Keathley makes a valiant and attractive attempt in this book to reconcile Calvinism and Arminianism, defending a strong, Calvinistic view of divine sovereignty and election while advocating a libertarian view of human freedom and an Arminian view of faith. The supposed key to the reconciliation is Molinism, the view that God has what is called “middle knowledge.” This refers to God’s apparent knowledge of what creatures with libertarian freedom would do in counterfactual circumstances. According to Keathley, “The Molinist model is the only game in town for anyone who wishes to affirm a high view of God’s sovereignty while holding to a genuine definition of human choice, freedom, and responsibility” (p. 6).

With this in mind, Keathley offers his alternative to the famous Calvinist TULIP acronym, ROSES: Radical depravity, Overcoming grace, Sovereign election, Eternal life, and Singular Redemption. The rest of the book comprises mostly a discussion and defense of these points in contrast to those other views. Again, Keathley’s book offers a worthwhile attempt to accomplish this goal. In the end, however, I’m afraid it fails to deliver.

Because of space limitations and because my own expertise is in philosophy, I will focus my remarks mainly on philosophical and conceptual matters. First, in chapter 1, Keathley makes his biblical case for Molinism. As part of that discussion, Keathley rightly defines contingency as “the notion that something could have been otherwise … something that happens to be true but obviously could have been false” (p. 28). Of course, the term “could” here is notoriously ambiguous. And Keathley appears to exploit this ambiguity to argue for creaturely libertarian freedom in Scripture. For example, Keathley writes, “When Samuel informed Saul that God had rejected him as king, he told him that it could have been otherwise” (p. 29). But neither this nor any of Keathley’s other examples goes to show that Scripture supports libertarian freedom. All that follows from the Samuel/Saul story is that if Saul had acted differently, he would have remained king. But it does not at all follow that Saul could have acted differently! For Keathley to insist otherwise is to beg the question against a compatibilist notion of human freedom.

Keathley insists, “God’s statements about destroying Israel and starting over with Moses were contingently true and describe real possibilities” (p. 33). Yes, but what kind of possibilities? I would argue that they are simply logical possibilities, which is to say that there is a possible world in which God destroys Israel and starts over with Moses. But this possibility says nothing about whether Moses had libertarian freedom to act otherwise or even that God could have done otherwise given his plans and intentions. Keathley appears to think that true contingency and conditionality rule out necessity and determination, but this is not the case. In the statement “If an asteroid hits the earth, all land animals will die,” the consequent is genuinely contingent on the antecedent. Nevertheless, given the antecedent, the consequent is physically necessary.

Second, Keathley rightly points out that the Bible frequently utilizes counterfactual conditionals in describing and predicting human behavior. But he too quickly concludes from this that the agents involved have libertarian freedom. This simply does not follow. A compatibilist can easily grant the truth of these conditionals but insist that this is best explained by these conditionals not being counterfactuals of freedom. Rather, God can know and utilize these counterfactuals in ordering his providence precisely because the agents have only compatibilist freedom.

Third, it seems that Keathley is unaware of the problems that Molinism has when understood (as it usually is) as an account of meticulous providence. Quoting William Lane Craig, Keathley states that via his middle knowledge, “God can plan the world down to the last detail and yet do so without annihilating creaturely freedom” (p. 39). But this is true only if God’s plan for the world develops logically after his cognizance of what counterfactuals are true. I have argued elsewhere (see “Molinism, Meticulous Providence, and Luck,” Philosophia Christi 11:1 [2009]: 156–69) that if God has a plan for how he wants the world to go that is logically antecedent to his cognizance of the truth-values of counterfactuals, then he would be astronomically lucky if there turned out to be a feasible world that matched his plan (because what worlds are feasible is determined by creaturely choices, not God). I call this version of Molinism Scheme A Molinism. For Molinism to be credible, the Molinist has to adopt Scheme B Molinism and believe that God first takes note of what counterfactuals are true, and thus which worlds are feasible, and then comes up with a plan for how he would like things to go in the actual world, given his severely delimited options. But I argue (ibid.) that this hardly deserves to be called a meticulous view of providence. On Scheme B Molinism, the extent of God’s providential control in history may be likened to a man choosing which movie to play on his DVD player where all the playable movies are written, directed, and acted by others.

Fourth, in chapter 2, Keathley objects to the Calvinist version of the idea that God has “two wills.” Among other things, he believes that this view is logically incoherent. But I believe that his objection amounts to a straw man. Why may we not say (roughly following the logic of John Piper, whom Keathley criticizes) that God has a prima facie (all things being equal) desire to save all, but an ultima facie (all things considered) desire to accomplish some other good ends that overrides his prima facie desire to save all and thus necessitates saving only some? There is nothing at all incoherent about this distinction. But Keathley would object that this is inconsistent with God having genuine love for the unsaved. Not so. Suppose that a physician, out of love and compassion, desires to go to a foreign land to combat a deadly disease. But he gets word that the same disease has infected many people in his homeland. All things being equal, he desires to work to save both groups of people. But all things are not equal. All things considered, he concludes that the best course is to remain home and save the people of his homeland. This is a perfectly coherent story in which a person has two desires but one overrides the other. And notice also that in making the choice to stay home, it doesn’t follow that the doctor doesn’t love the folks in the foreign land. Likewise, God can love the non-elect and genuinely desire to save them, but choose not to in order to achieve a putative greater good (e.g., magnifying his own glory).

Moreover, it is not clear that Keathley’s own preferred solution to the problem escapes serious objection. He argues that God does indeed have two wills but that they are his antecedent and consequent wills. The former is God’s genuine will to save all sinners; the latter his will to damn all those who refuse his offer of salvation. Put another way, God desires to save all but only on the condition of their faith and repentance. In response to the objection that this view portrays God as powerless and waiting in regard to the salvation of sinners, Keathley says simply that God is waiting but not powerless. It is not clear what the intended force of this response is supposed to be. Clearly, on Keathley’s view, God lacks to the power to effectively bring a sinner to faith and repentance without violating his freedom. But in that case it is logically possible that God’s offer of salvation would be universally rejected. It’s hard to see how God could be more powerless in the salvation of sinners than that!

Fifth, in chapter 3, Keathley argues that human libertarian freedom follows from God’s freedom. Against my (and John Feinberg’s) contention that even God lacks libertarian freedom, Keathley argues, following Thomas Flint, that this means that God “had to create this particular world. Then this is the only possible world” (p. 71). I deny the entailment. I grant that God, given his plans and purposes for the world, could not have created any other world than this one. But it does not follow that there is no possible world in which God’s plans and purposes are different. Yet even if I’m wrong about this, I fail to see the problem. Orthodox Christianity is committed to the freedom of God especially as exercised in creating the world. However, I see no non-question-begging reason why orthodox Christianity has to be committed to understanding God’s freedom in libertarian terms. On a compatibilist view of freedom, God is free in creating if the world he creates results from his own desires and intentions. Whether he could have created a different world is irrelevant.

Sixth, in response to the charge that libertarianism makes our choices simply a matter of chance or luck, Keathley, following Bob Kane, argues that being undetermined doesn’t mean being uncaused. Perhaps. But it does mean that agents still choose (say) doing x over doing not-x for absolutely no reason! When answering the question why Smith chose to play ball rather than stay home, ultimately the libertarian (in order to avoid determinism) has to say that that is just what Smith chose to do. If pressed as to why he so chose, the answer has to be: for no reason at all! I contend that such freedom, if we have it, undermines rather than supports our moral responsibility.

Seventh, in the chapter on Overcoming Grace, Keathley argues that salvation can be monergistic (and thus all of grace) even though saving faith precedes regeneration and is not an irresistible gift. To understand this position, he presents what he calls the “Ambulatory Model” of grace:

Imagine waking up to find you are being transported by an ambulance to the emergency room. It is clearly evident that your condition requires serious medical help. If you do nothing, you will be delivered to the hospital. However, if for whatever reason you demand to be let out, the driver will comply.… You receive no credit for being taken to the hospital, but you incur the blame for refusing the services of the ambulance. (p. 104)

This model allegedly shows how salvation can be purely monergistic and completely gracious. According to Keathley, a person does not do anything to acquire saving faith; all a person has to do is refrain from acting. But even though you don’t cause saving faith, you do control whether or not you get to saving faith in that you have the power to resist the offer.

I’m afraid this model just won’t do. For one thing, though Keathley insists that this model avoids attaching human merit to saving faith, he does not answer this question: Why does person S not resist the offer of grace, but person S* does? Though it occurs at a different level, the age-old problem that faces Arminianism plagues the Ambulatory Model too—it cannot explain (without appeal to some kind of merit) why one person exercises saving faith and another does not. Keathley’s point that his model requires a person only to refrain from acting has no teeth. It is obvious that refraining in a case like this requires an active choice! “Refraining” here means something like “acquiescing,” and acquiescing is doing! My gaining saving faith, on the Ambulatory Model, clearly requires my cooperation, and cooperation on the part of the sinner is the antithesis of monergism.

Eighth, chapter 5 addresses the topic of election. There are several things worth commenting on here, but I will limit myself to two. First, Keathley presents the Molinist view of election as an unconditional election though reprobation is conditional. He explains,

According to Molinism, our free choice determines how we would respond in any given setting, but God decides the setting in which we actually find ourselves.… In other words, the Molinist paradigm explains how it is possible for there to be a decree of election without a corresponding decree of reprobation.… The Molinist model presents an asymmetric relationship between God and the two classes of people, the elect and the reprobate. (p. 154)

Regardless of its merits for handling the problem of reprobation, the salient question to ask here is how this model provides for an unconditional election. Why is person S elect? Because God (via middle knowledge) saw that he would choose to have faith in the possible world that God made actual. It may be true that on Molinism whether or not I find myself in a world in which I am elect is up to God, but that does not lend itself to my election being unconditional in any relevant sense. Yes, according to Molinism, God chooses which world we live in, but even as William Lane Craig admits, “it is up to us whether we are predestined in the world in which we find ourselves” (p. 154). My election is conditioned on my choice. Moreover, whether or not I am electable at all depends on what I would do in the various settings God could place me in. Indeed, whether there are any feasible worlds in which I am electable is entirely up to me! It is not plausible to call this model a version of unconditional election.

Lastly, Keathley misconstrues the grounding objection to middle knowledge. He says, “Implicit in the grounding objection is the denial that God has the ability to create creatures with libertarian freedom” (pp. 162–63). This must come as a bit of a surprise to open theists who are among those who think the grounding objection is a serious problem for Molinism. Why not turn this around and say that Molinism implicitly restrains God’s sovereignty by saying he can’t create creatures with compatibilist freedom? No, the grounding objection has nothing to do with whether or not God can create libertarianly free creatures. It’s about whether or not counterfactuals have truth-value, and thus about whether or not God can know them. There are reasons to think that counterfactuals lack truth-value, which means that middle knowledge is incoherent (see my “The Grounding Objection to Middle Knowledge Revisited,” Religious Studies 39:1 [2003]: 93–102).

Salvation and Sovereignty provides a thought-provoking and challenging discussion of soteriology. It is definitely a worthwhile read. But for the reasons stated (and many more that I’ve had to pass over), it does not provide the promised reconciliation between Calvinism and Arminianism.