Friday, July 2, 2010

The Foreknowledge Argument and Hyper Libertarian Free Will

Here's William Hasker's version of the foreknowledge argument.

(B1) It is now true that Clarence will have a cheese omelet for breakfast tomorrow. (Premise)
(B2) It is impossible that God should at any time believe what is false, or fail to believe anything that is true. (Premise: divine omniscience)
(B3) Therefore, God has always believed that Clarence will have a cheese omelet for breakfast tomorrow. (From 1, 2)
(B4) If God has always believed a certain thing, it is not in anyone's power to bring it about that God has not always believed that thing. (Premise: the unalterability of the past)
(B5) Therefore, it is not in Clarence's power to bring it about that God has not always believed that he would have a cheese omelet for breakfast. (From 3, 4)
(B6) It is not possible for it to be true both that God has always believed that Clarence would have a cheese omelet for breakfast, and that he does not in fact have one. (From 2)
(B7) Therefore, it is not in Clarence's power to refrain from having a cheese omelet for breakfast tomorrow. (From 5,6) So Clarence's eating the omelet tomorrow is not an act of free choice.

(Hasker. God, Time and Knowledge. p.69)

The weak point in the argument is B4 - although it certainly sounds bad to say it's in my power to bring it about that God has not always believed something He has always believed. So I think best strategy for the libertarian is to grant the argument and the truth of B4, but explain why the argument rules out hyper-libertarian free will (link) and not libertarian free will.

Some libertarians have rejected B4 as false and I don't blame them - I used to be one of them. B4's phrase "power to bring it about" is equivocal and can be understood in a way in which B4 is false. If by B4 all we mean is that Clarence has the ability to refrain from eating an omelet tomorrow such that if he were to refrain from doing so, God would have always believed he would refrain from doing so, then B4 is false. If B4 is understood in this so called 'counterfactual power over the past' sense, we should reject B4. But Open Theist are usually unsatisfied with a counterfactual power over the past sense of 'bring it about' and seek something more.

I think that by "power to bring it about God has not always believed...", Hasker means the power to change the proposition that "God has not always believed..." from possibly true to actually true. In that case, we have no such ability and if that's what freedom means, the argument validly rules out such a freedom. But LFW doesn't require such a freedom only the Open Theist's hyper-libertarian free will requires the ability to change propositions from possibly true to actually true.1 So the argument rules out hyper-LFW, not LFW. Proponents of the foreknowledge argument face a delema: either reject B4 as understood in a 'counterfactual power over the past' sense or accept B4 in a hyper-LFW sense. Either way, LFW is untouched by the foreknowledge argument.


-------------------------------------------
1Let's contrast 'counterfactual power over the past' verses the power to change a proposition from possibly true to actually true once more. According to counterfactual power over the past, Clarence has the ability to refrain from eating an omelet tomorrow such that if he were to refrain from doing so, God would have always believed he would refrain from doing so. This isn't a retro-causal ability. Clarence can't cause or uncause the past. Clarence's causal abilities here are to eat the omelet or not. We must avoid the temptation of confusing Clarence's causal abilities with the logical implications of his using those abilities. By talking about what would have been true if Clarence were to use his abilities, we are moving away from talking about what is actually true to what would be hypothetically true. We are talking about hypotheticals, not actuals.

On the other hand, if Clarence has the power to change the proposition "Clarence refrained from eating an omlet" from possibly true to actually true, we are talking about an actual ability Clarence has. We are talking about actuals, not hypotheticals. Any strangeness about the idea of the past having been different seems to melt away when we keep in mind that we are talking about a hypothetical scenario rather than Clarence's actual ability.

13 comments:

Steven said...

I don't understand this distinction you are creating between hyper libertarianism and plain ol' libertarianism. What is your philosophy of free will, and why does it not lead you to open theism?

Godismyjudge said...

Steven,

LFW is the ability to choose A or not. Hyper-LFW includes the ability to convert the propostion Bob choose A from possibly true, to actually true. For more, please see the post on 'additions to free will'. LFW avoids the foreknowledge argument, but hyper-LFW does not.

God be with you,
Dan

Steven said...

"LFW is the ability to choose A or not."

How are we understanding "ability" here? Hypothetical or absolute?

Godismyjudge said...

Steven,

Using your terms, absolute, I think. I take a hypothetical ability to mean something like Bob is hypothetically able to choose chocolate in world W at T, if in some world W* at T, Bob is absolutely able to choose chocolate.

God be with you,
Dan

Steven said...

Okay, thanks Dan. :-)

You'll notice I said, "Thus, it had to be possible for him to actually do A, and actually do B (where A and B are two abilities of his which are distinct) at t, given the way the world turned out up to t." That is what having two absolute abilities at t looks like--two ways he can go, given the way the world has turned out.

But it seems that's not true if God has extensive foreknowledge of future propositions. There's only one way the world go at t consistent with the way it turned out up until then. And your posts recently on the divided and compound sense admit as much, when you say that in the compound sense, the agent cannot do otherwise. It is precisely the compound sense that is being discussed in the argument; and none of us exist except in the compound sense. We don't exist in states of affairs abstracted from the course of time and God's foreknowledge; we exist in a universe where God had knowledge of our actions long before we ever performed them.

If that's right, then no one has more than one absolute ability. We have hypothetical ability to do otherwise, and the other worlds are very similar except for facts about what God believed; but if you buy into the compound/divided sense, then you must admit that no one has more than one absolute ability.

Godismyjudge said...

Steven,

You'll notice I said, "Thus, it had to be possible for him to actually do A, and actually do B (where A and B are two abilities of his which are distinct) at t, given the way the world turned out up to t." That is what having two absolute abilities at t looks like--two ways he can go, given the way the world has turned out.

Fair summary (assuming you are not including with ability the ability to change propositions about A & B from possibly true to actually true). This is freedom in a divided sense – plain old LFW, not hyper-LFW. Bob can choose chocolate or not. This is an actual ability Bob has, not a hypothetical one. No causal activity prior to T renders impossible Bob’s choice of chocolate or not; nor does any truth prior to T conflict with Bob’s ability to choose chocolate or not.

But it seems that's not true if God has extensive foreknowledge of future propositions. There's only one way the world go at t consistent with the way it turned out up until then. And your posts recently on the divided and compound sense admit as much, when you say that in the compound sense, the agent cannot do otherwise. It is precisely the compound sense that is being discussed in the argument; and none of us exist except in the compound sense. We don't exist in states of affairs abstracted from the course of time and God's foreknowledge; we exist in a universe where God had knowledge of our actions long before we ever performed them.
The main problem here is your first sentence (i.e. But it seems that's not true if God has extensive foreknowledge of future propositions.) There’s no conflict between LFW (as described above) and God’s foreknowledge. You said it yourself, the foreknowledge argument discussed the compound sense (not the divided sense described above). The divided sense is about ability (Bob can choose chocolate), the compound sense is about the hypothetical use of that ability (If Bob chooses chocolate, then…). The divided sense is about personal abilities (Bob can…) the compound sense is about impersonal possibilities (it is impossible that, given “Bob will not choose chocolate” is true, Bob will choose chocolate). The divided sense is about causal possibilities, the compound sense is about logical compossibilities.

Not accounting for these distinctions, leads you to see a conflict between two absolute abilities and God’s foreknowledge, where none exists. Of course, if you added to “absolute ability” the idea that we can change a proposition from possibly true to actually true, then you would have the weapon needed to smash the distinction between the compound and divided sense entirely. But without that, I think you need to live with the idea that no prior truth is in conflict with Bob’s ability to choose chocolate or not, even if some prior truths are in conflict with his hypothetically using the ability that he will not use.

God be with you,
Dan

Anonymous said...

(from Steven)

Being able to do otherwise in the divided sense is not the same as having absolute ability to do otherwise. Having an absolute ability involves what you are able to do, what options are open to you, given the way the world has turned out. But there is no "way the world turned out" in the divided sense, because the divided sense is just an abstraction, it is just considering whether the physical facts (and, presumably, mental facts) determine the action the agent will do.

(ii) No one exists in the divided sense. The divided sense does not a whole to help the libertarian cause, because what libertarians typically want is real world, real life ability to do otherwise. This would be ability to do otherwise in the compound sense--after all, we live in the compound sense.

(iii) You don't seem to be a typical libertarian. You don't think real world, real flesh and blood ability to do otherwise, given the way the universe has turned out thus far, is required. That's fine and dandy, but you have to give up PAP

Godismyjudge said...

Steven,

Being able to do otherwise in the divided sense is not the same as having absolute ability to do otherwise. Having an absolute ability involves what you are able to do, what options are open to you, given the way the world has turned out.

So does the divided sense.

But there is no "way the world turned out" in the divided sense, because the divided sense is just an abstraction, it is just considering whether the physical facts (and, presumably, mental facts) determine the action the agent will do.


That’s not quite accurate. The divided sense includes all facts and causes preceding the action. Remember, it’s not God’s foreknowledge and our ability that is in conflict. It’s God’s foreknowledge and the hypothetical, counterfactual use of that ability that’s in conflict.

(ii) No one exists in the divided sense. The divided sense does not a whole to help the libertarian cause, because what libertarians typically want is real world, real life ability to do otherwise. This would be ability to do otherwise in the compound sense--after all, we live in the compound sense.

Strictly speaking, we don’t live in a ‘sense’; either divided or compound. Some propositions are in a divided sense (i.e. Bob can choose chocolate) and some are in the compound sense (i.e. it is impossible that God knows Bob will not choose chocolate and the proposition “Bob will choose chocolate” is true). But we live in the world, the real world, not in a ‘sense’. However, you are arguing for hypothetical abilities (when the person is unable to do the thing in the real world) and you are arguing based on the results of hypothetical counterfactual futures, so I find your ‘real world’ objection odd.

(iii) You don't seem to be a typical libertarian.

Well I have quoted Arminius, Molina, Suarez, some of the schoolmen and WLC saying what I am saying. I could have quoted others as well. But those are the main theologians that have tried to reconcile God’s foreknowledge with freewill.

That's fine and dandy, but you have to give up PAP

Why?

God be with you,
Dan

Greg said...
This comment has been removed by the author.
Greg said...

Forgive me for commenting an old post, but I am wondering: wouldn't it be better to define LFW as 'the ability to choose without sufficiently determining factors'? Thus, to be free, an alternative choice would not have to be actualizable just not determined.
In that case, one could just accept Hasker's argument as sound but no problem for LFW.

Godismyjudge said...

Hi Greg,

Thanks for your comment. I wasn't 100% clear on this comment:

"wouldn't it be better to define LFW as 'the ability to choose without sufficiently determining factors'? Thus, to be free, an alternative choice would not have to be actualizable just not determined."

I take this as the idea that so long as we are ultimatly responsible and our actions are not causally determined, it doesn't matter if we have alternative possiblities or not. Again, I wasn't sure if that's what you meant.

If that's right, I suppose I have a few reservations about that approach. First of, the abilty to choose seems to imply alternative possiblities. We choose between alternatives.

Second, I tend to agree with Aristotle that "when the origin of the actions is in him, it is also up to him to do them or not to do them” (1985, Ethics. Book III). In other words, ultimate responsiblity or freedom from sufficent causes implies alternative possibilities.

In an case, I hope I have not missread you.

God be with you,
Dan

Greg said...

It seems you did understand me, even though, upon re-reading, I see that my original comment WAS unclear.

Think of the Frankfurt counterexamples. Consider a scenario where, unbeknownst to you, someone implanted a chip into your brain that allowed him to know your thoughts and control your actions whenever he wanted. Suppose he intended to control you only if he saw you were heading toward a certain alternative A rather than B (perhaps a necessary but not sufficient telling sign) or you could even suppose the guy somehow had counterfactual knowledge from God but not foreknowledge. In any case, you do not show the sign, so he does not control you. You choose B freely, even though he would have forced you if you had entered the circumstances in which you would have freely chosen A in the libertarian sense.
Now, you can (1) hold the PAP and deny LFW, (2) modify LFW to deny PAP, (3) keep PAP and LFW but deny there are any true creaturely counterfactuals (I know you are a Molinist),(4) deny that such a scenario is possible for some other reason or (5) say I've misunderstood the terms and concepts. To me the reasonable option seems to be (2). You appear inclined to (4) or (5), but why?

As for the two things you mentioned already: (1)under normal circumstances we are able to choose either of the alternatives but if there is an intervening determining factor, we are not free to do what the factor determines although we are free to do what no factor determines and (2) My point is that if you choose B without sufficient causes, prior determining factors, enacting that choice, the origin IS within you. It was not determined by any sufficient causes, only by your LFW. Even if such causes would not allow the alternative choice A.

I need to study much more, but those are just my current thoughts.

Godismyjudge said...

Hi Greg,


Good, questions. Your right, I would go with 4 since I do have some problems with the Frankfort example; I think the whole issue is “the tell” and the intervener is somewhat of a red herring. In many ways “the tell” smuggles elements of the foreknowledge argument into the Frankfort example. The tell may be construed as either deterministic or indeterministic. If it’s deterministic (whether the tell itself makes us vote republican or simply implies that something determines that we vote republican) then the Frankfort example question begs in favor of determinism before you get to the intervener. Thus the irrelevance of the intervener.

For this reason some folks reject the tell as impossible. But I can suppose some indeterminic ways to understand the tell. What if the tell is counterfactually dependent on what I would do (i.e. if I would have voted democrat the tell would have said democrat)? Again, this assumes AP and again, the interesting part is settled before we get to the intervener.

The tell is in the past. Since the past sometimes is part of the way in which we understand LFW, yes this does get into questions of how we define LFW. We don’t have retro-causal powers, that violates our ideas of causation and time and the inalterability of the past. So I can’t cause a republican tell past to become a democratic tell past. But it’s not so clear that a counterfactual dependence runs into this problem since only under a hypothetically different future do we have a hypothetically different past. And even then, only the elements of the past that logically depend on our future choice would be different in this hypothetical future in which we “do otherwise”; it’s not like the causal antecedents of our choice would have been different as compatiblism requires.

The same happens when we replace the controller for God’s counterfactual knowledge; that presupposes either AP or determinism and therefore eliminates the need for the controller. Take Molinism; God’s knowledge is logically ordered into 1) what can happen (possibilities or natural knowledge), 2) what would happen (hypotheticals or middle knowledge) and 3) what will happen (the future or free knowledge). Middle knowledge is God knowing what you would freely do under any circumstance. So let’s say you in the circumstance that you are in an ice cream shop, God knows you can choose chocolate or you can choose vanilla. That’s at least two possible worlds (not counting a third world in which He prevents you from entering the shop or a fourth in which you are in the shop but choose not to eat…). Via God’s middle knowledge, God knows you would choose vanilla, so of the two possible worlds we are down to one feasible world. Thus middle knowledge implies at least two possibilities.

Given determinism you could only eat vanilla and there is only one possible world in which you eat vanilla. In that case God doesn’t need middle knowledge and in fact there is no difference between God’s natural knowledge and His middle knowledge. Could God still have some sort of deterministic counterfactual knowledge in such a system? I suppose. So then we have either middle knowledge, which implies AP or deterministic counterfactual knowledge, which implies determinism. Either way, the question of AP is decided before we even get to the Frankfort example.

As for your last paragraph, I guess I view alternatives as an essential aspect of choosing. No alternatives, no choice. Also, as for self-determination to one and only one thing, I would say that if it’s true the we were built in such a way that we determine that one thing and ultimate responsibly would be traced through to however we got that way.

God be with you,
Dan