingent).[3]

On this way of thinking the two sorts of knowledge are exclusive and exhaustive. There is no room for a third category, middle knowledge. “There is nothing in the nature of things which is not possible or future; nor can future conditional things [viz., the contents of the supposed middle knowledge] constitute a third order.”[4]

Though this distinction between natural and free knowledge is a mere distinction of reason (or order, as Turretin put it) and so does not represent a temporal distinction in the eternal mind of God, yet it might not be altogether inaccurate to think of God’s natural knowledge as the knowledge of infinite possible worlds, each world a maximal possible state of affairs,[5] and his free knowledge as the knowledge of that world which through the divine decree is created, and so becomes actual.[6] It would not be inaccurate, that is, if we thought of the divine decree as an act of selection from among the infinite array of possible worlds known naturally by God in virtue of his omnipotence, even though the act of selection is nonetheless timelessly eternal, not an action in time.

2. Professor Tiessen’s Proposal

Besides the natural and the free knowledge of God, which he recognizes, Professor Tiessen proposes to insert a category of divine knowledge in between them, a form of conditional knowledge, and in that sense a form of middle knowledge. Yet he is emphatic that this is not the middle knowledge of Molina or of modern proponents of Molinism such as W. L. Craig or Alvin Plantinga, whose proposal is strongly motivated by the need to make provision for causally indifferent or indeterminately free human choices. Tiessen is resolutely opposed to such an account of freedom, and so the need to safeguard it and provide for it forms no part of his proposal.

As Tiessen explains things, middle knowledge (as he understands it) is God’s knowledge of what possible creatures would do, and this “is significantly or categorically different from his knowledge of what they could do (necessary knowledge) but is logically prior to God’s knowledge of what actual creatures will do” (p. 347; emphasis in the original).[7] Both God’s knowledge of what possible creatures could do, and what they would do, is prior to any decree that God makes as to what they will do.

As I understand it, Professor Tiessen’s proposal amounts to the following: God has knowledge of what A would do in circumstances C.[8] What he would do, given who A is, depends upon A’s dispositions, desires, plans, etc., at that time, what the circumstances are at that time, and how A appreciates these circumstances. This is what A would do. For Tiessen this is distinct, “significantly or categorically different,” from whatAinC could do. What A could do is more abstract. What A would do in C is or may come to be an element of an entire possible world, in fact of more than one possible world. In contemplating the possible world which is to become the actual world God may insert the segment, what A would naturally do in C, as part of that world, along with his own activity in intervening in the life of A in that world (if such activity is judged to be wise). Because of the compatibilistic character of A’s free action, God would know how A would behave if his circumstances were to be affected by such divine activity. Tiessen is not very specific about the character of this intervention, but it presumably could include miracles, and other supernatural influences of various kinds. So what A would do in C is a relatively independent subset of a possible world, and in becoming part of the possible world that God decrees, it may or may not be modified, depending upon the wisdom of God and so forth.

God is then able to decide whether he will simply permit people (or angels) to do what they would naturally do in those circumstances, or whether he needs to introduce activity of his own Spirit, either by way of dissuasion . . . or by promptings of his own. It is only in God’s decree that he decides which of the huge number of counterfactuals will be actualized, and how they will come to be, whether they are actively permitted (as in the case of evil) or brought about through more specific divine activity (as in the case of all the good which redounds to God’s glory). (p. 352)

Segments such as what A would do in C have a purely naturalistic character, a state of affairs whose accurate description contains no reference to God’s activity, a character which God is appraised of, and which may then, by the divine decree, form part of the actual world. Tiessen repeatedly distinguishes between such counterfactuals (understood as actualizable conditionals) and the possible worlds in which they occur (e.g., p. 363).

So the first feature of his account of Calvinist middle knowledge is that such knowledge is composed of logically prior elements, such as what A would do in C, together with how God would, if at all, modify this element in incorporating it into the world he will create, in the light of his wise purposes for A and for others.

The second feature of the status of this middle knowledge is that its “moment” or “moments” are temporal moments, as the following passage makes clear.

Thus, in my middle knowledge Calvinist model, before God decides what he will do by way of creating a world and ordering its history, God knows how particular creatures would act if they found themselves in particular sets of circumstances. This knowledge enables God to choose a world whose history is exactly the history that he wisely purposes, but to bring about that history through a combination of the morally responsible actions of rational creatures and God’s own actions. (p. 352; emphasis in the original)

So the components of divine middle knowledge are features of the divine mind involved in a temporal process of deciding which world to actualize. As Tiessen puts it, there is a further stage, beyond God’s natural knowledge, in which God

would not simply be intuiting, he would be deliberating, analyzing, calculating “what if “scenarios that include both the creatures’ actions and his own. . . . I judge that this deliberative process of assessment, and of playing out scenarios, moves beyond what God knows simply because he knows himself and all that he could do consistently with his own nature. This deliberative process makes use of his natural knowledge, but it moves beyond it, and yet is prior to his decision to choose a particular world. (p. 355)

This is an avowedly temporal process, a “deliberative process,” involving the activities of analyzing and calculating. Such deliberation must be a temporal event or stage in the divine mind, not just an intermediate logical moment which we may introduce for the purposes of clarifying our own thought about the order of elements in the divine mind, but a temporal moment. But if the divine life has temporal moments, then in the preparation of the possible world that he will decree God is in time, he is temporal.

So what are the benefits of such an innovative proposal? Those proposed by Tiessen may conveniently be distinguished between benefits for God himself, and benefits for our understanding of God’s relation to evil. As regards God himself Tiessen says,

God’s knowledge of what particular creatures would do in all possible circumstances is immensely useful to God. It enables him to choose the world history best suited to his nature and purposes for the world without having to force his will upon his creatures. In knowing how a particular creature would act in a given set of circumstances, God is able to choose the world in which the combination of the actions of God and his creatures would eventuate in exactly what God desired, but this can be brought about with minimal direct intervention. Thus, God need not force or coerce his creatures in order to have things turn out as he wished. (pp. 347-48)

And the benefits for us?

This has valuable implications for understanding how God is genuinely responsive to human prayers and how God meticulously governs human history in a manner that preserves his absolute sovereignty without taking away the moral responsibility of the human beings who bring about that history. It helps us to understand why God is not morally responsible for the evil acts of his creatures, even though these are all part of his eternal purpose and are, therefore, done according to God’s will. (p. 348)

The idea is that in decreeing what A in C would do, in rendering it actual, God utilizes the choices and actions that A in C would naturally make, make in an unforced or uncoerced way, what he would do. God does this by (in a sense) “finding” A in C. A in C is already “there,” already “present,” present as a possibility, what God would do, part of his middle knowledge.

3. Comments

There is a lot here. I shall confine my response to an attempt to show the unsatisfactoriness of this proposal by making three distinct points, and then by answering a possible objection.

First, I shall probe the distinction (which Tiessen emphasizes, and which is central to his account) between what A in circumstances C could do, and what A in circumstances C would do. Second, I shall query the intelligibility of these purely naturalistic segments. Third, I shall say something about what motivates Professor Tiessen in offering this account.

Could and Would. Tiessen thinks that God’s knowledge of what A in C would do is “significantly or categorically different” from what A in C could do (p. 347). However he nowhere explains, as far as I can see, why there is this significant difference, but simply asserts it. But he most certainly needs to provide an argument for this claim if he is ever going to have a chance of convincing the sceptic. Because it seems obvious that if there is some particular thing that A would do in C, some particular choice that he would make, for example, then it is also the case that A could make that choice. If he would do it, then he could do it, and if he can’t do it then he wouldn’t do it by willing it or wanting it or by bringing it about in some other way. Generally, what a person would do is a subset of what that person could do.

Given Tiessen’s failure to provide a reason to treat would as significantly different from could, the divine knowledge of what A would do in C, however we may suppose that this knowledge is gained, cannot contribute to anything that is significantly different from what the divine knowledge of what A in C could do, however this knowledge is obtained. And the knowledge of what A could do in C is part of God’s natural knowledge.

For on the classic view of God’s knowledge, God by his natural knowledge knows what A in all possible states of his mind could do in all possible sets of circumstances. The contents of all possible worlds containing A include propositions about what A would do if C were to be the case, of if D were to be the case, and so on. These are not so much counterfactual, as pre-factual. In the divine mind the states of mind and sets of circumstances of a possible agent A form elements in innumerable possible worlds, all of the possible worlds in which A exists. But given that he provides us with no argument it is not at all clear on what basis Tiessen claims that there is a significant difference between God knowing what A could do in C and knowing what A would do in C. It is obviously not sufficient for him simply to assert that Calvinist middle knowledge requires this distinction between could and would, for the prior question must be, is there this distinction in the first place?

It is striking that despite the significance of the distinction for his case, Tiessen does not say much about its basis. But perhaps he provides a clue to his thinking in the contrast that he draws between “the knowledge God has of things which are possible by virtue of their consistency with God’s own nature (his natural or necessary knowledge) and his knowledge of what creatures would do in particular circumstances” (p. 347). Here Tiessen appears to be thinking on the one hand of what A in C could do considered only in the light of God’s own nature, as being somewhat abstract or unspecific, and on the other hand what it is concretely possible for A to do in sets of circumstances such as C, what A would do. This is borne out by the later claim that by his natural knowledge God has the knowledge of logical relations, causal relationships, and so on, that ground “his more particular knowledge in the middle stage” (p. 365; emphasis added).

But there are two things problematic about such a suggestion. One is the problem of what an account of this more abstract relation of A to God’s nature would look like, and the other is whether Tiessen is giving an accurate account of the natural knowledge of God as this has been understood in the tradition. Tiessen also appears to think that his view of God’s natural knowledge is prefigured by something Richard A. Muller says about the orthodox Reformed view, which he quotes. Muller refers to this account of natural knowledge as indefinite, “inasmuch as its objects are possibilities, not actualities” (p. 346 n. 8). But this does not mean that all the objects of natural knowledge are abstract or that they in any respect lack the specificity of the objects of his free knowledge, actualities.[9] Some of them certainly are abstract, including the hosts of necessary truths that God knows. For in addition to knowing all possibilities, God by his natural knowledge knows all necessities, propositions that are true across every different possible world he knows. In this sense the necessary truths God knows may be said to be more abstract than the possibilities God knows. But Professor Tiessen does not seem to have these in mind here.

It is characteristic of the account of God’s natural knowledge, not that it concerns (merely) all possible beings, such as A, but that God by his natural knowledge knows with full specificity what A (in all possible states of his mind and body) would (or could) do in all possible circumstances. The schedule of the immense range of A’s states of mind and body placed in the immense range of circumstances in which he could be situated would be doubly immense, but still relatively small in comparison to the remaining sets of possibilities in which these immensities could (or would) in turn be placed as they form the elements of possible worlds.

It is in such terms that it is plausible to understand the traditional account of the natural knowledge of God. Turretin, for example, says, “Natural and free knowledge embrace all knowable things and entities.”[10] God by his natural knowledge knows all knowable possibilities, not merely sets of individuals and sets of circumstances in abstraction. And at one point Tiessen himself says that God “knows everything that could be, he knows all possible worlds” (p. 347). But if God’s natural knowledge includes all possible worlds, then he knows (in complete detail) all the possible worlds in which A exists, with all their differing circumstances from world to world. So it remains doubtful whether there is any distinct category of what would be as against what could be that could (or would) form the basis of a category of divine middle knowledge of a Calvinist kind.

Secondly, as I understand what he says, Tiessen believes that what God knows about what A would do in C, where what A would do in C is considered “naturalistically,” precedes his knowledge of what A would do in C*, where the circumstances are altered by some specific divine intervention(s). As he says, there is first what such as A would do “apart from any intervention on his [viz., God’s] part,” and then there is “what they would do if God changed the situation in some way by actions of his own” (p. 352). But the value of discriminating in this way between A in “natural” circumstances, C, and then as A and his circumstances are immediately affected by the possible intervention of God, in circumstances C*, is once again not at all clear. For on the traditional account of creation, God brings to pass all creatures and their actions, willingly permitting whatever is evil. These causal influences include, in the case of some, the enlightening and regenerating activity of his Spirit.

The Temporalist Implication. It is important to note that divine temporalism is essential to Tiessen’s Calvinist middle knowledge. Often when reflecting on the components of the divine decree theologians distinguish a first and a second aspect, a before and an after, recognizing all the while that these are distinctions of reason that are made in order for us to understand things a little more clearly than otherwise. But as far as Tiessen is concerned the temporalist language must be an essential part of the description of the process by which God analyzes and deliberates, an essential part of the character of his middle knowledge. It is hard to see how any attempt to eliminate such tensed terminology could succeed without resulting in the “middle knowledge” becoming part either of God’s natural knowledge or of his free knowledge.

Tiessen is sensitive to the issue of divine timelessness and temporality but it is not sufficient to parry the objection to say, as he does, “The Reformed tradition already allows for a process or function or activity involved in God choosing, from the immense array of possibilities, what he will bring about (directly and indirectly)” (p. 364). But there is a significant difference between a logical distinction and a temporal process. Further, it is not a question of moving from middle knowledge to free knowledge, but of what goes on within middle knowledge. God’s deliberating and calculating as he draws conclusions as to his middle knowledge entails that his acquisition of middle knowledge entails a temporal process, and hence God must be in time.

As far as I can see there is nothing incoherent in the supposition that God is in time, nevertheless it is a proposal that runs flatly against the orthodox Reformed theological tradition and ought not to be abandoned without very good reason.

But the introduction of a temporalist understanding of the divine life is not the only novelty. Tiessen’s proposal also carries implications for the character of divine natural knowledge, as we have seen, and for the traditional account of divine omniscience. On the traditional account of divine omniscience, God eternally knows all creatures and their actions, and he does this by being immediately and intuitively aware of the contents of his own mind (natural knowledge) and what he has freely decreed to come to pass, his free knowledge. But Tiessen’s proposal requires significant changes in the way in which God knows, and these changes imperil divine omniscience. They represent God’s knowledge of his own mind regarding what world he decrees to create as a result of a process, and as a process that involves deliberation. The process involves coming to certain conclusions on the basis of that deliberation. So there are periods in the deliberative process, during the deliberation, when God does not know what to do for the best until he has concluded the deliberation. This is clearly inconsistent with the traditional account of omniscience.[11] Professor Tiessen says that the outcome of the deliberation adds nothing to God’s knowledge in any way that could be identified as an improvement (p. 365). But improvement or not, it does involve God in learning, in gaining knowledge that he earlier lacked.

Tiessen’s second novelty is based upon the previously noticed misunderstanding of the idea of God’s natural knowledge which, because it contains the knowledge of all possible worlds, involves all those possible worlds in which God would intervene in one way or another. Given such orthodox accounts of the natural knowledge of God, he does not have to contemplate, at some intermediate, middle stage, how his interventions might work out; he immediately and intuitively knows how they would work out, as part of his natural knowledge of all possibilities, which include possible worlds in which he does not intervene in any way and possible worlds in which he intervenes in some way or another. I shall return to this shortly.

We here witness a familiar phenomenon. The traditional “grammar” of God, largely bequeathed to the Western church by St. Augustine, is a package deal, a highly integrated conceptual scheme. Proposing a modification to an aspect of this scheme inevitably has knock-on effects. We see that the temporalist character of Professor Tiessen’s proposal regarding the character of Calvinist middle knowledge has at least the following domino consequences: it changes the character of God’s natural knowledge, it creates a category of divine knowledge in which God learns, so imperiling divine omniscience, and it places divine immutability in jeopardy.

So we must now ask, given the considerable costs that Professor Tiessen’s proposal would impose upon the Reformed theological tradition, is it worth the benefits that he claims for it (p. 348)?

In trying to answer this question I shall not dwell on the way in which the tradition safeguards the character of God against the charge that being sovereign God is the author of evil, by arguing that evil is a privation, or by invoking divine willing permission, or by utilizing the distinction between primary or secondary causation. Instead I shall focus simply on the alleged benefits of Professor Tiessen’s proposal. These accrue because on this scheme “God need not force or coerce his creatures in order to have things turn out as he wished” (p. 348).

Tiessen writes of God knowing how a particular creature would act in a given set of circumstances, by becoming acquainted with what A would do in C. But in the suggestion that God “finds” this possibility, there is a certain kind of imaginative self-deception, if this is indeed what Tiessen means. God is not first offered a blueprint of what A in C would do, and then adapts it in order to make possible his wise purposes for A.[12] But suppose that he was offered such a blueprint. What would the source of the blueprint be? Where would it originate? One possibility is that its source, whatever exactly it is, is independent of the divine mind. But this suggestion does not look to hold much promise, as it would undermine God’s sovereignty in a big way. The only alternative possibility (as far as I can see) is that the source is A himself. But how could it be that A is the source of what A in C would do? For A does not yet exist, he is only a possible person. How could a possible person be the source of such information? Because, on Tiessen’s assumptions, the freedom of all God’s possible human beings is compatibilist in character, God’s knowledge of how such freedom could or would be exercised is straightforward and uncontroversial. He would know, intuitively and immediately, what A in C would do. Thus God would not need to resort to middle knowledge.

For as part of his natural knowledge God has the idea of A in C as a possibility, along with his knowledge of A possessing innumerable different beliefs and desires in innumerable different sets of circumstances, and (given such knowledge), in his wisdom God creates A in C, creates him down to the last atom and molecule, evil apart, and immediately sustains his life nanosecond by nanosecond, even as, while being sustained by his Creator, A in C perpetrates evil. Such a state of affairs does not necessarily involve divine coercion or compulsion, not in the usual senses of these words, though it will if God in his wisdom decrees to create A as being in some circumstances not responsible for his actions. It is not that God in his wisdom permits possible persons such as A to exist, rather he brings it about that they exist by decreeing that they do, and (in his wisdom) he permits their perpetration of evil.

The absence of coercion is part of what it means for the divine decree and human responsibility to be consistent or compatible. How this happens, how it happens that what someone is decreed to do he may nonetheless be responsible for doing, is somewhat mysterious, as are all points where the divine nature intersects with the creaturely. As we have noted, Professor Tiessen is partly motivated in his account of Calvinist middle knowledge by the thought that it will help us to see further than the tradition has so far taken us in understanding how God is genuinely responsive to petitionary prayer, and how he meticulously governs history in a manner that preserves his absolute sovereignty without taking moral responsibility away (p. 348). It is a natural human instinct to want to have that mystery lessened or alleviated. But we are seeing that the hope of shining more light on these opaque areas by positing Calvinist middle knowledge is illusory.

4. An Objection

The implication of the line of argument that I have been deploying against Tiessen is that the proposed “Calvinist middle knowledge” is in fact a part of divine natural knowledge, and that it is both problematic, and theologically costly, to posit a separate category of middle knowledge. In his article Tiessen considers an argument against this line, one provided by David Werther, and so before ending this discussion I must say something about this objection, and the confusion that it contains.

As we have noted, various points in Tiessen’s article reveal a possible misunderstanding of divine natural knowledge as understood in classical Calvinism. To add to those already noted, in discussing David Werther’s claim that Calvinist middle knowledge is in effect a part of God’s natural knowledge (pp. 354-55), Tiessen asks whether all the worlds that God could possibly create include the proposition that it’s necessarily the case that A in C will do X; is such a proposition necessarily part of God’s natural knowledge? Tiessen’s answer is that he thinks that God’s natural/necessary knowledge of himself includes the principles of causation that underlie such a statement, but that other than this “I propose that God’s contemplation of worlds containing moral agents which he could self-consistently create, and of how the history of each of such worlds would unfold if left to itself, and how it would unfold if he acted personally within the world, indicate that a further stage, beyond God’s necessary knowledge, has been reached” (p. 355), the further stage being the “moment” of Calvinist middle knowledge. The grounds for making such a proposal, other than (once again) the desire to safeguard the Calvinist middle knowledge thesis, are not made clear. The “proposal” looks to be purely stipulative, without any independent merit or rationale. It’s hardly a strength of a proposal that the only reason for holding to it is that without it the position being advocated would be undermined!

Further, Tiessen asserts that to suppose that “necessarily, A in C will do X” as part of God’s natural knowledge, would be to treat what has traditionally been deemed God’s free knowledge as necessary “and so all of God’s knowledge, including his knowledge of the actual future, would be necessary knowledge” (p. 356). So if Werther is correct about middle knowledge then God’s free knowledge is also called into question.

But there is a serious confusion here, the confusion between

(i)	Necessarily (God knows that A in C will do X)
	and
(ii)	Necessarily, A in C does X.

God’s necessary knowledge is the idea that God necessarily has knowledge of all possibilities and of all necessities. But it does not follow from this that all that God knows has the status of necessary truths. The propositions in the mind of God of the form “A in C does X” are not necessary truths, but sets of possibilities which God necessarily knows. Put another way, when one of the possibilities that form God’s natural knowledge is freely decreed, then what is decreed is logically contingent; it might not have been decreed. Of course among the possibilities that God necessarily knows, and may decree, are causally necessary propositions of the form, “If A were to be in circumstances C he would (as a matter of causal necessity) do X.” So it is important to bear in mind the distinction between logical or metaphysical necessity on the one hand, and causal necessity on the other, as well as the distinction between “God necessarily knows all possibilities” and “All possibilities known by God are necessary.”

5. Conclusion

So I judge that Professor Tiessen’s proposal regarding middle knowledge suffers serious defects. As a proposal it is unclear, and the distinction between what A in C could do, and what A in C would do, looks to be a distinction without a difference. The account of the character of God’s middle knowledge commits a Calvinist to divine temporalism of a sort that both imperils God’s natural knowledge as this is classically understood, and surrenders divine omniscience. Finally, the idea that the proposal casts new light on how divine sovereignty meshes with human responsibility is based upon an account of how God learns possible truths that in turn rests upon an illusion.[13]

II. Tiessen: “No, But . . .”

I am grateful to Paul Helm for his very helpful comments on my article in Westminster Theological Journal.[14] He has identified some places where I did not state myself clearly as well as some matters concerning which I now believe it wise to revise my earlier position. At other places, I welcome further discussion because I remain convinced that my proposal has something useful to offer a Reformed theological understanding of God’s providential work.

1. God’s Knowledge Of “Could” And “Would”: Necessary Or Middle?

I had posited that the distinction between God’s necessary knowledge and his middle knowledge is a distinction between his knowledge of “everything that could be,” that is, of “all possible worlds” and his knowledge of “what creatures would do in particular circumstances, which may or may not ever occur, depending on which of the many possible worlds God decides to actualize.”[15] Helm complains that I have not explained why the difference between these is significant and he doubts “whether there is any distinct category of what would be as against what could be that could (or would) form the basis of a category of divine middle knowledge of a Calvinist kind.”

I have defended divine middle knowledge for a few years now, but since at least 2005 I have been pondering the possibility that God knows counterfactuals of soft-determinist freedom as an aspect of his necessary knowledge.[16] When I wrote my article for WTJ in 2007, I still saw value in distinguishing a “middle” knowledge, but even then I stated that “it is more important to me to reach agreement among Calvinists that God makes significant use of his knowledge of counterfactual (or true hypothetical events) than it is to reach agreement about when, logically, God has this knowledge.”[17] Reflection on Professor Helm’s recent comments has finally brought me to the conclusion that he is correct on this point. God’s knowledge of counterfactuals is not different from his knowledge of possibilities; it is therefore part of his necessary knowledge.[18] My WTJ article would better have been entitled: “Why Calvinists Should Affirm God’s Deliberate Use of His Knowledge of Counterfactuals in His Wise Decree, Although They Reject Molinism.”

As I change my mind, I have naturally wondered why I took what I now view as an unhelpful turn. As I reflect upon my earlier error, two contributing factors come to mind. First, it is possible that I blurred the difference between divine and human knowledge. If we wish to predict and to influence another person’s behavior, it is much more useful to know with certainty what that person would do in a particular set of circumstances than to know what they could (or might) do. But one of the great differences between God and us human beings is that, however well we know another person, we never know them well enough to predict with certainty exactly how they will act. They are capable of surprising us. Such is not the case for our omniscient God. He knows people completely, understands circumstances exhaustively and, because God has given moral creatures the freedom of spontaneity (rather than libertarian freedom), he knows exactly what a particular sort of person would do in each hypothetical set of circumstances. Thus, the difference that exists in our human experience, between what we deem it possible for a person to do and what they will do, is a problem caused by imperfect knowledge not a matter of different kinds of knowledge.

Given the difference between God’s knowledge and ours, it is clear that there is an important difference between our knowing what a person could (that is, might) do and God’s knowing what that person would do. I erred seriously, therefore, in proposing that there is a difference between God’s knowledge of what people could and what they would do. This is because God has not made moral creatures libertarianly free; they do not have the power of contrary choice.[19]

I consider God’s knowledge of counterfactuals to be very helpful in our understanding of the compatibility between God’s meticulous sovereignty and morally responsible creaturely freedom. Assuming the validity of the grounding objection to Molinism, if creatures were libertarianly free, the most that even God could know is the probability that a creature would act in a particular way; he could not know with certainty how they would act. Having that certain knowledge of how any moral creature would act in a set of circumstances is what enables God to decree a world history in which his will is always done, even though his creatures act as they choose.

The second factor that probably contributed to my error was the fact that it was from Molinists that I learned to appreciate the usefulness to God of his knowledge of counterfactuals. For some time, I have been acting on the belief that these advantages went with middle knowledge but that the Molinist construction needed to be revised to remove their error concerning the nature of creaturely freedom. I believed that compatibilist middle knowledge was the way to preserve what they had. What I had failed to see, during that time, was that the sole rationale for positing middle knowledge is to give room for libertarian creaturely freedom as a fact of the world God chooses to actualize. I now believe that rejection of the Molinist construction because of its faulty understanding of freedom also entails rejection of the concept of divine middle knowledge. If (as I believe) creatures are not libertarianly free, there can be no difference between God’s knowledge of what creatures could do and what they would do, that is, what “A would do in C,” as Helm puts it, and God knows this important truth necessarily.

Despite the claim in the title of my article, that Calvinists should reject Molinism, it is now obvious to me that I described the situation as Molinists would see it. For them a distinction does exist between what people could and what they would do. This is because they believe that, for creatures to act with genuine (morally responsible) freedom, they must have the power of contrary choice. From their perspective, A could act in more than one way, in circumstances C. What A would do is therefore the product of an act of will on A’s part which is not determined by who A is. It would not suffice for God to know the nature of all possible creatures, he would also have to know their free choices in all possible worlds.

As Helm has rightly discerned, no distinction exists between what A could do and what A would do, in circumstances C, if soft-determinism pertains so that A has the freedom of spontaneity but not of contrary choice. Helm is also correct in his observation that my attempt to differentiate between God’s (necessary) knowledge of possibilities as abstract and God’s (middle) knowledge of counterfactuals as specific is invalid and fails to take into account adequately the traditional Reformed understanding of what God knows necessarily.

2. Effects Of The Rejection Of Middle Knowledge Upon My Model Of Providence

Thankfully, conceding that God knows counterfactuals as part of his necessary knowledge (rather than as a middle knowledge) has virtually no effect on my account of divine providence. What has always been most important to me is that Reformed accounts of God’s providential work in the world take into full account the usefulness to God, in his establishment of the decree, of his knowledge of counterfactuals.

One reason for deeming “middle knowledge” a helpful name for the knowledge of counterfactuals that I suggest God used, in wisely deciding which particular world history he would create, was that this term had already been used by Molinism in arguing this point. I discern that the primary cause of resistance to the Molinist proposal on the part of seventeenth-century Reformed theologians was that it was developed as part of a synergistic soteriology,[20] but I have stated clearly that I consider Molinism incoherent and that I affirm a classically Reformed monergistic soteriology.

Some of the other classic Reformed objections to the concept of divine middle knowledge depend on its Molinist, synergist, formulation. But some of those objections cut more deeply; they express a concern that God’s use of his knowledge of counterfactuals makes him dependent upon the creature.[21] Consequently, if I concede that God’s knowledge of counterfactuals is an aspect of his necessary knowledge but continue to argue that this knowledge is used by God when he decides upon which world he will create, much of my apologetic for what I have previously dubbed “middle knowledge Calvinism” is still necessary. Nevertheless, that apologetic should be easier when the position being defended is disassociated from “middle knowledge” with its Molinist and synergistic association.

Bruce Ware is another Calvinist theologian who has frequently expounded a “compatibilist middle knowledge” understanding. He may continue to affirm divine middle knowledge but he need not do so. This is apparent if one reads his presentation of “compatibilist middle knowledge and divine providence” in his chapter for the four views book that he recently edited.[22] In my opinion, Ware succeeds in identifying biblical evidence that God knows counterfactuals and that he uses them in planning his work in the world. Nevertheless, if one were to replace Ware’s references to “middle knowledge” with “God’s knowledge of counterfactuals,” and to grant that God has this knowledge necessarily, Ware’s fundamental argument would be unaffected.

Richard Muller makes this pertinent remark in his discussion of middle knowledge: “If the issue were simply the divine knowledge of future possibility—even of possibilities arising out of the contingent interaction of finite creatures—it could be easily understood under the rubric of the divine scientia necessaria or necessary knowledge of all possibility.”[23] If the necessary knowledge of God as understood in the Reformed tradition suffices for his wise decision to actualize a particular world, then the traditional Reformed view that God has only two kinds of knowledge is correct. I granted this possibility in my WTJ article,[24] but I now believe it to be the reality.

My response to the traditional Calvinist objections to divine middle knowledge, when their point of concern is God’s deliberative use of the knowledge of “what A would do in C,” is unaffected by a change in my terminology regarding God’s knowledge. I now find myself in the line of theologians (like Gomarus, Walaeus, and Richard Baxter) who (as reported by Muller),

though repelled by the Pelagianizing impact of this [i.e., Molina’s] view, adapted the argument of Molina to refer, not to a scientia media between knowledge of the possible and knowledge of the actual, but to a scientia hypothetica prior to all of the divine determinations. In this view, God rests his decretum upon his knowledge of how the world order is to be constructed in its most minute hypothetical workings. The decree, therefore, establishes the freedom of secondary causes and allows for or permits the eventuality of sin and evil, though only in a hypothetical sense, namely, as events that will occur, given the actuality of the circumstances preceding. The point, in other words, is not that God learns from or reacts to a future possibility, but that God actualizes a particular concatenation of possibilities in which, given the particular set of circumstances directly willed, certain events will occur by reason of secondary causes, including the exercise of human free choice. The free choices belong, therefore, to the particular world order that God wills to actualize. As for God’s “foreknowledge” of all such actual events, it is necessary, certain, and determinate as it follows the decree and rests on the certainty of the divine causality.[25]

This sounds significantly similar to the intentions of my own proposal. What I have spoken about when I conceive of God’s choosing a world to actualize is essentially what Muller describes as the “particular concatenation of possibilities in which, given the particular set of circumstances directly willed, certain events will occur by reason of secondary causes, including the exercise of human free choice.” Similarly, the point most important in my previous construction is well described in Muller’s description of the views of Walaeus, Gomarus, and Baxter: “God rests his decretum upon his knowledge of how the world order is to be constructed in its most minute hypothetical workings.”

Given this strong similarity of intent, it may be that a more appropriate name for my model would be “hypothetical knowledge Calvinism.” That would have the advantage of not immediately attracting the negative response characteristically triggered in the Reformed tradition by reference to divine middle knowledge; it simply draws attention to a truth already affirmed by the tradition, the significance of which had not been sufficiently recognized.

Since the purpose of Molina’s affirmation of divine middle knowledge was to include in God’s deliberation leading up to his decree his knowledge of future counterfactuals (i.e., subjunctive conditionals) of libertarianly free acts, use of that term within a monergistic construction is counter-productive. A term such as “hypothetical knowledge” (or counterfactual knowledge) has the advantage that it focuses on the kind of knowledge God uses in formulating his decree rather than its place in the logical order. If that kind of knowledge is understood to be an aspect of God’s necessary knowledge, then we need only unpack its usefulness to God; we need not postulate that his deliberative use of the knowledge is itself a new kind of knowledge.

3. God’s Relationship To Time

Helm writes: “But as far as Tiessen is concerned the temporalist language must be an essential part of the description of the process by which God analyzes and deliberates, an essential part of the character of his middle knowledge. It is hard to see how any attempt to eliminate such tensed terminology could succeed without resulting in the ‘middle knowledge’ becoming part either of God’s natural knowledge or of his free knowledge.” He is troubled because “the status of this middle knowledge is that its ‘moment’ or ‘moments’ are temporal moments.”

Perhaps my acknowledgment that God knows counterfactuals necessarily and does not, therefore, need middle knowledge, will alleviate Helm’s concern. I am continuing to assert that God analyzes and deliberates, that he considers possible worlds and chooses to actualize one of them, but this is not a new form of knowledge (as middle knowledge would be), it is the wise use of what God knows necessarily. God needs nothing beyond that necessary knowledge to make his decision to create this particular universe, something which God does freely, not out of necessity.

Despite my use of temporal language in speaking about God’s knowledge and decision, I would argue that my understanding of God’s use of his necessary knowledge of counterfactuals in forming his decree does not depend upon essential divine temporality. After all, when we consider God’s decree, we are speaking of God without a world. Even theologians who have argued for God’s temporal immanence in the universe have posited that God was timeless without the universe.[26] I admit that it is very difficult for me to conceive of the dynamic relations between the persons of the Trinity and of a logical order of God’s decrees in completely atemporal ways, or to think that God has no experience of a time before he created the world, or that the Son has no experience of pre-incarnate existence. Nevertheless, I acknowledge Millard Erickson’s wisdom when he pleads agnosticism about whether God experiences succession and sequence within himself, because any assumption that God does would derive from an assumption that God’s existence and experience are of just the same nature as ours and this would be folly.[27]

Professor Helm finds it “ironic that Molinist middle knowledge, . . . whatever else it may be burdened with, is not burdened with a temporalist account of God.” This is a very important observation because it demonstrates that even those who originally speculated that God has middle knowledge, a logical moment in which he contemplates what libertarianly free creatures would do in possible worlds, did not consider that deliberation to require divine temporality. Consequently, now that I have rejected middle knowledge, I see no reason why my description of God’s deliberation upon his necessary knowledge of counterfactuals should pose any problems for complete atemporalists.

Francis Turretin writes: “The decree is ascribed to God not inasmuch as it is the effect of previous deliberation and consultation with reasoning passing from one thing to another (of which he has no need ‘to whose eyes all things are naked and most open,’ Heb. 4:13), but by reason of the certain determination concerning the futurition of things (according to which he does nothing rashly, but designedly, i.e., knowingly and willingly).”[28] I do not know how God’s determining the future “designedly” rather than “rashly” differs essentially from his determining it with “deliberation,” but I do not perceive my proposal to be different in essence from that which Turretin affirms.

Furthermore, the temporality of my language concerning God’s contemplation of possible worlds, when he wisely decides which one he will actualize, could function metaphorically within Turretin’s own thoroughly atemporal concept of God’s eternal decree. He wrote: “Although some decrees may be said to be prior or posterior to others, it does not follow that they are not eternal in themselves because this is not said on the part of God (for so they are one only and a most simple act in God), but with respect to our manner of conception (who, on account of the distinct objects, cannot conceive of the decrees except distinctly by priority and posteriority).”[29]

4. Conclusion

Clearly, Professor Helm’s critique has been helpful to me. It has led me finally to abandon the attempt to incorporate divine middle knowledge into my Calvinist understanding of God’s eternal purposing of the history of the universe, in all its detail. Since I do not share the Molinist desire to make libertarianly free human decisions a matter of God’s knowledge distinct from his knowledge of himself, I have no need to affirm divine middle knowledge. Nevertheless, I continue to believe that God’s knowledge of counterfactuals is useful to him in his wise decree concerning the futurition of everything that happens in the universe God creates and governs for his own glory.

Notes

1. Terrance Tiessen, “Why Calvinists Should Believe in Divine Middle Knowledge, Although They Reject Molinism,” WTJ 69 (2007): 345-66. Professor Tiessen advocated this position in his Providence and Prayer: How Does God Work in the World? (Downers Grove, Ill.: InterVarsity, 2000). For a similar view, see Bruce Ware, God’s Greater Glory: The Exalted God of Scripture and the Christian Faith (Wheaton, Ill.: Crossway, 2004).
2. Francis Turretin, Institutes of Elenctic Theology (3 vols.; ed. James T. Dennison Jr.; trans. George Musgrave Giger; Phillipsburg, N.J.: Presbyterian & Reformed, 1992), 1:212-13.
3. Ibid., 1:207.
4. Ibid., 1:214. For a brief modern account of the traditional view of God’s natural and free knowledge in the context of a discussion of Molinism, see Thomas P. Flint, Divine Providence: The Molinist Account (Ithaca, N.Y.: Cornell University Press, 1998), ch. 2.
5. For a standard modern account of possible worlds, see Alvin Plantinga, The Nature of Necessity (Oxford: Clarendon Press, 1974), ch. 4.
6. Tiessen himself favors this way of representing the distinction: “He [God] knows everything that could be, he knows all possible worlds” (“Why Calvinists Should Believe,” 347).
7. In what follows page numbers in parenthesis marks in the text are references to Professor Tiessen’s article.
8. In what follows, each time I use “what A would (or could) do in C” as my central example, A represents a person, with an array of beliefs, desires, and objectives, and C stands for the circumstances A is in, some features of which at least A is aware of.
9. We noted earlier that Turretin contrasts the indefiniteness (indefinitam) of God’s natural knowledge with the definiteness (definitam) of his free knowledge. But this contrast is not made in terms of the specificity of the knowledge but of its temporal fixedness.
10. Turretin, Institutes, 1:214.
11. It is ironic that in its original form Molinist middle knowledge, from which Tiessen distances his own account, whatever else it may be burdened with, is not burdened with a temporalist account of God.
12. According to the evidence provided by n. 12 on p. 348, Bruce Ware also suffers a very similar deception. Ware asserts, “Because God knows the natures of each person perfectly, he knows how those natures will respond to particular sets of factors presented to them. But in no case does he cause the evil to be done.” But God does not know by learning from these possible beings. No, they are a necessary part of the contents of an omniscient mind. (As Tiessen himself recognizes; see p. 361.)
13. Thanks to Oliver Crisp for his comments on a previous draft.
14. Tiessen, “Why Calvinists Should Believe.”
15. Ibid., 347.
16. In an email message to John Frame on January 12, 2005, after we had discussed this issue in a brief exchange, I wrote: “I have come to see that the concept of MK, as such, is not as essential to my model as it is to the Molinist model because of my rejection of libertarian freedom. . . . I still see God’s knowledge of counterfactuals as important in this construct. But, I am now less sure that it matters whether he knows this as part of his essential knowledge or, distinctively, as part of a logically (and perhaps not completely non-chronologically) separate ‘moment’ or act of knowing.”
17. Tiessen, “Why Calvinists Should Believe,” 346.
18. Of course, this means that I was wrong to reject David Werther’s case for including counterfactuals of compatibilist freedom in God’s necessary knowledge (“Why Calvinists Should Believe,” 354-56) and that I now agree with the perspective I quoted from John Frame (Doctrine of God [Phillipsburg, N.J.: Presbyterian & Reformed, 2002], 502-3; cited in “Why Calvinists Should Believe,” 354 n. 32).
19. I am aware that Helm believes the nature of human freedom to be a matter not revealed to us by Scripture (in Bruce Ware, ed., Perspectives on the Doctrine of God: Four Views [Nashville: Broadman & Holman, 2008], 50), and that he thinks theologians who endorse without qualification a philosophical position on this issue make their theology “hostages to fortune” (Perspectives, 126). This is a serious concern and I do not take it lightly. At this point, however, I am greatly impressed by the explanatory usefulness of the concept of soft-determinist freedom in understanding Scripture’s teaching. It is very clear to me that God is meticulously sovereign and equally clear that certain of God’s creatures (angels and humans) are morally responsible. I find the compatibilist account offered by Jonathan Edwards, and widely appropriated by later Calvinists, very persuasive. If am wrong about this, however, the effect of the error would be much more far reaching in its effects upon my theology than the concession that God does not have middle knowledge. Indeed, I would be left with much less to say of an explanatory nature and more appeal to mystery. My doctrine of divine providence would become apophatic. Furthermore, much of the construction in the other models of providence described in my book Providence and Prayer: How Does God Work in the World? would have to be dismissed as unjustified speculation.
20. Tiessen, “Why Calvinists Should Believe,” 356.
21. Ibid., 356-58.
22. Ware, Perspectives, 109-20.
23. Richard Muller, The Divine Essence and Attributes (vol. 3 of Post-Reformation Reformed Dogmatics: The Rise and Development of Reformed Orthodoxy, ca. 1520 to ca. 1725; Grand Rapids: Baker Academic, 2003), 418.
24. Tiessen, “Why Calvinists Should Believe,” 356.
25. Muller, The Divine Essence and Attributes, 420.
26. William Lane Craig, for instance, puts together three factors, that God exists in time, that time had a beginning, and that God did not have a beginning, and Craig concludes that “God must be causally, but not temporally, prior to the Big Bang. With the creation of the universe, time began, and God entered into time at the moment of creation in virtue of His real relations with the created order. It follows that God must therefore be timeless without the universe and temporal with the universe” (Time and Eternity: Exploring God’s Relationship to Time [Wheaton, Ill.: Crossway, 2001], 233). John Frame posits that “God’s experience of time, as Scripture presents it, is more like the atemporalist model than like the temporalist one” (The Doctrine of God, 557). Yet, in light of God’s temporal omnipresence, Frame concludes: “So God is temporal after all, but not merely temporal. He really exists in time, but he also transcends time in such a way as to exist outside it” (559).
27. Millard Erickson, God the Father Almighty: A Contemporary Exploration of the Divine Attributes (Grand Rapids: Baker, 1998), 276-77.
28. Turretin, Institutes, 1:311 (4.1.3).
29. Ibid., 1:315 (4.2.6).