William Lane Craig--molinism And Open Theism

William Lane Craig—Molinism and Open Theism INTRO The follow are informal thoughts I’m working on in response to comments William Lane Craig makes regarding open theism here (http://www.reasonablefaith.org/site/News2?id=6803&page=NewsArticle). Just to set the stage. In 2001 IVP published Divine Foreknowledge: Four Views. Perhaps the best part of the book is the interaction between Greg Boyd (open theist) and Bill Craig (Molinist). In that volume Greg tried to build a bridge between Molinism and open theism by expanding God’s middle knowledge to include “might-counterfactuals” in addition to the “would-counterfactuals” Molinists argue constitute God’s middle knowledge. Open theism, Greg suggested, might be viewed as a kind of Neo-Molinism. Then in 2002 Greg wrote “Neo-Molinism and the Infinite Intelligence of God” (Philosophia Christi, 2 [2003]). You can find this article on Greg’s site (http://www.gregboyd.org/essays/warning-eggheadessays/neomolonism-and-the-infinite-intelligence-of-god/). Greg’s Neo-Molinism argument saw several published reviews, all negative. In my own conversations with Greg over time I know that Greg’s changed his opinion on the validity of such a ‘Neo-Molinist’ strategy. He tried to bridge the gap between Molinism and open theism. Greg kept working on his own articulation of counterfactual semantics and open theism and in 2006 Greg and I co-authored with Alan Rhoda “Open Theism, Omniscience, and the Nature of the Future” (Faith and Philosophy 23 [2006]: 432–459). You can find that article on Alan’s site (http://www.alanrhoda.net/papers/opentheism.pdf). People still continue to respond to Greg’s “Neo-Molinism” argument as if it’s a strategy he’s still pursuing. But that’s not accurate. Anybody who has read our 2006 Faith and Philosophy article would know we don’t view “might and might not” propositions as “counterfactuals.” Moreover, Greg added an explanatory footnote (5) to his Neo-Molinist article available on his site stating, “This essay was written in 2002. Now (in 2008) I would not refer to ‘would’ and ‘might’ counterfactuals, for I now see that this way of speaking presupposes there’s a settled future that is factual which these statements are ‘counter’ to. I would thus rather speak of ‘would-factuals’ and ‘might-possibilities’.” Unfortunately even this qualification requires some explanation. The point we made in 2006 was that ‘might and might not’ propositions are as factual as any other sort of proposition. Or if one is describing the states of affairs themselves, ‘would’ and ‘might/might not’ are both equally grounded in actual states of affairs. So to speak of ‘would’ props as ‘factual’ in a way that ‘might and might not’ props are not (because they posit possibilities) essentially misses the point we argued in 2006. But not a lot of people have taken notice of this added footnote on Greg’s Neo-Molinism paper. In addition, David Hunt (from the 2001 Four Views book) has contributed a chapter on open theism (in a book edited by Bill Craig and Paul Copan) in which he references our 2006 F&P article and Greg’s NeoMolinist strategy, responding to both and making some inaccurate statements. Parts of Hunt’s chapter can be viewed on Amazon: (http://www.amazon.com/Contending-Christianitys-Critics-AnsweringObjectors/dp/0805449361/ref=sr_1_6?ie=UTF8&s=books&qid=1249930621&sr=1-6). If our 2006 F&P article along with Greg’s onsite comments to his Neo-Molinism strategy aren’t enough, Greg’s most recent publication should set the record straight for Craig and Hunt. This most recent article

1

is published by Religious Studies (2009). A copy of it (NOT the final version) can be found at (http://www.ctr4process.org/affiliations/ort/2006/BoydG.pdf). CRAIG’S QUESTION OK, having said all that, on his site (http://www.reasonablefaith.org/site/News2?id=6803&page=NewsArticle) Bill Craig shares and attempts to answer the following question posed by a reader: Question: …I would like to hear your input on one particular aspect of which Boyd calls Neo-Molinism. He introduces the existence of might counterfactuals as well as would counterfactuals. On his website and in his books he says: On a counterfactual square of oppositions, the logical antithesis of the statement “agent x would do y in situation z” is not the statement “agent x would not do y in situation z.” This is a contrary proposition, not a contradictory proposition. The logical antithesis of “agent x would do y in situation z” is rather the statement “agent x might not do y in situation z.” This latter statement also has an eternal truth-value and hence must be known by an omniscient being. The point is that would-counterfactuals do not exhaust the category of counterfactuals: there are also might-counterfactuals. Propositions about both categories of counterfactuals have an eternal truth-value that must be known by God. Hence I see no reason to restrict God's middle knowledge to knowledge of would-counterfactuals, or, what comes to the same thing, to conclude that all might-counterfactuals are false. There is much that can be said in the defense of the open view of God’s omniscience and what Boyd has to say on the matter, but I feel that much of it can be boiled down to this statement above. Do you disagree with his statements about the existence of might-counterfactuals? MY RESPONSES In what follows I’d like to engage Craig’s answer a step at a time. Craig’s comments are taken directly from reasonablefaith.org. Craig: At the ETS meeting Boyd explained his neo-Molinist alternative, which, he said, hasn’t generated many followers. Tom: Greg has essentially abandoned this strategy. He came to see that attempting to define open theism as a kind of “neo-Molinism” was in the end not possible. Craig needs to distinguish between Greg’s older “Neo-Molinism” argument and the more current arguments in Rhoda’s 2006 F&P article. Craig: His claim is that although God lacks middle knowledge of “would” counterfactuals of creaturely freedom such as “If Jones were in circumstances C, he would freely do action A,” God does have middle knowledge of “might” counterfactuals of creaturely freedom like “If Jones were in circumstances C, he might freely do action A.” Tom: Calling “might and might not” propositions “counterfactuals” was part of Greg’s strategy then. Now he doesn’t call “might/might not” propositions “counterfactuals.” The argument from the Square of Opposition, and conjointly true “might and might not” propositions, however, is an argument Alan,

2

Greg and I argued in 2006 and continue to stand by today. Alan has continued to publish articles along this line (cf. his website: www.alanrhoda.net) Craig: Thus, God knows logically prior to His decree to create a world what any person He could create might or might not do in any set of circumstances in which God should place him. Tom: Right. Except these remain “might/might not” props even after God creates. Craig: By exploiting His middle knowledge of such might-counterfactuals, God is able to plan how He would Himself respond to any choice which any person might make in any set of circumstances. Thus, although God gambles in creating a world of free creatures, in that He knows neither how creatures would choose were He to create them, nor how they will choose in the actual world, nevertheless, He is never caught off guard or unprepared for their choices, for He has already decided how He would respond to any action they might take. Tom: Essentially, yes. Craig: Moreover, He is so intelligent that He knows that however creatures might choose He will so respond as to ensure the realization of His ultimate purposes. Tom: If by “ultimate” Craig means the end-state of creation, in eschatological terms, then yes. Craig: Such a move on Boyd’s part appears at first blush to represent a significant step in the direction of Molinism, since there is a significant intuitive difference between what a person might do in a set of circumstances and what he could do. Tom: Note the introduction of “could” propositions. This will become important. Craig: For example, when Adolf Hitler spoke to the Nazi rally at Nuremberg in 1937, he could have broken into an oration in praise of Winston Churchill—but doubtless this is not something that he might have done. Thus, there is an important difference between what a person can do and what he might do in any given set of circumstances. Tom: On the one hand “can” may refer to mere logical possibility. Hitler’s breaking out in praise of Churchill is a logically consistent state of affairs. Imagining Hitler doing this doesn’t generate a logical contradiction per se. In this sense Hitler “could” break out in praise of Churchill. But more than mere logical consistency contributes to creaturely choice. We also use “can” to describe what lies within the natural range of free choice given all the contributing factors that determine the choice at some particular time, i.e., causally possible futures. That’s the ‘Z’ in “X would do Y in Z.” Z is all the factors that constitute the entire context in which Hitler is the particular person under consideration. This is not straight-forward ‘logical’ possibility. Given all the determining factors at the time (state of mind, depth of character solidification, etc.), Hitler (I’d suggest) “could not” break out in praise of Churchill at that particular time. So let’s keep these distinctions in mind. Craig: If Boyd means to capture this intuitive sense of “might,” then by admitting the truth of mightcounterfactuals logically prior to the divine decree, Boyd seems to have quietly abandoned the most common and forceful objections to the doctrine of middle knowledge, namely, objections based on the lack of grounding of such propositions. 3

Tom: Not at all. The grounding objection isn’t abandoned. Greg does believe the truth of “might/might not” propositions is grounded prior to the divine decree. Grounded in what? In God of course. God is the ground of all that is true prior to creation. What else could ground the truth of what “might/might not” (or of what “will” or “would”) be, or any truth at all? Craig apparently misunderstands our grounding objection to Molinism. The argument does not run as follows: that a) God can ground NO truth prior to the divine decree, that b) all truth must be grounded, and that c) therefore true “would” counterfactuals (having no ground prior to the decree) cannot be true. That’s not our argument at all. What we (Alan, Greg and I) argue is that God MUST be the grounds for whatever is true prior to the decree since (a) all truth is grounded and (b) God is the only available ‘pre-decree’ reality around to serve as grounds. What’s important to note is that God’s being the eternal ground of what we “might/might not” choose is not problematic for creaturely freedom. It’s being eternally true before we determine ourselves that we might/might not determine ourselves in a particular manner is not incompatible with our being libertarianly free. What IS problematic for our behaving freely, however, is God or any actuality other than ourselves being the eternal ground of what we “would” freely do in this or that circumstance (or, more to the point, if we are not the grounds of the truth about what we freely do). Craig: On Boyd's neo-Molinist view there appear to be truths about what persons might or might not do under any set of circumstances, truths which seem to go beyond mere possibilities, which are known by God logically prior to His decree which world should be actual. Tom: Right. But these not problematic in the sense that eternally true “would” counterfactuals are. Craig: But if might-counterfactuals can be true logically prior to God's decree, then why not also wouldcounterfactuals? Tom: Because “would” posits a determined result of choice prior to our own actual self-determining act while “might/might not” only posits the range of possible results. An agent can be libertarianly free though the truth of his possible choices precedes the actual choice. But the matter is entirely different with “would.” What is problematic is that the truth of how we “would” determine ourselves precedes us and is not, consequently, grounded in us. Craig: Boyd doesn't seem to appreciate that on the Molinist view would-counterfactuals logically imply might-counterfactuals, so that both are true and known to God. Tom: Boyd perfectly appreciates the truth that every true “would” (or “will”) entails a corresponding true “might” and every true “would not” entails a corresponding true “might not” in the case of a true “would not.” This in fact is precisely what Alan, Greg and I argue (and Greg more recently in illustrating the Square of Opposition): true “might” stands in a subaltern relation of entailment to a true “would.” Similarly, true “would not” stands in a subaltern relation of entailment to a true “would not.” So if it’s true that “X would do Y in Z” then (Alan, Greg and I agree) it follows that “X might do Y in Z” is true. But let’s press on a bit. What Craig fails to realize is that in the Square both “would” and “would not” are contrary propositions. THIS was the question posed to Craig on his website and which he never actually answered. “Would” and “would not” stand left and right across the top of the Square while “might” and 4

“might not” stand left and right across the bottom of the Square. Now, by definition, the truth of subalterns is only entailed by the truth of their alterns. So “would” only entails “might” when “would” is true, not when it’s false. What Craig fails to appreciate is that contraries “would” and “would not” may be one true and the other false or BOTH FALSE (but not both true). That’s the definition of the contrary relation. Subalterns (“might” and “might not”) on the other hand may be one true and the other false or BOTH TRUE (but not both false). The question then is what sort of future is being described if it’s the case that both “would” and “would not” are false and both “might” and “might not” are conjointly true? The answer leads directly to an open future so long as “might” and “might not” are not limited to meaning mere logical possibility (which they cannot be reduced to). Craig: Boyd is fundamentally mistaken if he imagines that Molinism aims “to restrict God's middle knowledge to knowledge of would-counterfactuals, or, what comes to the same thing, to conclude that all might-counterfactuals are false.” The Molinist maintains that both kinds of counterfactuals are true and known to God via His middle knowledge. Tom: By “might-counterfactuals” Greg means conjointly true might/might not propositions, in which case (and given that “would” and “would not” are contrary props) Greg’s point is that the Molinist who posits true “would” props of creaturely freedom has to agree that whatever the “contradictory” of “would” is, it is false. If Craig wants to say that “would” and “would not” are contradictories (which I understand is his view) then he faces serious problems. What would Craig suggest is the ‘contrary’ of “would”? How would Craig lay out these four props (would, would not, might, might not) on the Square? How might he draw the contrary vs contradictory lines? Craig: It is Boyd, then, who is trying to restrict God's knowledge by limiting it to might-counterfactuals. Tom: No, he only restricts the description (and so knowledge) of future free choices (and other contingencies) to might/might not props. If there are in fact any true “would” props prior to the divine decree, then we argue that (a) God is the ground of their truth (just as he is for true might/might not props), and (b) they do not describe future free acts (but would instead assume in their “if” clauses the acquired character and other contextual factors necessary to entailing their “then” clauses). Craig: But, having abandoned the typical objections to middle knowledge, what justification does Boyd have for this restriction? If he claims that would-counterfactuals are incompatible with creaturely freedom, then he has forgotten the difference between what one could do and what one might do in any set of circumstances. Freedom requires only that in a given set of circumstances one be able to refrain from doing what one would do; it is not required that one might not do what one would do. Tom: Then how does “might not” contradict “would” (as it must given its logical place on the Square)? There’s a bit of equivocating on the meaning of “can/could” in Craig’s argument. Craig: It’s important to understand that on the traditional semantics for counterfactual conditionals, might-counterfactuals are simply defined—contrary to their usage in ordinary language—to be the negations of would not-counterfactuals. Tom: Fine with me. But walk through this slowly. That “might not” contradicts (or “negates”) “would” is all Alan, Greg and I are arguing. If Craig admits this much he’ll face problems in a moment. 5

Craig: That’s what Boyd recognizes when he mentions in the passage you cite, Ephraim, that “On a counterfactual square of opposition, the logical antithesis of the statement, ‘Agent x would do y in situation z’ is not the statement ‘Agent x would not do y in situation z’ . . . [but] ‘Agent x might not do y in situation z’.” [my emphasis] Tom: Does Craig agree that “might not” contradicts “would” and that “might” contradicts “would not”)? If so, then he concedes that “would” and “would not” are contraries and may both be false, in which case both “might” and “might not” would be true. Craig: Boyd seems to confuse the ordinary language use of “might,” which is fraught with connotations of freedom, with its technical meaning in counterfactual semantics, which has no bearing at all on the freedom of the choices made. Tom: Not at all. Let’s stick with Craig’s admission that “might not” contradicts “would” and that “might” contradicts “would not.” That’s all we need. What Craig really wants to argue (I’m guessing) is that in counterfactual semantics “might” is equivalent to “merely logically possible that….” But in this case how can “might” contradict “would not” or “might not” contradict “would”? It’s no contradiction to “X would do Y in Z” to say “It’s logically possible that X not do Y in Z.” To contradict a “would” you have to posit more than just the mere logical possibility of “might not.” So “might” has to mean more than “merely logically possible that….” Craig: (In fact this conflation of what one could do with what one might do is a confusion that runs throughout Boyd's writings.) If Boyd is willing to accept true might-counterfactuals, then I see no reason remaining to deny the truth of would-counterfactuals as well. Tom: Cf. explanations above. It’s clear why Greg, Alan and I would argue “might/might not” props which are true prior to the divine decree are unproblematic for freedom while true “would” props are problematic as descriptions of libertarian choice. Craig: Unfortunately, by denying the truth of any would-counterfactuals, Boyd's view seems to collapse the distinction between “might” and “could.” So long as might-counterfactuals are implied by true would-counterfactuals, we can maintain the distinction between what someone might do and what he could do under any circumstances. It is true that Hitler might not give an oration in praise of Churchill because he would not; but still, he could do so because there are possible worlds (less similar to the actual world than those in which he might not do such a thing) in which he does. Tom: Logically possible worlds, not historically/causally possible worlds. This is where Craig reduces “might” (i.e., ‘possibility’) to logical possibility. But if that’s all Craig will allow “might” to mean, then there’s no falsifying “would not” with “might.” Craig: But if would-counterfactuals are uniformly false, then it is false that he would not do such a thing and so by definition it is true that he might do such a thing. Tom: Would-counterfactuals are not uniformly false. They may be true. But if true, they describe that which follows compatibilistically from the apodosis (the “if” clause). Thus, it may be eternally true that “IF Hitler becomes the sort of person who did in fact stand in Berlin to give the speech in question…THEN he ‘would not’ break out in praise of Churchill.” In that circumstance, THAT particular 6

Hitler “could not” praise Churchill. He’s not libertarianly free to intend such praise at that time given the person he is at that time. Craig: No matter how outlandish an action we pick—like Hitler's delivering his oration while standing on his head—, there will, in the absence of any true counterfactuals to the effect that he would not take such an action, be worlds among those most similar to the actual world in which he takes that action, and so it is true that he might do such a thing. Tom: But central to our thesis is that if “might” is to do the contradictory work expected of it on the counterfactual Square, then it must mean more than “merely logically possible that….” This is the “might” which we posit. Craig and Hunt will beg the question if they insist that “might” does not mean this when philosophers gather to play the logical necessity/logical contingency word games. We’re on an entirely different field of play. “Might” may certainly refer to historical possibility and have implications for freedom. That IS, as he says, the commonsense notion of “might.” As such, it’s more likely to be the biblical notion as well. Craig: Hence, Boyd's apparent move toward Molinism turns out to be a mere feint. Since mightcounterfactuals collapse to counterfactuals about what someone could freely do, God knows nothing more than mere possibilities prior to His creative decree—exactly what traditional open theism claims! [my emphasis] Tom: How could the future of any world that God hasn’t yet committed to create be anything more than merely possible? This doesn’t mean there are no true would-counterfactuals prior to the decree. It can be true prior God choice to create that “if” God creates world W and Hitler becomes such and such a person, “then” Hitler would not break out in praise of Churchill in such and such a circumstance. But even here, Hitler’s not breaking out in praise of Church (well, Hitler very existence for that matter) is (as the “if” states) a possibility. God may create world W and Hitler never come to exist, or Hitler may come to exist but not become a demonic tyrant. But IF he becomes the demonic tyrant he in fact became, THEN he would not be the sort of person capable of breaking out in sincere praise of Churchill. Craig: Openness theology is at a loss to explain this coalescence of human freedom and divine sovereignty. Ironically, Boyd is forced to revert to Calvinistic determinism to account for God’s providence and thus actually winds up destroying human freedom. Tom: Not sure that’s fair or accurate. See the next paragraph… Craig: Boyd likes to compare God to a Grand Master in chess, who is able on the basis of his knowledge of his own prowess and his opponent's weakness to predict exactly when and with what move he will checkmate his opponent. Tom: Greg does not compare God to this. It’s not claimed that Chess masters are always able to “predict exactly when and with what move he will checkmate his opponent.” It’s acknowledged that most of the time the master does not know what exact moves his opponent will make (or even would make). In some rare cases one can box an opponent in and limit his options and on that basis predict future moves. But the chess master analogy is only meant to demonstrate that one can exercise a sovereignty necessary to guaranteeing ultimate victory without knowing all the exact moves an opponent will make. If I play Kasparov (and Kasparov is in good health, not insane or drugged), Kasparov “will” win the game. No might/might not about it (in my view). And this truth is not grounded in any knowledge Kasparov has 7

of what I “would” do in this or that situation. It’s only grounded in his superior grasp of the possibilities. And yet Kasparov’s winning is not jeopardized by this. Craig: The analogy is an engaging one; unfortunately, on Boyd's view God is not so brilliant a chess player as to be able to know that His plans will probably succeed. Tom: On the contrary. Nobody doubts the outcome of Kasparov’s playing me a game of chess. Craig: For He failed to achieve the universal salvation He desired and regretted having created man. Tom: If indeed in the end God fails to achieve universal reconciliation, then universal reconciliation cannot take the place of what we’d call “winning” in chess. The ultimate non-negotiable victory would have to be defined in terms of reconciling and enjoying forever “whosoever” will, in which case God’s ultimate victory does not require universal reconciliation. But this is hardly worse than Craig’s alternative in which ‘who’ gets saved is determined by neither creature nor God but is, as Craig’s admits, simply a metaphysical surd floating eternally alongside God. Craig: Those are not the moves of a Grand Master! So how could He possibly know before the foundations of the world, for example, that His plan for Christ to be crucified through the free agency of Pilate and Herod would be fulfilled? Tom: Do the texts require us to interpret “plan” in terms of specific persons and names and shoe sizes? I think not. Craig: At the ETS meeting Boyd's answer was that God's plan did not include Herod, Pilate, or even crucifixion, but merely the fact that Christ would die for our sins. Why, in the absence of compelling philosophical arguments, compromise biblical teaching in this way? Tom: What in the text is being compromised? ---------------------------FEB 2010 UPDATE * I understand Greg Boyd and Bill Craig and each sharing in a new Four Views on Providence. Should be good! I suspect Greg and Bill will each address these issues and clarify their positions for us. Looking forward to it. *Let me clarify a point regarding whether we affirm the existence of “might-counterfactuals” at all. In my view true “might” props that are true as subalterns of true “would” props are indeed counterfactual, as counterfactual as their “woulds.” But this only regards TRUE “would” props, since for every true “would” there is a corresponding true subaltern “might” (as there is a true “might not” for every true “would not.” But a) these don’t describe libertarianly free choices, and b) when “might” and “might not” are conjointly true (as they may be), both “would” and “would not” are false. THIS “might and might not” prop is not (and cannot be) a counterfactual proposition so far as I can tell. For the causal possibility it describes IS the fact of the matter.

8

JULY 2010 UPDATE Greg’s also addressed these questions in “Two Ancient (and Modern) Motivations For Ascribing Exhaustively Definite Foreknowledge to God: A Historic Overview and Critical Assessment” (Religious Studies 2009) available here: http://www.gregboyd.org/wp-content/uploads/2009/11/twomotivations.pdf. Funny note: It’s supposed to have been a “historical overview” (an overview dealing with the history of the issue) and not a “historic overview” (which means the overview itself is an important and influence event in history). Just thought that was cute!

9