Terrance Tiessen is an advocate of what, for convenience, I dub Reformed middle knowledge—which is a compatibilist variant of Molinism. His book on Providence & Prayer, which is well worth owning, presents a comparison and contrast between competing positions on divine providence. Terry also has a forthcoming article in the WTJ in defense of Reformed middle knowledge.
I'll quote some passages from his book, and then comment on the excerpts:
Without restating the argument, I believe that the demonstration of the compatibility of foreknowledge and human freedom, even libertarian freedom, which was offered by William Lane Craig (Molinist model) is completely adequate. Foreknowledge is not causative, not would it be an instance of backward causation.1i) It's hard to square this admission with Terry's stated position elsewhere, in which he denies that foreknowledge and libertarian freedom are compatible (e.g. 317).
ii) I also don't see how his conclusion follows from the premise. It's true that foreknowledge doesn't determine the future. So it's not "causative." But foreknowledge implies a determinate future. The relation is logical rather than causal. But even on logical grounds, foreknowledge is incompatible with an indeterminate future—since an indeterminate object of knowledge cannot be an object of knowledge.
What God knows ahead of time (from our perspective) is the future that comes about through the decisions of free human agents (and would be so, even if they were libertarianly free, which I do not believe is the case, though Craig does). If people decided differently, God's knowledge would have been different. This does not entail that the future must be undecided; it affirms rather than the future that actually comes to be is determined by responsible agents at that time, that those actual decisions have truth value, and that God knows their truth value.2This statement is not so much wrong as it is misleading, for it's quite ambiguous, and skates over the deeper issues. For those theological traditions that affirm foreknowledge, the question at issue is not whether God knows the future, but how he knows the future—consistent with other attributes or actions.
It's true that human beings, as secondary agents, have an instrumental role to play in causing certain things to happen. It's also true, in a sense, to say that if people decided differently, God's knowledge would have been different. But this is all equivocal.
From a Reformed standpoint, human beings don't "determine" the future. God determines the future. God's determination determines what human beings will do. This includes their contribution to the future as secondary agents. There is a sense in which human beings effect or eventuate aspects of the future. We are temporal creatures, acting in time and space, so that our past or present actions (or even inactions) impact the future. To some extent, we bring it about. A system of providence involves secondary agents (personal agents) as well as secondary agencies (inanimate forces).
That, however, doesn't mean that we "determine" a future event, any more than an electrical storm "determines" a future event. Lightening may cause a forest fire, but it doesn't "determine" that outcome—not in the way that Terry is using the term.
Now, you might say that it determines the outcome in terms of physical determinism. It's a physical determinate, in the sense of cause and effect (e.g. a chemical reaction). But I don't think that's the kind of determinism that Terry has in mind.
Rather, he is taking the position that the possible future which God determines to occur is, itself, determined by his knowledge of what we would do in one possible world or another. That's not the same thing as physical determinism.
Rather, it's a ways of saying that God's reason for choosing to instantiate possible world A over world B is supplied by his knowledge of what we would do in alternative scenarios.
From a Reformed standpoint, our contribution to the future is determined by God. Yes, we contribute to the eventuation of the future. But that is because God has determined that future (as an end), and has also determined our causal contribution (as a means to that end).
Now, it's consistent for Terry to say that we determine the future since, for Terry, God is opting for possible world A over possible world B in light of what we would do in either situation. And that, in turn, introduces the specter of retrocausation—as if God's knowledge of what we would do is caused by what we would do.
Indeed, this implicitly flirts with the specter of preexistence, as if God is making his choice from a platonic plenum of free-floating possibilities that subsist apart from the nature and being God—like a cosmic mail-order catalogue.
Calvinism doesn't deny that God knows what we would do. But what is the source of his knowledge? In Calvinism, God's knowledge of hypotheticals and counterfactuals is an aspect of his self-knowledge. God knows what we would do, because God knows what he would do with us.
In Reformed theology, hypotheticals are hypothetical decrees. God knows the alternative outcome because he knows what would occur if he decreed the alternative outcome. God not only knows the outcome of the actual decree, but he also knows the outcome of every hypothetical decree.
The correct metaphor wouldn't be a mail order catalogue, but a novel. A novelist knows what his characters will do, and he knows what they would do. He knows that, in part, because he assigns to each character its defining characteristics. It would be out of character for a character to act in a certain way. And it would be out of character because the novelist has vested the character with certain characteristics.
He also knows what each and every character will do in the future because he has written their future for them. And he knows what they would do in some alternative future, because he has imagined alternative endings for his novel.
Do the characters determine the plot? No. Do the characters determine the author's knowledge of the plot? No.
The characters carry out the plot. They are part of the story. They make certain things happen—things that wouldn't happen if that character didn't exist. And the author's knowledge of the story would be different if the story were different.
But who would make the story different? The storyteller or the storybook characters? If the author chose to write a different story, then that would affect his knowledge of the story, but that's an aspect of the author's self-knowledge. He determines which possible storyline will become the actual storyline.
The characters don't write the plot. Rather, they are written into the plot. The novel isn't written from the inside out—like The NeverEnding Story.
Okay, I'm using a metaphor. And it has certain limitations—like any metaphor. But I think it helps to explicate and expose certain inchoate assumptions and intuitions that are feeding into middle knowledge.
Unlike fictional characters, human beings are conscious beings. And we are genuine agents. We cause certain things to happen.
Yet, when we talk about possible persons in possible worlds, we are talking about the equivalent of fictional characters. Possible worlds are fictions—fictions which inhere in the mind of God. It's just that one of those otherwise imaginary scenarios will come true—in the actual world which God chooses to instantiate.
What is a possible person, anyway? Is a possible person something over and above God's idea of a possible person?
A real person is something over and above God's bare idea. In making a real person, God objectifies his idea in space and even. Even in that case, a real person exactly corresponds to the possible person—which he concretely exemplifies.
And in what sense is a possible person God's idea of a possible person? Are there possible persons, in the sense of freestanding possibilities—of which God has an idea? That would be Platonism rather than Christian theism.
Rather, God's idea is constitutive of a possible person. A possible person is a contingent set of properties which God predicates of a common subject. They don't have to go together to form that particular set of properties, which—in turn—constitutes a possible person. God could consociate a different set of properties.
Indeed, God entertains every possible combination. That's a feature of his necessary knowledge. He knows all possibilities and all compossibilities. Hence, God knows a possible person by knowing his own idea of a possible person—and his idea is constitutive rather than derivative.
By contrast, Terry is implicitly—if not intentionally—suggesting the reverse: that the possible person is constitutive of God's idea.
From a Reformed standpoint, God knows a possible person by knowing himself. He knows what he thinks, and he knows what he can do. A possible world is a subset of divine omnipotence. A possible world is a circumlocution for what God could possibly do. And a possible person is a special case of a possible world.
So what makes a possible person possible is divine omnipotence. Possibility isn't ultimately an attribute of an ideal person, but of the God whose omnipotence is what makes anything possible, and whose omniscience is constitutive of the object.
That's a long answer to a brief quotation, but it will expedite the analysis as we proceed.
I, too have become increasingly convinced that God's knowledge of what would happen in hypothetical situations is an essential element in his wise planning and predestining of the future of the world's history.3I don't have any problems with this statement, per se. My problem, rather, is with Terry's assumption that this condition can only be satisfied by middle knowledge. But hypotheticals are possibilities. As such, they are already covered by God's necessary knowledge. God knows what would transpire if he decreed an alternative state of affairs. So this doesn't demand recourse to middle knowledge.
God must know what a person would do in every possible situation, and one cannot know that if the person's decision is ultimately indeterminate, awaiting the apparently arbitrary choice of the free agent when the moment of decision arrives.4I agree, which is why I don't how what he denies on p317 is at all consistent with what he affirms on p315. Here he says one cannot know that if the person's decision is ultimately indeterminate, awaiting the apparently arbitrary choice of the free agent when the moment of decision arrives, but there he says the demonstration of the compatibility of foreknowledge and human freedom, even libertarian freedom, which was offered by William Lane Craig (Molinist model) is completely adequate.
This tension is reinforced by his definition of contracausal freedom (365) and libertarian freedom (366) in the glossary, which converges with his denial on p317, but diverges from his affirmation on p315.
If God simply decided the future in one logical moment without regard to the possible responses of creatures to his own initiatives and the wisest responses that he could make to those creaturely decisions, then any appearance of significance in those human decisions is thoroughly illusory.5Once again, this is, at best, ambiguous.
i) How does a denial of middle knowledge imply that God doesn't take the consequences of each alternative scenario into account when he chooses which possible world to instantiate? God's necessary knowledge will suffice.
ii) Now his necessary knowledge would be inadequate if you think that God's knowledge of counterfactuals is dependent on what the creature would do. But I reject that assumption for reasons I gave in my earlier reply to Terry, where I said (to quote myself):
In Molinism, God's knowledge of what the creature would do is a mediate knowledge, mediated by the creature.iii) Since, moreover, Terry is on record as endorsing compatibilism, that would be the obvious strategy to pursue in anchoring and explicating the significance of human decisions.
God knows what the libertarian agent would do because what the libertarian agent would do is the cause of God's knowledge. God isn't causing the agent to do this, for if he were causing it, the agent would cease to be free in the libertarian sense.
In Calvinism, by contrast, God knows what the agent would do because he knows what the agent would do if he caused the agent to do this or that in some alternative scenario. So, in Calvinism, God's knowledge of hypotheticals or counterfactuals is immediate rather than mediate.
Thus, Calvinism preserves divine aseity and impassibility, whereas scientia media compromises divine aseity and impassibility by making God's knowledge of hypotheticals or counterfactuals contingent on what the creature would do, rather than what God would do with or to the creature.
iv) The problem with Terry's position doesn't lie in the responsiveness of the creature, but in making God react to the creature—as if God must adapt to a realm of autonomous possibilities that present themselves to him or confront him with a preset range of choices from which he is allowed to make his selection.
Yet hypotheticals are not a preexistent, coeternal vending machine, alongside God, as if God inserts the coin, makes his selection from a preset menu, then pushes the button, and out pops the actual world.
1 T. Tiessen, Providence & Prayer (IVP 2000), 315.
2 Ibid. 315-16.
3 Ibid. 316.
4 Ibid. 317.
5 Ibid. 319.