Does Compatibilism Entail Determinism? A Pragmatic Argument From Purpose in Evil

Here is the paper I presented at the 2014 annual meeting of the Evangelical Philosophical Society in San Diego.

Does Compatibilism Entail Determinism?
A Pragmatic Argument From Purpose in Evil
by Guillaume Bignon

The debate on the nature of free will and divine providence, which has been showcased for centuries in the theological world between so-called ‘Arminians’ and ‘Calvinists’, is often (and appropriately) seen in the philosophical world to be between ‘libertarians’ and ‘compatibilists’.
On the one side, ‘Arminians’ (understood broadly enough to include open theists, simple-foreknowledge Arminians and Molinists), are libertarians. They uphold ‘libertarian free will’ affirming that human free choices are not determined by God’s providential decree. Rather, given the totality of God’s providential dealings at the moment of choice, the free agent remains able to choose either way; he is not determined to pick one option over another. Most libertarians are also ‘incompatibilists’, that is, they think that determinism is incompatible with moral responsibility. It means that as libertarians they believe that determinism is false, and as incompatibilists, they add that if it were true, there would be no moral responsibility: no praise or blame for human choices.
On the other side of the debate, as is required by at least some of the 5 points of Calvinism, and following the Westminster Confession of Faith, Calvinists affirm that God determines everything that comes to pass, including human free choices—the good and the bad. They are determinists, in that sense, but I should point out that the focus of Calvinists is really on human free choices, and not so much on whether quantum physics involves some kind of indeterminism. If a particle were to oscillate indeterministically in an isolated corner of the universe while absolutely all human choices were determined by God’s providence, we would still say that Calvinist determinism obtained. Finally, Calvinists are also called compatibilists to describe their belief that this sort of determinism is compatible with human moral responsibility.
And so it is, that Arminians and Calvinists are mostly accurately pitted to be ‘libertarians versus compatibilists’.
But while the thesis of libertarianism straightforwardly entails indeterminism as is affirmed by Arminians of all three flavours, compatibilism on the other hand does not commit its Calvinist proponents to the truth of Calvinist determinism, only to its compatibility with human moral responsibility. Given this, even if Calvinist philosophers were to successfully establish compatibilism, the question of determinism would theoretically remain. Supposing the compatibilist arguments are successful, God could determine human choices without doing violence to human moral responsibility, but has He? Typically, Calvinists find support for this stronger contention in biblical exegesis, or in the traditional philosophical arguments against indeterminism. The present paper aims to offer an alternative philosophical route, based upon the purposelessness of evil in an indeterministic world. I shall argue that God, as a maximally great being would only permit (or risk) evil with morally sufficient reasons in the form of compensating goods, and that if indeterminism is true, at least some evil results from the misuse of libertarian free will alone, serving no other good purpose. It follows that unless libertarian free will is itself a necessary safeguard of moral responsibility or something very much like it, God would not commit to it with its attached price tag of otherwise purposeless evil.

In practice, this produces a straightforward deductive argument for the Calvinist determinist view that goes as follows:
(1) God only ever permits (or risks) the existence of evil when He has sufficient reasons in the form of compensating goods.
(2) If God leaves human choices undetermined, the world contains (or has been risked to contain) at least some evil whose sole rationale is to avoid determinism.
(3) There is no compensating good that comes from avoiding determinism.
(4) If God leaves human choices undetermined, He permits (or risks) the existence of evil without any compensating good (follows from (2) and (3)).
(5) God does not leave human choices undetermined (follows from (1) and (4)).
(6) Calvinist determinism is true (follows from (5)).

Now, as it stands, I have little doubt as to which premise will be rejected by non-Calvinists: that would most likely be premise (3). Arminians overwhelmingly believe that there is a compensating good that comes from libertarian free will over against determinism, namely the great good of moral responsibility. Since they believe that moral responsibility is incompatible with determinism (which I don’t) and they believe that moral responsibility is a great good (which I do), it follows on their view that a great good comes out of avoiding determinism—namely, moral responsibility—and it very well could compensate for a great deal of otherwise purposeless evil. Therefore, premise (3), if left unsupported, is somewhat of an assumption of compatibilism. Accordingly, one may ask, ‘why assume the most controversial premise? Is it not bordering on circular reasoning?’ Let me make two responses. First, I think very good arguments in fact can be marshalled for the compatibility of moral responsibility with Calvinist determinism. It is the main burden of my PhD dissertation. But secondly, and the reason I shall not discuss these arguments here, is that the purpose of the present argument is not so much to prove Calvinist determinism in one fell swoop, as it is to simply raise the stakes on the question of compatibilism. It is an argument that closes the gap between compatibilism and determinism, so that if an Arminian theologian were ever pushed to admit the more modest thesis that determinism would not hurt moral responsibility, then given my additional modest premises, it would automatically take him all the way to Calvinist determinism. If God could determine human choices thereby securing purpose in evil while preserving human moral responsibility, then He very probably has. It remains a rather bold and valuable conclusion; so with these stakes in mind, let us examine my allegedly ‘modest’ premises.

Premise (1) is hardly controversial even for libertarians. All parties in this debate hold that God, in His perfect goodness, would not permit any evil, which He could otherwise prevent, unless He had overriding reasons—compensating goods—to motivate His permission of it. All libertarians affirm this much, and posit that libertarian free will is one of those great goods in question; one that justifies whatever amount of otherwise purposeless evil will, would, or could, result from its misuse. Accordingly, there isn’t much further I need to add by way of argumentation, and instead the two interesting questions become on the one hand whether libertarian free will is indeed one such compensating good, and on the other hand, how much and what kind of evil results from indeterminism depending on one’s view of divine foreknowledge and providence. These are the questions raised by premises (2) and (3), to which we now turn.

Premise (2) is the grand claim that no matter which one of the three libertarian positions one adopts, it commits one to say that some purposeless evil results from it. The three flavours of libertarianism in view are open theism, simple-foreknowledge Arminianism, and Molinism. Of course, they differ in the sort of divine providence they affirm, but as I will now establish, each one ultimately restricts God’s control of human choices in such a way that some room is made for otherwise purposeless evil.
On open theism, God does not determine the outcome of free choices, nor does He even know ahead of time how those free choices will turn out. Evidently enough, this leaves plenty of room for things to happen outside of God’s control: He might very well try to influence things as best He can, but if free will is libertarian and God lacks even foreknowledge of what humans will do with it, it straightforwardly follows that the outcome of those choices falls outside of His control: He simply cannot guarantee righteous outcomes and good purposes behind all the evil that they shall contain.
On simple-foreknowledge Arminianism, God knows ahead of time what humans will freely do, but as I have argued elsewhere, this mere foreknowledge does not afford God any more control over those choices than He would have had on open theism. The knowledge of what will happen in the future cannot be used in order to control those same future events, because if God foreknows that choice X will be freely made in the future, He cannot use that knowledge to prevent X, or X wouldn’t be the future after all, which is absurd per hypothesis. Now I realise that some advocates of simple-foreknowledge have argued that it does provide some providential advantage over open theism. They argue God can engage in prophecy about the future, or can for example infallibly win a game of rock/paper/scissors against the devil. These claims may or may not hold, but I need not evaluate them here, because even if they are successful, they do nothing to secure God’s control of human free choices, and hence it remains that if God is equipped with simple foreknowledge, He cannot guarantee the outcome of libertarian free choices, and hence cannot secure good purposes behind all of them.
On Molinism—the final libertarian view before us—the picture is a bit brighter. God not only knows what will happen before it happens, but He also knows what would happen in any hypothetical circumstances, before He decides whether or not these circumstances shall come about. This knowledge gives Him a providential opportunity to bring about or refrain to bring about events involving human libertarian free choices, depending on which envisioned scenario He prefers. That is a serious providential advantage, but as Molinists themselves explain, it still involves inevitable limitations on divine control. On this view, free will is still libertarian, and so God still does not determine the outcome of human choices. He can control them to a certain extent by choosing which circumstances He actualizes, knowing what would result if He picked those circumstances, but He does not control the matter of what would follow from any given set of circumstances. He does not determine the truth of those counterfactuals of freedom, which state what humans equipped with libertarian free will would do in those circumstances. Given this, it follows that while God has much more control over the outcome of human free choices than on the previous views, He still cannot securely guarantee good and righteous outcomes behind all of them. At least some of those counterfactuals will fall on the wrong side, wholly outside of God’s control, and prevent Him from actualizing certain good states of affairs, thereby forcing Him to include lesser outcomes, partially spoiled by the presence of otherwise purposeless evil.
In conclusion, we see that no matter which view of foreknowledge and providence Arminians adopt, as soon as they maintain the libertarian view of free will, they must affirm that God cannot securely prevent the occurrence of purposeless evil, that is, states of affairs featuring moral evil, which God would definitely have prevented if only He could have done so without doing violence to the libertarian free will of His creatures, but in fact couldn’t.
Now let me acknowledge (and agree with) the following objection one might raise about premise (2): just because God cannot guarantee that purposeless evil doesn’t come about, it doesn’t follow that purposeless evil does come about. It could be that things turn out right even though they weren’t fully secured by God’s providential control. Let me make three responses.
First, even though it is theoretically possible, it is unbelievably improbable, bordering on virtual impossibility, that every single indeterminist free choice would by chance alone turn out to feature only evils for which God has a purpose, even though He didn’t determine it to be such. As I just explained, the sort of luck that God would need on the Molinist view differs from that which He would need on simple-foreknowledge Arminianism or open theism, but on each of these views, free will is such that at least some and probably many human free choices turn out differently than God would like if only He had His way.
Secondly, you need not take my word for it, as Arminians themselves provide abundant support for this premise every time they criticize Calvinism for allegedly failing to explain why God would determine certain sorts of evil if determinism were true. It concedes that they believe those instances of evil have no good purpose except to make room for libertarian free will. John Sanders writes:
When a two-month-old child contracts a painful, incurable bone cancer that means suffering and death, it is pointless evil. The Holocaust is pointless evil. The accident that caused the death of my brother was a tragedy. God does not have a specific purpose in mind for these occurrences.[1]
This is rather explicit: it is employing libertarian free will not merely as part of a free will defence against the problem of evil, but very much as part of a theodicy. These evils actually occur, outside of God’s control, and hence for no good purpose. In light of their view of providence, this much is or should be affirmed for at least some evil, by all libertarians, which clearly concedes the truth of premise (2).
And thirdly, even if we now suppose that the all-improbable lucky draw still happened, forgetting that it is virtually impossible, if we suppose that by chance alone, the stars aligned to secure good purposes behind all the evil that happens even without God securing this outcome, it still remains that premise (2) is true, because purposeless evil has been risked by God, even though it didn’t obtain. The truth of premise (2) is secured if purposeless evil obtains, or has been risked to obtain. It’s all that is needed for the argument to go through, because if no compensating good can come out of libertarian free will, then even risking purposeless evil would be a useless piece of reckless gambling on the part of God, which of course is unacceptable: if nothing is to be gained, why gamble? So premise (2) stands, and we must affirm that indeterminism entails that some purposeless evil has occurred or has been risked to occur. And of course, as I acknowledged already, it is not truly evil without a purpose, but it is evil whose sole purpose is to avoid determinism—to make room for libertarian free will.
So we can now turn to premise (3), and ask the question, ‘was it all worth it?’

Premise (3) claims that there is no compensating good that comes from avoiding determinism. That is what compatibilists normally claim: you can have all the benefits of free will, that is, meaningful, morally responsible choices, without sacrificing determinism and hence purpose in evil. As I mentioned above, my argument will assume compatibilism in premise (3), rather than establish it, but we still need to ask: is incompatibilism the only reason to affirm indeterminism? In other words, is it the sole benefit of libertarian free will that it allegedly rescues moral responsibility? Or could it be that even if compatibilism were true, other benefits of libertarian free will would remain and justify that God give it to us at the cost of all this otherwise purposeless evil?
There are two possible ways to maintain such a view: one would either have to say that libertarian free will is an intrinsic good, or that it is an extrinsic good. Either it’s good in itself, or it is good because of something else that it permits and maybe secures.
Let’s consider the latter first. Is libertarian free will the necessary ingredient of a great good outside of itself? Since we are assuming compatibilism, the great good of moral responsibility is no longer at stake, so what other greater good could hang on the question of indeterminism? I think the best candidate for such a thing would be the rescuing of God’s righteousness. Libertarians almost universally press this second argument against determinism: if God determines all that comes to pass, that includes evil, and that allegedly makes Him evil for being the so-called ‘author of sin’. Let me make two responses. First, let me just say—albeit too quickly—that this inference is unjustified. The traditional arguments that aim to show by analogy that God would be evil in that case, fail to establish the relevantly analogous features between God and whichever evil manipulator is taken as allegedly analogous. Refuting such arguments in detail happens to be the second burden of my PhD dissertation, but more simply, and for the purpose of the present modest argument, I will just say this: if compatibilism is true—as the present argument openly assumes—then it strongly undercuts this second alleged problem of determinism. That is because the two arguments are related in a way: incompatibilists typically argue that if determinism is true, humans are not morally responsible for their sins, and since they are not, God is. They press for a transfer of moral responsibility from the determined sinner back onto the determining God. But if compatibilism is true, and determined humans can be fully responsible for their evil, then it seems the buck could stop there—though I realize it may not have to (God could be found guilty even if the determined human takes on some of the guilt), but still, there would be less of a reason to seek another culprit in God. A fuller discussion of the relationship between determinism and divine involvement in evil takes us too far outside the scope of this paper and into my doctoral thesis, so for our present purposes, I will just modestly conclude that if libertarianism isn’t necessary for moral responsibility, it is probably not necessary for securing God’s righteousness either.
Another possible benefit of libertarian free will that could be proposed is its alleged necessity to permit ‘significant love relationships’. If I love God because God determined me to, the argument goes, then this love is not ‘genuine’, it is not ‘meaningful’. Libertarian free will could then be a means to the greater end of independent, undetermined love.
We can all agree that if libertarian free will is necessary for significant love relationships, this end could justify much evil on balance. Love is a highly desirable greater good. But the necessity of libertarian free will for such remains doubtful. Independent considerations could probably be brought in by determinists to reject the necessity of libertarian free will for significant love, but I contend that we can even find the tools to undermine the thesis with the mere assumption of compatibilism already made in the present argument. The assumption of compatibilism undermines both the claims that libertarian free will is necessary for moral responsibility and that it is necessary for significant love relationships, because both allegations are premised upon the same concept, namely the necessity of one’s choice being ‘free’, and issuing from their ‘true self’. In this, the love objection is really not very different from that of moral responsibility: love is significant only when it is freely chosen. And so we are back to the freedom condition, which on compatibilism is by definition compatible with determinism. If human free choices that are determined can be free enough that humans are morally responsible, then most convincingly they are free enough that a love that issues from them is significant. The same freedom of the will is required for moral responsibility and significant love. To maintain one without the other would seem arbitrary. From this, it doesn’t follow that libertarian free will is not necessary for significant love, but it does follow that it is no more necessary for significant love than it was for human responsibility and divine righteousness, which is all the present argument needs to assume.
So we must press the question: if determinism still permits full human moral responsibility (and significant love), and doesn’t compromise the righteousness of God—the two most difficult problems of determinism—then what else could be wrong with it? And accordingly, what could be so great about its avoidance by God at the cost of libertarian free will and its attached price tag of some otherwise purposeless evil?
I suppose one could say that yet other problems remain beyond those two on the laundry list of what is wrong with determinism, and have yet to be discovered, but what exactly would those further problems be? Libertarians are free to suggest additional problems in the future and these shall be put on the table for examination, but I will say this for now: given the long history of debates on the matter, and the voluminous literature on free will and determinism from the indeterminist and Arminian perspectives, how likely is it that a moral defect of determinism has yet to be discovered? It seems plausible instead to affirm with premise (3), that if compatibilism is true, there probably remains no moral problem with determinism. In other words, there is no compensating good that comes from avoiding determinism.
But now, what if indeterminism were great, not in virtue of anything else it permits, but in virtue of being an intrinsic good? Maybe God’s decision to give humans libertarian free will isn’t motivated by the goodness of something else, but by the sheer fact that being undetermined is just good, or good in itself.
Given the widespread disagreements that exist among philosophers of all centuries on the question of how one should define ‘intrinsic value’, let alone measure whether something in fact has it, I cannot reasonably hope to establish that indeterminism isn’t in fact intrinsically valuable on any of these controversial accounts. What I can do instead, is simply to leave it to you and your own understanding of intrinsic value, by asking this: is it plausible that indeterminism is good because it just is good? It doesn’t seem to me to be that way. Free will, certainly might be thought to be like that; moral responsibility, maybe as well, but there doesn’t seem to be anything good or bad in itself, about being determined or undetermined apart from considerations of freedom and responsibility. So it seems rather plausible that if an indeterminist will is not valuable in virtue of a greater good outside of itself, no motivation remains for God to give it to us, risking all the otherwise purposeless evil in the world.
What this all means, is that if someone does a sufficiently good job at defending determinism against its known objections, it commands the truth of premise (3).
And then, since (4), (5) and (6) follow from (1), (2) and (3), it now yields the conclusion that determinism is true: if God can determine human choices and preserve moral responsibility, then just as Calvinists have always affirmed, He has.

Of course, for all this paper has said, the entire battle for the truth of compatibilism has yet to be fought, but to assess the value of the present argument, let me suggest in conclusion that its dialectical force is very much like that of the ontological argument for the existence of God.
The ontological argument contends that if it is even possible that God exists, then it follows that God exists. Using a simple formulation of the argument in terms of possible worlds, it goes like this: if it is possible that God exists, then He exists in some possible world; and in virtue of being (by definition) a maximally great being, if God exists in some possible world, then He exists in every possible world—because necessary existence is better than contingent existence. And of course, if He exists in every possible world, then He exists in the actual world. So the ontological argument gives us the fascinating conclusion that if it’s possible that God exists, then God exists.
It is the same sort of interesting conditional that my present argument has aimed to establish: given moral responsibility, if Calvinist determinism is merely possible, then it is true.


[1] John Sanders, The God Who Risks: A Theology of Divine Providence, 2nd ed. (Downers Grove, IL: InterVarsity Press, 2007), 272.


  1. Brilliant!! Thank you, and God bless.


  2. (4) If God leaves human choices undetermined, He permits (or risks) the existence of evil without any compensating good (follows from (2) and (3)).

    Of course this is true for the libertarian who also denies foreknowledge of future contingencies. But wouldn’t the molinist deny risk based upon their claim that God knows such CCFs?


Post a Comment