Bnonn Tennant (the B is silent)

Where a recovering ex-atheist skewers things with a sharp two-edged sword

About Uncategorized

What to do when skeptics attack libertarian free will—become a Calvinist

By on

5 minutes to read This is a continuation of the discussion started with Stuart McEwing in his article ‘Openness Theology (Part Two)’, exploring the ramifications of libertarian free will, the principle of alternative possibilities; and how an Arminian theology ultimately collapses into either a Reformed or Open theology, depending on how you push it.

By way of backstory…

This is a continuation of the discussion started with Stuart McEwing in his article ‘Openness Theology (Part 2)’. I realize it’ll go over the heads of some, and I apologize for that—but I think these issues are interesting and important enough to warrant bringing them to the front page. Interesting because, for more philosophically-inclined Christians, they raise questions about our own natures and our relationship to God; important because the answers to these questions have a lot of ramifications for not just our theology, but also our apologetics.

For example, a fairly standard line of attack for skeptics is to draw out the inconsistencies between holding to both God’s definite foreknowledge (DFK) and libertarian free will—which many Christians do. As a skeptic of LFW, though a believer in DFK, I took this line of attack in the comments thread of Stuart’s article:

P = “God knows that an agent S will choose A rather than ¬A”
Q = “S will choose A rather than ¬A”
[A] = the principle of accidental necessity (PAN)
[L] = the principle of logical necessity

  1. [A]P
  2. [L](P → Q)
  3. [A]Q

This precludes the possibility of S’s choosing ¬A. Since LFW typically relies on the principle of alternative possibility (PAP), this argument suffices to disprove the standard libertarian view.

Stuart, however, resolves the difficulty by rejecting the principle of alternate possibility while still holding to libertarian freedom: namely, that our choices are causally unrestrained. To justify rejecting PAP, he cites a hypothetical scenario where it seems that PAP is false, but agent S still has free will. This kind of scenario was first proposed by a philosopher named Harry Frankfurt, and is so called a Frankfurt Counterexample.

At this point, I’m gonna start talking to Stuart directly:

Continuing the discussion…

Stu: I think it’s interesting that you object to PAP using a Frankfurt Counterexample. Frankfurt being a compatibilist and all (: But I take it you’re adopting the Molinist position, ala William Lane Craig.

I think that’s problematic, because ultimately it collapses into a pure Reformed theology. PAP is necessary to liberterian free will (LFW), because without it there’s no obvious distinction between incompatibilism and compatibilism; and without that, there’s no reason to believe in LFW and be a Molinist!

For example, imagine a choice between A and ¬A, where God foreknows the outcome A. Compatibilists, who hold to theological determinism, believe something like the following:

  1. Principle of Volition (PV): Agent S can consciously contemplate A or ¬A and choose one
  2. Principle of Accidental Necessity (PAN): S cannot choose ¬A because his choice of A is accidentally necessary
  3. Principle of Compatibilistic Free Will (CFW): S freely chooses A

But what’s the difference between these beliefs, taken together, and what a libertarian would believe sans PAP? Perhaps you’d say (2) is incomplete, and that completing it creates the relevant distinction:

    2C: S cannot choose ¬A because his choice of A is accidentally necessary AND causally restrained
    2L: S cannot choose ¬A because his choice of A is accidentally necessary though NOT causally restrained

But the difference being suggested here only gains its force by trading on an equivocation in the concept of causality. (2L) cannot be true as a blanket statement under traditional Christianity. And (2C) need not be true, depending on what kind of causation you have in view.

If any kind of causation is in view, then presumably the libertarian and the compatibilist would both agree that (2C) must be true, and together reject (2L)—because the mechanics of God’s creative act necessitate at least three causal restraints on contingent choices:

    CR1. Prior to creation, God surveyed all possible worlds and chose to create this one (call it W1)
    CR2. God initially instantiated W1 in reality by speaking it into being
    CR3. God continually upholds the instantiation of W1 in reality moment to moment

Any Christian must believe all three of these propositions, and all three of them constitute causal restraints on our choices.

A bit of explanation re these three causal restraints Christianity implies

Statement (CR1) entails a causal restraint on our choices, because God’s ability to know true facts about choices in worlds which have not been instantiated logically entails that his knowledge is not grounded on any choices’ actually obtaining. But if his knowledge is not grounded on the choices’ obtaining, yet he still has definite foreknowledge of their outcomes, it follows they must be causally determined. Were they not—were they indeterminate—then by definition he could not know their outcomes.

Statement (CR2) entails a causal restraint on human choices, since S’s choice of A is conditioned on God’s instantiation of W1. Indeed, every choice made in W1 occurs inevitably as God determined when he chose to instantiate W1.

Statement (CR3) entails a causal restraint on human choices, because we know that God alone instantiates things in reality. This instantiative power is a kind of causation, though not a natural causation (aka secondary causation). It’s an existential or primary causation. By definition, only God has this power; it’s sui generis, and a non-communicable attribute. Were God not exercising this power continually, the universe would simply fail to exist. Thus we know that whenever something is real, God alone instantiates it in reality; and since S’s choice to A is real, God alone therefore instantiates it in reality. It’s arguable whether this is merely a restatement of (CR2) or not; I don’t have a considered opinion on that.

The upshot (which is threefold):

Firstly, we must be careful when, in (2C) and (2L) above, we talk about S’s choice being “causally restrained”. Do we mean that it’s restrained in a natural sense, in an existential sense, or both? Any Christian must, of necessity, acknowledge that our choices are existentially causally restrained. But then there is no disagreement between the libertarian and the compatibilist, and their views appear to be the same. On the other hand, if we’re only talking about natural causal restraint, the compatibilist need not (to my knowledge) affirm that our choices are restrained at all; ie, he may agree with the libertarian that the only causally relevant factor in S’s choice is the action of S’s own will.

Secondly, because libertarianism without PAP implies a closed future, and acknowledges God’s definite foreknowledge even of non-instantiated worlds, it therefore necessarily entails theistic determinism:

    TD. Theistic determinism is true just in case for an agent (S) choosing whether A, the outcome A or ¬A is actualized inevitably because of a prior action on the part of God.

Thirdly, libertarianism with PAP necessarily entails the opposite: ie, it implies an open future, which in turn requires a denial of God’s definite foreknowledge, since there is literally nothing for him to know about human choices logically prior to their obtaining.

Make a choice: Calvinism or Open Theism

This is why an Arminian theology will either collapse into a Reformed theology or an Open theology when you push its premises to be consistent with one another. Once you’ve discarded PAP you’re most of the way there, since you’re essentially adopting a compatibilist view already—making theological determinism a lot easier to swallow.

On the other hand, if your intuitions were to refuse to let you discard PAP—as I’ve seen be the case for many Arminians, despite the PAP counterexample God conveniently provided for us right in the Bible itself (Exodus 7ff)—then if you want to align all your beliefs to be consistent you have to let go of God’s definite foreknowledge.

I look forward to your thoughts (:

 7 comments

Jordan Fishel

Dominic, you say “libertarianism with PAP necessarily entails the opposite: ie, it implies an open future, which in turn requires a denial of God’s definite foreknowledge, since there is literally nothing for him to know about human choices logically prior to their obtaining.”

This seems to me a non-sequitur. How exactly did you come to the conclusion that LFW and PAP necessitate a denial of God’s definite foreknowledge? Your reasoning that there is “nothing for [God] to know about human choices logically prior to their obtaining” seems either contrived or unnecessary. Is it really impossible, given LFW and PAP, for God to not know all future events? Is it truly so irrational or contradictory if we define Predestination to include all known human choices? As I understand it, Foreknowing a thing is not the same as Necessitating a thing (i.e., it is not the same as denying PAP).

Moreover, if I am understanding, by “prior to their obtaining,” you are referring to those choices being actualized in the progress of time, correct? If so, then, are we also to understand that your argument is presupposing a particular Theory of Time?

Dominic Bnonn Tennant

Jordan, I explained the reasons in the article. As you say, foreknowing a thing is not the same as necessitating it; but it does make it determinate. And if it is determinate, then it is not indeterminate. Indeterminism is a requirement of libertarianly free choices.

Jordan Fishel

Dominic, it is evident to my mind that there is a problem with your argument. For it seems to indicate that knowing agent S will inevitably choose A inexplicably causes A’s capacity for choosing ¬A to vanish. But it is also evident that I am not familiar enough with the philosophical material to sufficiently explain or rebut. So I asked a good friend of mine, a Prof. of Philosophy and Molinist sympathizer, if he might help me understand your arguments better. With permission, here is the response he gave me:

— begin quote —

There are a couple of problems – one associated with the very thesis he wants to accomplish and one with the manner in which he makes his case. The first one first.

He wants to argue for the conclusion that “when skeptics attack libertarian free will—become a Calvinist.” But the thrust of the article is to complain that such libertarian free will actually DOESN’T exist on pain of being an Open Theist!

On the second point, he makes the same mistake that both philosophical Calvinists and Openness thinkers make: He equivocates either on what it means to be “inevitable” or what it means for some action to be “accidentally necessary.” Usually, Openness thinkers and fatalists opt to call it “temporal necessity” but never mind for the moment. The linchpin of his argument is to argue that the modal

X will necessarily obtain

is grounded in the necessity of CAUSATION. In other words, any event E will necessarily obtain solely because of God’s decree for it to happen. Note that the libertarian free will to cooperate in E’s obtaining is entirely missing! But when this Calvinist sees E’s obtaining, he automatically asserts that it must be because God has so causally made it to happen (hence his confusion in unpacking CR1-CR3 – where in every case it’s causal restraint). In short, the author here neglects the distinction made very early by Plantinga between “weakly actualizing” and “strongly actualizing.” The former alludes to creating the environment/circumstances sufficient for E’s obtaining. The latter is God’s direct causing of E’s obtaining. Notice that in both cases E’s obtaining is “inevitable.” However, only in the second case (Theological Determinism) E’s obtaining is caused exclusively by God. To be inevitable entails either that E will obtain or that E must obtain. The Molinist will affirm the former but deny the latter. But what of the Molinist view here lapses into theological determinism? Nothing! For the Molinist affirms that the agent who WILL actualize event E does not HAVE to actualize event E.

In addition, the original argument for theological determinism self-destructs:

1. [A]P
2. [L](P ? Q)
3. [A]Q

Suppose these variables take on the following content:

1*. Necessarily, Smith truly believes that Jones will buy a guitar.
2*. Necessarily, if Smith truly believes that Jones will buy a guitar, then Jones will buy a guitar.
3*. Necessarily, Jones will buy a guitar.

Thus, as long as Smith truly believes that Jones will buy a guitar (which is easily conceivable and true of many situations that we truly believe in) then Smith’s belief determines Jones’ purchase of the guitar! This argument approach has been unused for millennia precisely because these arguments (originally used by Greek fatalists) made human beliefs suddenly “causally restrain” (or make fated to occur) events. In fact, the pattern can be replaced with all sorts of alternatives that make the argument absurd. Another example:

1**. Necessarily, it is true that I will use the restroom.
2**. Necessarily, if it is true that I will use the restroom, then I will use the restroom.
3**. Therefore, necessarily, I will use the restroom.

Suddenly the mere truth of my using the restroom somehow makes it fated to occur that I will use the restroom!! The truth is, it’s premise 1 that is unintelligible (e.g., the notion of “accidental necessity”). Everyone admits that it’s logically possible for events to be spontaneous. However, they are said to be constrained. But by what? Fate? That’s just a convenient title for an unexplained constrainer! If my actions are free, how are they constrained simply by God knowing about them?

Dominic Bnonn Tennant

Hey Jordan, I appreciate the lengthy critique. But I think there are a few problems:

He wants to argue for the conclusion that “when skeptics attack libertarian free will—become a Calvinist.” But the thrust of the article is to complain that such libertarian free will actually DOESN’T exist on pain of being an Open Theist!

I must confess I don’t see the problem here. If it is true that LFW logically entails open theism, how is that problematic to my thesis, assuming that open theism isn’t a viable option (which I’ve taken as given in the context of the discussion I was having with Stuart)?

He equivocates either on what it means to be “inevitable” or what it means for some action to be “accidentally necessary.”

But to the contrary, one of the purposes of my article was to elaborate on the distinction between logical and accidental necessity, precisely because some people make this equivocation.

The linchpin of his argument is to argue that the modal X will necessarily obtain is grounded in the necessity of CAUSATION. In other words, any event E will necessarily obtain solely because of God’s decree for it to happen.

This is incorrect. In fact, it’s very confused. Firstly, necessity is not the issue here. Determinism is. It’s already been granted that all human decisions occur of accidental necessity; what I’m addressing in the CR statements is the notion that, while this obviates PAP, it still leaves room for libertarian free will.

The CR statements aren’t arguing that any event will obtain causally necessarily because of God’s decree; and neither are they arguing that any event will obtain solely because of God’s decree. I explicitly grant that there’s no reason to make this assumption; I explicitly leave room for secondary causes.

What I’m arguing in the CR statements is that given certain actions of God, the events of the world cannot fail to happen as they do. When God chose to create this world (W), this functioned as a causal restraint on the events within W.

Weak actualization doesn’t solve this problem (as far as I can see), because (CR1) doesn’t require strong actualization to begin with. (CR2) and (CR3) arguably do; but (CR1) simply states that given God’s actualization of W, all events within W are causally constrained to occur as God knew they would. I actualize my decisions inevitably, because of God’s prior action in actualizing the world in which those decisions are actualized:

1. If my decisions were causally unrestrained, I could make decisions that aren’t actualized in W.
2. But I cannot make decisions that aren’t actualized in W.
3. Therefore, my decisions are not causally unrestrained.

That seems sufficient to disprove LFW, but of course there’s always the chance I missed something.

Suppose these variables take on the following content:

1*. Necessarily, Smith truly believes that Jones will buy a guitar.
2*. Necessarily, if Smith truly believes that Jones will buy a guitar, then Jones will buy a guitar.
3*. Necessarily, Jones will buy a guitar.

But this isn’t the argument I gave; that much ought to be obvious if you compare it with the notational form! Several problems present themselves:

A. I explicitly stated that P stands for “God knows” etc. I never gave allowance for the argument working in the case of another human being. The reason for this is less because the argument fails, and more because it becomes absurdly hypothetical. No human being could have this kind of knowledge, unless he were a time traveler.

B. To make the argument clear, the definition of knowledge needs to be much more plainly stated. “Truly believes” ain’t sufficient, because it doesn’t deal with the grounds of the belief. For example, to say that God truly believes something doesn’t assure us that he believes it because it is true (ala Gettier, he could believe it coincidentally; or ala open theism he could believe it probabilistically). That might seem pedantic, but the force of the argument is lost if this isn’t clarified. We’ve gotta be clear that the relationship between God’s belief, and the truth of that belief, isn’t a weak one. God doesn’t merely believe such-and-so truly. He believes such-and-so because it is true.

C. Of course, the biggest problem with your professor’s version of my argument is that it completely obliterates the distinction between logical and accidental necessity! Where I write [L] or [A], he just writes “necessarily”. That kind of equivocation, though, is hardly going to present the argument in a fair light, right?

So lemme plug in the variables and give the argument in plain English, as it should look:

1′. It’s accidentally necessary that God believes-because-it-is-true that Jones will buy a guitar.
2′. It’s logically necessary that, if God believes-because-it-is-true that Jones will buy a guitar, then Jones will buy a guitar (by definition).
3′. Therefore, it’s accidentally necessary that Jones will buy a guitar.

There’s nothing strange about that conclusion, though. Notice, for a start, that (1′) and (2′) are simply truisms. If it is indeed true that Jones will buy a guitar, and if it is true that God has definite foreknowledge, then how could he not know that Jones will buy a guitar? He must know it. Not that it’s logically necessary, but it is accidentally necessary by virtue of Jones’s action.

Furthermore, it’s clear that despite what your professor says, God’s belief doesn’t determine Jones’s purchase. On the contrary, if God knows that Jones will buy the guitar precisely because of Jones’s buying it, then it is Jones’s action that in some sense determines (or at least selects) God’s belief.

Finally, it’s also clear that Jones’s buying the guitar is accidentally necessary. It follows directly from the premises.

In fact, the pattern can be replaced with all sorts of alternatives that make the argument absurd. Another example:

1**. Necessarily, it is true that I will use the restroom.
2**. Necessarily, if it is true that I will use the restroom, then I will use the restroom.
3**. Therefore, necessarily, I will use the restroom.

I’m not sure what your professor is trying to show here. Premise (1**) and (3**) are functionally identical; so all he’s showing is that an argument that starts out assuming the necessary truth of using the restroom can’t prove the necessary truth of using the restroom. (Or that if it’s necessarily true that he’ll use the restroom, then it’s necessarily true that he’ll use the restroom.) Neither of those are interesting statements, though…

With that noted, it seems almost superfluous to observe that the argument doesn’t follow the form of the argument I’ve given—it omits the question of definite foreknowledge. But it’s not like that’s an optional extra—the exact point of my argument is that definite foreknowledge renders propositions in the future accidentally necessary. Leaving it out dissolves the argument entirely. And again, no distinction is drawn between accidental and logical necessity either. So it’s hard to see what your professor is getting at here. This is how the argument should look:

1”. It’s accidentally necessary that God believes-because-it-is-true that I will use the restroom.
2”. It’s logically necessary that if God believes-because-it-is-true that I will use the restroom, then I will use the restroom.
3”. Therefore, it’s accidentally necessary that I will use the restroom.

It seems more like your professor is wanting to take issue with the principle of accidental necessity itself. But then recouching my argument is pointless. It just begs the question. He should simply take issue with the following statement:

PAN: if it is true that Q, then it is accidentally necessary that Q

But PAN seems intuitively obvious. Let’s say Q stands for “Columbus discovered America”. This fact is true. It’s also unavoidable, because it’s in the past. We can see that it’s not logically necessary, and we can see that prior to its becoming true it could just as equally have become false. But now that it is true, it’s unavoidably true. And that’s all that’s meant by “accidentally necessary”.

However, if PAN is intuitively obvious, then it also seems clear Q can be accidentally necessary even if it’s in the future, simply by merit of a conditional statement. If it is unavoidably true that tomorrow Q, then it is accidentally necessary that tomorrow Q.

And that’s where God’s foreknowledge comes in: it makes the conditional certain. Given definite foreknowledge of Q, Q becomes accidentally necessary. Thus definite foreknowledge of our actions renders them accidentally necessary prior to our taking them.

Jordan Fishel

Thank you for that wonderful explanation, Dominic. It is just what I was hoping for: something a bit more palatable. I realized that this article was a response contextualized to your other debate. As such, that LFW may entail open-theism is a fair assessment. But whether or not LFW logically necessitates open-theism is certainly the issue here. I will mull over your replies. To say the least, this will be good material for an engaging discussion with my friend. Thanks for your time, Dominic, and God bless.