Wednesday, January 08, 2014

Is aftknowledge equivalent to foreknowledge?




The latest Unbelievable? podcast featured a prominent Calvinist debating William Lane Craig on middle knowledge. I have posted my summary along with a link to the MP3 file here:
http://winteryknight.wordpress.com/2014/01/07/william-lane-craig-and-paul-helm-discuss-calvinism-and-molinism-on-the-unbelievable-radio-show/

Of course, Josh is far too partisan to ever consider weaknesses in his own position, but he overlooks a fundamental asymmetry between knowledge of past choices and knowledge of future choices. Given libertarian freedom, when facing forward, there's a forking path of two or more possible choices ahead of us, any one of which may become actual. There's no one choice to foresee. Rather, we see various alternate routes fanning into different timelines. 

But looking back, there is only one actual choice. Only one path was taken. There is only one choice to be known. 

So, of course it's possible to know which choice was made once that's past, once that's over and done with. Time's passage is, itself, is a process of elimination. It hardly follows that it's equally possible to know which choice will be made in advance of the fact. 

For that matter, it's far from obvious, given freewill theism, that God is automatically be privy to past human choices. 

44 comments:

  1. Libertarian freedom isn't an issue. It's truth about the future that's the issue. If the future is open, then it can't be known until it's settled truth. Once there is settled truth, it can be known. But settled truth about the future is certainly compatible with libertarian freedom. It's just truth that I will libertarianly freely do certain things. What's not compatible with libertarian freedom is something ahead of time ensuring that it will be true. So the truthmakers of truths about future contingents would have to be in the future, not in the present. Then the only further issue is how God might know the truths that are grounded in my future choices, and atemporality solves that, as would some sort of ability to look into the settled future by a temporal foreknower in the present or receiving an infallible word from a being who could do so. You can't do it by means of looking to what the present will cause, as a determinist can do, but there are plenty of other ways to make foreknowledge epistemically certain without threatening libertarian free will. You just have to admit to the reality of the future. But that's no problem if you can admit to the reality of the past, which is certainly settled as of the present but not in a way that threatens my past free choices. I'm not a libertarian, but I see no incompatibility between libertarianism and foreknowledge provided you have one actual future that's distinct from the possible but non-actual futures.

    ReplyDelete
    Replies
    1. I'm afraid I don't see how divine timelessness relieves the tension. As Peter van Inwagen recently argued:

      Suppose I told a lie at 11:46 a.m. Eastern Standard Time, on 23 December 2006. An omniscient but non-temporal God has—given only that, for every time t, any proposition that is about states of affairs subsequent to t, is ‘already’ true at t or already false at t—the following ability: He is able (acting, of course, extra- temporally) to cause a monument to have come into existence ex nihilo in 1900, a great slab of stone on which the following words were inscribed: ‘On 21 September 1942, a human being named Peter van Inwagen will be born. On 23 December 2006, at 11:46 a.m. Eastern Time (or Eastern Standard Time, as Eastern Time will then be known), he will have to choose between lying and telling the truth. He will choose to lie.’ Suppose God has done this thing he is able to have done. Can it be that my lying—my telling a lie at 11:46 a.m. EST, 23 December 2006—was a free act? That is, was I able, on that occasion, to tell the truth? Well, was there, just before that moment, a possible continuation of the (then) present state of affairs in which I told the truth? Let us consider all the possible continuations of that state of affairs. It is true in every one of them that a monument (inscribed with just the words recorded above) came into existence ex nihilo in 1900—and true that its coming to be was caused by God’s extra-temporal act of creation. Is it true in any of the possible continuations of the then-present state of affairs that the words inscribed on the monument did not express a true proposition? No, for in that case God would either have been mistaken or have been a deceiver, and both are impossible. My act (my telling the lie), was therefore not a free act, for if it were, I should have had access to a possible continuation of the then-present state of affairs in which no monument (so inscribed) was brought into existence by God in 1900 or in which one was but the words inscribed on it did not express a true proposition—and no continuation of that state of affairs that answers to either of these descriptions exists. We may say that the monument in this imaginary case is a ‘Freedom-denying Prophetic Object’. The concept of a Freedom-denying Prophetic Object is a very abstract one: a divinely inspired human prophet who has foretold certain actions of human beings would also be a Freedom-denying Prophetic Object. It is no doubt true that very few human actions can be shown to be unfree on the ground that that they were ‘foretold’ by the statement that they should occur having been somehow being encoded in the structure of a Freedom-denying Prophetic Object—for (no doubt) very few human actions have been so foretold. But this seems irrelevant to the question whether human beings have free will. Suppose God should reveal to us that on a distant, uninhabited planet, he had created, before the existence of the first human being, a vast library that contained minutely detailed biographies (correct in every detail) of all the human beings who would ever live. Would the following be a reasonable reaction to this revelation: ‘Ah, then human free will does not exist. But if God had not created that library (and had not done anything relevantly similar), human free will would exist.’? Obviously not: although it follows logically from the story of the monument that my lie was not a free act, the fact that my lie was not a free act cannot be supposed to be a consequence of the existence of the monument.

      "What Does an Omniscience Being Know"? J. Kvanvig, ed. Oxford Studies in Philosophy of Religion (OUP 2008), 1:218-19.

      Delete
    2. Jeremy,

      How would you respond to this argument?

      If that Jones sits at t was true a thousand years ago, then there must have been something temporally nonrelational about the world a thousand years ago in virtue of which this was true...But if the thousand-years-ago truth of that Jones sits at t was made true by a hard feature of the world a thousand years ago, then whereas this truth depended on the world, it depended on a part of the world outside Jones’s control at t, namely, how the world was a thousand years prior!

      J. M Fischer & P. Todd, “The Truth about Freedom: A Reply to Merricks,” Philosophical Review 120 (2013).

      Delete
    3. There are two foreknowledge problems. The Aristotelian problem is whether there can be truth about contingent matters in the future. The Epicurean view is that Aristotle's problem shows the incompatibility of freedom and truth about future free choices. Carneades the Skeptic responded that the argument against the compatibility assumes that the fixity of the future is like the fixity of determinism, but he insisted that it is not. The first involves mere truth about the future, not predetermination or causation. It can be true that you do X but possible that you do not X. One is about actuality. The other is about which other possibilities could have been actual (even though they are not actual). This is a pretty important advancement in modal logic, but it's a pretty basic point nowadays. Something can be possible but not actual. Therefore something in the future can be contingent without being predetermined but nonetheless but what will truly happen. So truth about the future is compatible with its not being predetermined, and that means libertarian free will is compatible with truth about what you will freely choose.

      Augustine thought this was so clear and obvious that he spends not more than a few sentences explaining Cicero's acknowledgement of this point before moving on to the foreknowledge problem that really bothered him, which was how someone could know the future truth without knowing it the way the Stoics thought God would know the future. They thought it was merely a matter of tracing out causes to see what would be guaranteed to happen. Augustine spent many pages throughout City of God trying to sort through this, but it took Boethius to point out that Augustine's own view of atemporality could answer this question easily.

      As an answer to the second question, Boethius' response is pretty good. As an answer to the first, it doesn't help at all, because you'd still need to explain how a prophet in time given a message by God would be able to know it (never mind how it could be true to begin with). But the answer given centuries earlier by Carneades is all you need, and it simply requires two premises: (a) the B-theory of time and (b) truthmakers about future contingent acts are at the time they take place, not necessarily at the time you know them.

      Once you grant those assumptions, van Inwagen's argument and the Fischer/Todd point do not work. They assume the denial of those assumptions.

      Delete
    4. "But the answer given centuries earlier by Carneades is all you need, and it simply requires two premises: (a) the B-theory of time and (b) truthmakers about future contingent acts are at the time they take place, not necessarily at the time you know them."

      i) Don't freewill theists typically reject the B-theory of time in favor of some A-theoretical version?

      ii) Wouldn't the truthmaker for a future contingent choice be the future event? It's the future event in virtue of which a proposition about a future contingent choice is true? If so, how could the libertarian God access that truthmaker before it happens? For until it happens, it could go either way.

      iii) What about freewill theists like W. L. Craig who reject truthmaker theory?

      Delete
    5. (i) Those are independent questions, and there are a number of people who accept a B-theory of time and libertarianism about free will. Offhand, I can think of John Hawthorne, Linda Zagzebski, and Greg Ganssle. I'm sure there are lots of others. According to second-hand reports, Peter van Inwagen thinks the A-theory/B-theory dispute is meaningless (which means he doesn't hold to an A-theory, even if he also doesn't assert a B-theory). An example of a historical philosopher who held both is Cicero, and according to Cicero so did Carneades, who it seems by Cicero's report wasn't the thoroughgoing skeptic he's often made out to be. I'm sure that, at least at one point in the development of his thought, Augustine held both as well, although I think he ultimately rejected his libertarianism.

      (ii) You have to treat the future and the past symmetrically, and you haven't done so in this analysis. You can't say "until it happens, it could go either way". It's simply true that it could go either way, and that's true even afterward. Afterward, you say that it could have gone the other way. The modal status of being contingent is not just true beforehand but also afterward, and thus the other possibilities are real possible worlds both beforehand and afterward. Beforehand, you say it by saying it could go either way, and afterward you say it by saying it could have gone either way, but the reason you say it differently is because you know how it turned out. That's an epistemic difference, not a metaphysical one. The past and future are equally settled and equally unsettled, on this view. They're settled in the sense of being actual and unsettled in cases when they're contingent in the sense of there being a possible world where something else happens. That's as true of the past as it is of the future and as true of the future as it is of the past.

      (iii) They hold that something happens without any explanation at all. That's a proposition that even Epicurus denied, quite explicitly. I've not seen a philosopher before David Hume even remotely tolerant of such a notion, and Hume was wrong to go a different way on that question. Even all the libertarians before Hume wanted to maintain that nothing happens without some explanation. I have little patience for Craig's way out of the truthmaker objection to Molinism. I think that's an incredibly powerful objection to that view.

      Delete
    6. Jeremy Pierce:

      

"You have to treat the future and the past symmetrically, and you haven't done so in this analysis."

      Why do I have to do that?

      "It's simply true that it could go either way, and that's true even afterward."

      What about the accidental necessity of the past? Time's arrow tracks the order of causation.

      "Beforehand, you say it by saying it could go either way, and afterward you say it by saying it could have gone either way, but the reason you say it differently is because you know how it turned out. That's an epistemic difference, not a metaphysical one."

      It's both. Before the libertarian agent decides, it could go either way. After he decides, that closes the door on alternate routes.

      To my mind, libertarian freedom is like a series of hallways. You have various doors to choose from. Once you pass through one door, it locks behind you. You can't go back and try the other doors in that hallway. Rather, you find yourself in a new hallway. The process repeats itself.

      Delete
    7. Because you're evaluating a proposal I gave for how the two are consistent. You can't reject some of the key elements of that proposal and then claim that the proposal without the crucial elements doesn't do what it does when it does have those. That would be like saying theism has no resources for responding to the problem of evil on the grounds that theism includes belief in God, and you don't like belief in God.

      Delete
    8. Time's arrow tracks the usual order of causation. It need not track the only order of causation. Anyone believe in a timeless God better say there are other pathways of causation than just in time from earlier to later moments in time.

      For libertarians, the opportunity to make a choice for that person is lost afterward, but the contingency of the event is still present even afterward. And that's what I'm saying matters for its modal status. I think the hallway metaphor is fine on the epistemic level, and I think it's fine on the level of what's available to the agent. But even compatibilists can say that. Once I've voted, I can't vote again (legally anyway). There's nothing metaphysical about that other than that I am no longer in a position to influence what will happen. It doesn't require thinking the past and future are asymmetrical.

      Delete
    9. Jeremy Pierce

      "Time's arrow tracks the usual order of causation. It need not track the only order of causation. Anyone believe in a timeless God better say there are other pathways of causation than just in time from earlier to later moments in time."

      That fails to distinguish between primary and secondary causality. We've been discussing the historical process. That involves intramundane causality, where causes precede their effects.

      "There's nothing metaphysical about that other than that I am no longer in a position to influence what will happen. It doesn't require thinking the past and future are asymmetrical."

      When you say the agent "is *no longer* in a position to influence what will happen", you just stated a basic asymmetry.

      Delete
    10. Jeremy Pierce

      "Those are independent questions, and there are a number of people who accept a B-theory of time and libertarianism about free will. Offhand, I can think of John Hawthorne, Linda Zagzebski, and Greg Ganssle. I'm sure there are lots of others."

      It's unclear how they square the B-theory with access to alternate possibilities. According to the B-theory, the entire timeline is a given totality. The future is actual. So at that point an agent can't choose between alternate future outcomes.

      Perhaps they'd say God knows which future a human agent would choose, and God instantiates the B-theoretical timeline which matches what the agent would choose, if given the opportunity.

      If so, that simply shifts the libertarian conundrum from how God can know what a free agent will do to how he can know what a free agent would do.

      In addition, it's not clear on that scenario how God decides which human choice to instantiate. For if humans have the ability to do otherwise, then there's a possible world for each choice and its contrary.

      Delete
    11. Jeremy Pierce

      "Time's arrow tracks the usual order of causation. It need not track the only order of causation. Anyone believe in a timeless God better say there are other pathways of causation than just in time from earlier to later moments in time."

      Didn't Jesus *have* to die before he could be raised from the dead? Didn't Jesus *have* to be born before he could die?

      Delete
  2. So let's say someone puts forward the following:
    - God exists atemporally;
    - God's knowledge is infallible;
    - At some future time, I can freely choose between two options A & B in the sense libertarians want;
    - My freely choosing A at that future time is the truth-maker for the contingent truth that I choose A rather than B;

    Does God's knowledge ensure that I choose A rather than B? Seems so. Granted, God's foreknowledge isn't the truth-maker for my choosing A rather than B at some time; God's knowledge depends on my choice in time. Even so, given God infallibly knows - even through some sense of dependence on seeing what the creature freely does - it is nevertheless impossible for the agent to do otherwise *holding fixed* what God foreknows. So if God "foreknows" this atemporally, I couldn't have chosen B rather than A. And so it seems libertarianism is at issue again.

    Where's the error here, Jeremy?

    ReplyDelete
    Replies
    1. In what sense does God's knowledge ensure it? It ensures it in the sense that if God knows it, it's true. But that's trivial. If I know it, it's true. That doesn't mean my knowledge makes it happen or threatens anyone's freedom. If I know that right now you're libertarianly-freely choosing to write a comment in response to me, that doesn't mean you're not free in doing so. It means you are, or I couldn't know that you were doing it freely. But nevertheless my knowledge guarantees it in the sense that if I know it it's true.

      The problem here is really a modal fallacy. The following statement is correct:

      Necessarily: If God knows it, it's true.

      But the following is not:

      If God knows it, it is a necessary truth.

      Epicurus and Alexander of Aphrodisias were prominent defenders of this argument, and you can see Augustine and Boethius showing awareness that it was out there in their times as well, but Carneades' point seems to me to show the fallacy.

      Delete
    2. I think it is important to see that I nowhere claim that God's knowledge plays a *causal* role; I don't think any incompatibilist (about God's knowledge and libertarian freedom) requires that much. I also don't think you've captured the correct way of stating the incompatibilist's argument. Granted, you have presented one way an incompatibilist's argument can run - one in which there is a modal fallacy, which everyone I've read agrees is a bad argument. Let me give it another go.

      The libertarian is committed to this claim: given the actual past and the laws of nature, an agent is free with respect to some action/choice/decision (whatever) only if the agent can do otherwise.

      As you know, one way of filling out the consequent is to give a possible world's account in which two worlds have the same laws and past events/features, but are distinguished by the different choices the agent makes at branching moment. Let's assume this is the way the libertarian wants to understand the relevant ability to do otherwise.

      So given the actual world, Jones will arrive at a moment where he must choose to mow his lawn or not. Suppose he will in fact choose to mow his lawn. Then we know given God's omniscience that God believed (if you prefer, timelessly believes that) Jones will mow his lawn. That fact about God is something we are holding fixed along with all the other past events and laws that occur/hold with we ask whether Jones can do other than mow his lawn.

      So if Jones can do other than mow his lawn *given* what we are holding fixed, there is a world where the same past and laws hold, and that includes God's believing that Jones will mow his lawn. Is there such a world? (Time to talk about where God's beliefs are hard or soft, I suppose.)

      Delete
    3. But my whole point (and why this is an example of that same modal fallacy) is that you can't hold God's knowledge fixed. God's knowledge is what it is because of what God knows. If you mow your lawn, then in the world where you don't mow your lawn God knows that you don't mow your lawn. That world is a possible world, and therefore it's possible that you don't mow your lawn, and in such a case God would have known all along that you wouldn't mow the lawn. Libertarian freedom seems to me to involve two components, (1) that you are the cause rather than events beforehand (or afterward, or timelessly) and (2) that you could have done otherwise than what you actually do. I don't see why you don't have both of those in this case. The order of explanation is important here for the possibility claim. It would be different if God's knowledge explained why you do it, because then it would be legitimate to argue that God's knowing is explanatorily prior and thus is a given preventing you from doing otherwise. But the whole point of saying that God's knowledge is not explanatorily prior but rather is what it is because of what you choose is that you can't legitimately argue that it's a given in any way that explains your choice. What goes into the choice and makes it possible is logically and explanatorily prior to God's knowledge of it. Only if you take God's knowledge to be necessary in the sense that the particular content of what God knows is necessary do you get what you're saying. But that is precisely to the commit the modal fallacy I pointed to. So I don't think your argument is any different from that fallacious inference.

      Delete
    4. James is appealing to the fixity of the timeless realm. His argument would be better expressed this way:

      a. Nec(If God knows p, then p).

      b. Unpreventably (or unchangeably) God knows p

      c. Unpreventably (or unchangeably) p

      This undermines condition (2) above (btw, Kevin Timpe, for example, thinks your (2) is false).

      I know you object to (b), but my only point is that there's no modal fallacy going on here, at least not the one you suggest. James may take issue with the operators scoping (b) and (c), but again, my point is that his argument will have *something* like this form.

      Delete
    5. When you say that the sort of argument I am running - a Pike-styled argument - makes the modal fallacy, I understand you are accusing me of moving the modal operator to the consequent illegitimately. What I don't understand is the sort of necessity you are saying that I have to be assuming. For instance, you say, "Only if you take God's knowledge to be necessary in the sense that the particular content of what God knows is necessary do you get what you're saying." What sort of necessity are you saying that I am invalidly transferring to the consequent... that in every world does Jones mow his lawn?

      I have another question. Isn't your response to me here just the Occamist response without being explicit about *soft* facts or beliefs?

      Last question: When the incompatibilist (about freedom and determinism) accuses the compatibilist of allowing for "free" actions that must necessarily happen, is the incompatibilist relying on the sort of argument a fatalist uses?

      Btw., the first condition you give for the libertarian shouldn't be there, since either everyone will agree (in which case it is trivial) or it rules out event-causal libertarian accounts, which shouldn't be dismissed out of hand.

      Delete
    6. Fletcher is right that I am assuming something like the above argument. The sort of necessity in play is more like accidental necessity; so all I need is that the future action of the agent is fixed in the same way that the past is.

      He's also right that there's no modal fallacy here. If the best the incompatibilist has to offer is an argument with a modal fallacy, the debate would have been over a long time ago and there wouldn't have been a debate about the best way to respond to the incompatibilist. Fischer and (I think) Tognazzini have something like an 11 premise argument outlining the various ways that compatiblists can respond. One of them is Boethian, one of them is Occamist, etc. It's very implausible to treat all of the incompatibilist arguments as relying on such a simple confusion, and it's not very charitable. But hey, I'm not claiming to have certainty about which position is right.

      Delete
    7. I'm having a hard time figuring out what accidental necessity is supposed to be. If it's not necessity, and it's just the sense in which the past is fixed, namely that there are truths about everything that occurs in it, then I've already said I have no problem with the future being like that, as long as you recognize that the past is also like the future in the sense that the past has contingency like the future. Defining this accidental necessity operator in terms of whatever is fixed about the past and then arguing that this makes the same be true of the future isn't going to lead to anything I'd find problematic. But I'd resist calling it unpreventable. I would call it unchangeable. What seems to me as the modal fallacy is taking this operator to have the force of necessity. It then becomes that fallacy. If it doesn't have that force, the conclusion isn't problematic. Of course the future is unchangable after the fact (because that's what the past is). But it doesn't follow that it's unpreventable beforehand. It just follows that it won't be prevented, not that there's no relevant possible world in which it's prevented.

      The difference between what I'm saying (which I attribute to Carneades) and the hard-facts soft-facts view Plantinga and Adams find in Occam is that the Occamist view depends entirely on whether there is a hard fact in the present, and the pre-torture case and the time travel case I gave seem to involve hard facts. My approach doesn't find that problematic, because I'm distinguishing not between hard facts and soft facts but between facts about the present and past that influence the decision and facts about the present and past that don't. Carneades' point is that libertarian freedom isn't threatened unless there's a causal problem. The mere fact that something in the present requires its truth isn't enough to threaten libertarian freedom.

      As for agent causation, I do think you need it or something like it. I'm not sure how event-causal libertarian accounts will do it. The only one I'm familiar with that denies agent causation is van Inwagen's, and he doesn't actually have an explanation for freedom, so I think he just doesn't have an account. But certainly the ability to do otherwise isn't sufficient, because then an Epicurean swerving atom, which allows for the ability to do otherwise, would count as a free choice. But a randomly-swerving atom isn't a free choices. It's as much outside my power as predetermined causes would be. Being caused by randomness certainly isn't better than being caused by my own predetermined desires and character traits. You need something more than just alternative possibilities, to rule out randomness. That plus needing some explanation gets you either compatibilism or something like agent causation.

      Delete
    8. Jeremy Pierce

      "I'm having a hard time figuring out what accidental necessity is supposed to be."

      The future is causally open in a way that the past is not.

      Delete
  3. If the future is as settled as the past then I don't know how you can have PAP. When we reflect upon our past we may say we *had* PAP prior to the past event being over and done with. But there certainly isn't any power of contrary choice about the past now. And if the future is as settled as the past, then there is no PAP, period.

    I'm not sure this sort of "possible" in "possible but non-actual futures" would satisfy many libertarians. It doesn't give them any live options, does it?

    ReplyDelete
  4. Jeremy, when you say, "the only further issue is how God might know the truths that are grounded in my future choices, and atemporality solves that," are you saying that LFW requires divine timelessness? The majority of libertarians will not be happy about that.

    You say this: "as would some sort of ability to look into the settled future by a temporal foreknower in the present or receiving an infallible word from a being who could do so," is compatible with LFW.

    I'm not clear on how this is compatible with LFW. If God (an infallible being) tells temporal Timmy (say he etches the proposition in the mason work of Timmy's house) at t1 that Linda will sit at t10, how is it that Linda refrains from sitting at t10? How does Linda have the power to, as they say, "add to the given past"? That is, all Linda can do *now* is to make the world-as-it-has-gone-up-to-now include one thing or some other thing.

    Presumably, *if Linda has LFW*, then she can either sit at t10 or refrain from sitting at t10. Intuitively, then, Linda's power or ability to *either* sit *or* not sit at t10 consists in this: to make the world-in-which-Timmy's-home-was-inscrupturated-at-t1 now (at t10) include her sitting *or* her standing.

    Now, it seems metaphysically impossible that God have a false belief. The inscripturation on the side of Timmy's house is the inscripturation of the belief of an infallible being, a being for whom it is a metaphysical impossibility that he have a false belief. It does indeed seem that there is only one future consistent with this past, i.e., the future where Linda sits at t10. The inscripturation cannot be mistaken, and so how does Linda refrain from sitting at t10? If she can't then how, exactly, does she have LFW?

    Another way to press the problem has been brought out by Patrick Todd in (“Prepunishment and Explanatory Dependence: A New Argument for the Incompatibility of Foreknowledge and Freedom,” Philosophical Review, forthcoming). Suppose God punishes Linda in jail at t1 for sitting at t10. How does Linda refrain from sitting at t10? If she has LFW, then it appears she can so refrain. Then what? Does she make her past not include her sitting in jail from t1-t9, say? But that certainly seems like a hard fact about the past. Does she make it a fact that she was punished but *unjustly* so? But we Christians may suppose God never directly punishes someone unjustly. The upshot here is this: this pre-punishment seems to be *on a par with* God's pre-inscripturation (to state the obvious, we're thinking of infallible *prophecy* here and how that is compatible with LFW) of his belief on Timmy's house; or even his telling Timmy in a dream, and so Timmy believes at least one belief that seems to entail a fact about the future, for if Timmy holds a false belief, then God does, which is impossible). So here's the question: if you've already been justly punished for X, how do you refrain from X-ing? It seems we could apply your same reasoning for how LFW is consistent with God knowing future settled truths to reconcile divine pre-punishment with human freedom. But this doesn't seem right. So how, then, isn't human *LFW* a factor in all of this?

    ReplyDelete
    Replies
    1. I never said LFW requires timelessness, just that it's one way (perhaps the best way) that you can answer the second foreknowledge argument, the one asking how it is that God would know the truths about future contingents.

      On the metaphysical possibility issue, the point is that it is metaphysically possible for someone with LFW to do otherwise. In the world where they do otherwise, God knows that the other thing is what they do. What God knows depends on what they do. That's what it means for a truthmaker to be the act and not God's knowledge of it. I'm not sure why you think God's knowledge that p in the actual world means God would know something false in a world where p is false. Then it wouldn't be knowledge.

      Delete
    2. This comment has been removed by the author.

      Delete
    3. Well, the dependence relation has to be the right sort. Not just ant sort will due; and so merely claiming that God's belief "depends on what they do" isn't sufficient to claim some sort of dependence and leave it at that. Moreover, it's simply not true that the relation x depends on y "means" y is *the truthmaker* for x. How so? Here's *one* sense of God's belief "depending" on something we do: God believes S will sit at t *only if* S sits at t. But this won't cut it, for it's also true that God decrees S will sit at t *only if* S sits at t. So you need to spell out *the right sort* of dependence relation. And one objection is that *our* beliefs seem to "depend" on the world in the same way you've indicated God's do, e.g., I believe you typed the above *because* you typed the above. But why think the dependence at issue in *foreknowledge* or, at any rate, *omniscience* questions is of *the same sort* of dependence as mundane me or workaday you? Other, more nefarious, problems lie in this direction, see the Fischer Todd article Steve cited above.

      In any case, I'm confused by your response: I said God infallibly believes that Linda sits at t2 and he inscribes this belief on the brickwork of Timmy's house at t1. You say Linda has LFW and so can do otherwise in the world; that is, you say Linda has the power to *add to the given past*. Well, that past includes the inscripturation of God's belief on the brickwork of Timmy's house. This is, as we say, a "hard fact" about Linda's past. Moreover, the proposition on the side of the house cannot be wrong, since it's an inscripturation of *God's* belief. So, you say Linda does otherwise; fine, then Linda changes the past. I say that's a jarring result!

      Finally, the pre-punishment point is still relevant. Linda sits at t10, this is immoral for some reason, and so God decides to punish Linda at t1–t9 (in jail, say). We can add that God believes that Linda sits at t10 "because" Linda sits at t10; that is, we meet your unspecified (as of yet) "dependency" relation. Further, God punishes Lisa at t1 *because* she sits at t10—again, meeting your dependency constraint; that is, Lisa doesn't sit at t10 *because* she was punished at t1. So, pre-punishment is *on a par* with the dependency condition, a condition you think is all that is needed for the agent to be libertarian free in light of God's infallible beliefs about all her actions. So you should be able to apply this response to the pre-punishment issue. So here's the question: How does Linda do otherwise than X at t10 if she's been punished for X at t1? I already addressed one response: Linda makes it the case that the punishment wasn't *just*. But God doesn't punish anyone unjustly. Your response should be able to reconcile this problem. But, it seems that divine pre-punishment rules out Linda's freedom, and thus your response, the "dependence response," is at best incomplete at the moment. (Btw, you should see I've not claimed God knows something that is false.)

      Delete
    4. I'm not sure I understand your first paragraph. Are you saying there are multiple models for dependence, and some of them won't allow what I want? If so, that's not a problem for merely finding the consistency of LFW and foreknowledge, because only one model of dependence is necessary. All you need is that if I were to do something different God would have known something different. And I find that entirely unproblematic. It even follows from the definition of knowledge.

      I never said anything about changing the past or adding to the past. I find those notions incoherent. The past is the past, just as the future is the future. There's no changing either. But you can certainly do things that cause the future to be what it will be, and you can similarly do things to make the past be what it was. I would recommend Dummett's famous paper on retroactive prayer on this subject. It's absolutely excellent. Hard facts aren't the issue. It's whether the hard fact is the explanation for the free choice, and it certainly isn't. That's the point of Carneades' objection to the Epicureans. They were confusing truth and what makes it true.

      Something seems really wrong to me with the pre-punishment case, but I don't think I understood what you were saying when I read it last night. I have to think about that. I haven't read that paper, and I've never seen this argument before.

      Delete
    5. I'm saying that for all you've said the "dependence" point is irrelevant for solving the problem. I grant that it will come down to the right sort. Your sort doesn't seem to be the right sort, though.

      I take it that any conditional that would also show compatibilism to be true will not be accepted by the libertarian. When you say, "All you need is that if I were to do something different God would have known something different," then this gets the compatibilist off the hook. take theological determinism. If I were to do something different, then God would have decreed something different." Or, take D. Lewis: If I were to do something different, a law of nature wouldn't have been a law." So you need to say something more here.

      Furthermore, this isn't "all you need." It is indeed obvious that "If I were to do something different, God would have known something different." But the question is, "Can I do something different than X-at-t *given that* God knows that I will X-at-t." For all you've said, Boetheism seems irrelevant. For the Ockhamist can say the same thing. So here's an argument that shows it's not enough, where ∂ = accidentally necessary, unpreventably, etc:

      1. Nec(If God infallibly believes that Smith eats pizza at t, then Smith eats pizza at t).

      2. ∂(God infallibly believes that Smith eats a pizza at t)

      3. ∂(Smith eats a pizza at t)

      Now, now doubt it is true that *if* Smith doesn't eat a pizza at t, then God wouldn't have believed it. My question, though, is this: *Given (1) and (2)*, how do you avoid (3)?

      You said, "I never said anything about changing the past . . . I find that notion incoherent." But you said God could tell a finite knower one of his infallible beliefs. So I say he tells finite Timmy at t1 his belief that Linda sits at t2. He "tells" Timmy by inscripturating this on the side of Timmy's house. If Linda is libertarian free to sit at t2, then her power at t2 is to *add to the given past*, the past which includes the writing on the wall, so to speak. Linda's power is the power to make the world-in-which-Timmy's-home-was-inscrupturated-at-t1 now (i.e., at t2) include her sitting *or* her standing. But every world in which she stands at t2 is *not* the world-in-which-Timmy's-home-was-inscrupturated-at-t1. So if she stands, she'd have to make her world not include that past fact. But her world *does* include it. So she'd have to change the past.

      On the pre-punishment paper, it's out now: http://philreview.dukejournals.org/content/122/4/619.short?rss=1

      Delete
    6. Here's the problem with the pre-punishment case. It's treating the past the way we don't treat the future, and keep in mind that I don't think you can reconcile LFW and foreknowledge if you treat the past and future asymmetrically. The future and past are fixed in the same sense as each other and not fixed in the same sense as each other. When you have backward causation, time travel, or foreknowledge it creates all sorts of problems unless you recognize that.

      In this particular case, what's going on is that there's some fact about the past that is incompatible with my having done otherwise. OK. But that's there in ordinary cases, just in the future. When I mow the lawn out of LFW, 30 minutes later the lawn is mowed. That fact is not present in any world where I chose not to mow the lawn. LFW requires that it not be in any world where I chose otherwise. But it's in the actual world. It's just not a fact that influences my decision, and so it's not a problem for LFW. The same has to be true about facts about the past. If Judas' decision was out of LFW, then there is some possible world where he doesn't betray Jesus at that moment of decision. In that world, either someone else does, or he does it at another time, or no one does. It doesn't matter which (for this issue). And in that world, the prophecy might have to have been different. But if the prophecy plays no role in his having done it, then it's compatible with LFW that the prophecy occurred beforehand. It doesn't matter if it's a hard fact about the past. It matters if the hard fact plays a role in getting him to do it. If the prophecy does play enough of a role in bringing about what happens, enough to guarantee the choice, then it's not free. But if it doesn't play that strong a role, it's compatible with LFW.

      Delete
    7. "If Judas' decision was out of LFW, then there is some possible world where he doesn't betray Jesus at that moment of decision."

      That's not the case. It's not just that there has to be *some* possible world, i.e., *any* ole world where he doesn't betray Jesus," it's that there is a *nearby* world where he doesn't, and here, "nearby" means "the same laws and intrinsic past is held fixed up to the moment of the choice." If you include worlds where the laws (of decrees, or etc) leading up to the choice are *different*, and this is relevant for securing *LFW*, then *determinism* is compatible with *LFW*! For it's possible that the actual world is determined and I do X. But in some other determined world with a different (non-extrinsic/non-relational) fact about the past, it's possible I don't X.

      So you need to consider only worlds with *identical* intrinsic pasts (and laws, decrees, etc) leading up to the choice.

      "It doesn't matter if it's a hard fact about the past. It matters if the hard fact plays a role in getting him to do it."

      I disagree. The hard fact need not "play a role" in getting him to do it yet still prevent him from not doing it. In the argument I presented above, the hard fact at (2) need not play a (causal?) role in getting Smith to eat pizza, yet given it's existence Smith can't do otherwise than eat pizza. You're making many of the same claims as Merricks in his recent defense of the compatibility between F&FK. I believe Merricks' position (e.g., "dependence" and "hard facts don't matter) was answered here: J. M Fischer & P. Todd, “The Truth about Freedom: A Reply to Merricks,” Philosophical Review 120 (2013).

      Other than that, I fail to see how your response was relevant to the pre-punishment example. You say god could tell a finite knower his belief about what an agent will do in the future. if he could do that, why couldn't he punish the agent for what the agent does in the future, and do so *because* the agent does it in the future? But once he has done so, how does the agent do otherwise?

      Delete
    8. The way this stuff is often taught to undergrads, it's put as a compatibility issue between determinism and free will. But in practice, what matters is not whether determinism is true but whether an individual choice is predetermined. Obviously a deterministic world across-the-board means all choices are predetermined, but you could have a world that largely follows deterministic laws but has a few instances of indeterministic libnertarian choices.

      My argument is that the same point needs to be extended. Libertarianism is about causal determinism. In particular, free will is defined in terms of not being caused by prior events to do the thing you do. If that's right, then it's actually a separate issue whether the entire history of the world has to be constant across the relevant possible worlds. I think these weird cases of backward causation show that you can't use "the entire history of the world constant but a different choices possible" kind of test to see whether libertarian choice is possible. The main idea behind libertarianism is that you could have done otherwise, as a genuine possibility, plus usually something like agent causation as the actual explanation in lieu of one in terms of events. It's usually assumed that this requires the relevant possible world where you do the other thing than the actual thing to be exactly like this world all the way up to the choice, but I think that's wrong. I think we assume that because we assume all causation is forward, and the only reason something beforehand would have to be different in the alternative possibility is because it causes your decision, and a different decision requires a different cause. But here we have a case where it would have to be different because it's an effect of your choice. In other words, I'm modus tollensing your modus ponens (I think).

      Change to a time travel case. Suppose I have libertarian freedom, and I time travel to the past, to a time before I exist. But I also time travel to a part of the universe that could not cause anything different about me, because it's so far away that its effects would not reach this part of the universe until after my time travel occurs. So there's no way that my time travel can have a causal effect on my time travel. Yet nonetheless there's the existence of my future self (in terms of my personal time) existing in my past (in terms of external time). There's no arguing that the mere presence of myself in the past on that other planet interferes with my choices. So obviously it's got to be compatible with libertarian freedom. But that means we can't build into the definition of libertarian freedom that the relevant possible worlds have to be the same in every way up to the moment of decision. We need to restrict it to things that can affect our choices. They are the same in every way in terms of the things that can affect our choices. An argument to the contrary would be sort of like saying time travel is impossible for a three-dimensionalist, because it's a priori impossible that someone can be wholly present in two places at the same time. I would have thought that time travel cases simply tell us that a supposed a priori principle like that is false. The same is true of the supposed a priori principle that libertarian freedom requires an alternative possibility where the world is the same in every way up to the moment of decision.

      On the last point, the agent doesn't do otherwise. That's not required for libertarian freedom. All that's required is that it's possible to do otherwise, not that it's actual to do otherwise. And possibility doesn't require the past being alike in every respect to the actual world, just to ways that the actual world might influence the choice.

      Delete
    9. Not sure why, Word told me I had 4033 characters but my post can't be accepted by blogger because it says I have more than 4096 characters. So I'll break it up:

      First, in even the most advanced literature, philosophers talk in terms of "determinism." At any rate, if you have "a world that largely follows deterministic laws but has a few instances of indeterministic libertarian choices," then these choices were not determined. So it *is* an issue of "determinism" and "free will," that is, can a free choice be determined? So that's the question, is it possible that this conjunction is true: Sam freely chose to do X and Sam was determined to do X (there some nuances to make here re: derivative freedom, but nothing hangs on that distinction now)

      Second, to your point that,"But that means we can't build into the definition of libertarian freedom that the relevant possible worlds have to be the same in every way up to the moment of decision," I have several things to say:

      1. Notice I have said that the *intrinsic* facts about the past need to be the same. I didn't mean, and said otherwise, that even the extrinsic or relational or, what we call *soft*, facts need to be the same.

      2. It's simply the case that everyone in the literature recognizes this. And here's why:

      a. Consider the consequence argument. It has a premise about the laws of nature and "facts about the remote past." This is called "the thesis of determinism." This argument is supposed to show that, if determinism is true, no one "can do otherwise." So, on libertarian assumptions, they reason that if determinism is true, then Sam cannot do otherwise than he does. What do they mean here? That Sam can't do otherwise in a world with *different* laws and *different* propositions about the remote past? Of course not! Sam possibly *does* do otherwise. No, they mean that Sam can't "do otherwise" in *all* possible worlds with the *same* laws and the *same* remoter past. If it really sufficed *for the libertarian* that we do otherwise in a world with a past different from ours, or laws different than those in the actual world, then the compatibilist has an *easy* response. He does X in @ but ~X in W, where W ≠ @ (up to the choice)!

      b. Consider the "luck objection." This objection, as even Robert Kane puts it, is that libertarianism *requires* "doing other wise" in "the exact same circumstances". Indeed, this is why van Inwagen thinks free will is a mystery. If LFW *didn't* require doing otherwise "in the exact same circumstances" (cf. his remarks about "rewinding the world" a million times), there'd be no luck objection!

      Delete
    10. Part deux

      Third, on the last point: I'm asking at a time *before* t10, how does the agent do otherwise than sit at t10 given the pre-punishment at t1–9? If Linda has LFW at t10, then she can either sit or stand at t10, right? So, at t1–9 she has been pre-punished for sitting at t10. This *depends on* her sitting. She's "the source." She's "the truthmaker" for God's pre-punishing her. Everything you've said I grant. How does she do otherwise at t10? You say, "No, it's enough that there's *another* world, a world where she doesn't sit at t10, and thus isn't pre-punished at t1-9, for her to have LFW." Thus, it seems that you reject this:

      (FP) For any action Y, agent S, and time t, S can perform Y at t only if there is a possible world with the same “hard” past up to t as the actual world in which S does Y at t.

      Is this right? You deny (FP)?

      Second, you deny what Ginet, PVI, Plantinga, et al say about our freedom , viz., that it is the freedom to add to the given past.That is, consider the world where Obama was elected in 2008. Now, suppose you're deciding upon whether to sit or stand at t, where t > 2008. Intuitively, it seems that if you *can* either sit or stand at t, then your power is just this: to make the world-where-Obama-was-elected-in-2008 now, i.e., at t, include your sitting or your standing. Do you deny this too? That is, do you deny that this is how we should understand S's libertarian freedom, i.e., as a power to add to the given past in the above described way?

      Delete
    11. No, a choice is determined if it is caused in a way that is not compatible with any other choice. A particular choice can be determined even if determinism is false. If a neuroscientist puts a chip in someone's brain to control their choices, that would make none of their choices free, and that's so even if determinism is false.

      It occurs to me that the time travel case, the pre-punishment case, and the prophecy case do not involve hard facts if you define those the way you are. In the time travel case, there's someone walking around before I make my choice to time travel who has a whole bunch of memories duplicating mine and then some additional once including the choice to time travel. If you just look at intrinsic facts, there's nothing there to show that I will time travel, because it might just be someone who appeared out of nowhere looking like me and having memories that include mine plus more. In the pre-punishment case and the prophecy case, atemporalism actually contributes something to make it similar. If God isn't at that time, then the only thing that makes it pre-punishment is a cause outside time and space. So you have someone experiencing effects that in fact are caused by God choosing to pre-punish the person, but the only thing that makes it pre-punishment is not intrinsic to anything at that time. The same goes with a piece of prophetic literature or a prophetic pronouncement. What makes it genuine prophecy is an infallible connection with something at another time mediated by an infallible God outside time. Nothing at that time requires that the pronouncement even be true. So the things pointed to as hard facts at that time are not hard facts that guarantee their truth. The only thing that guarantees their truth is not intrinsic to that time and therefore would only be soft facts. So on further reflection I would say that the Occamist shouldn't worry about these cases at all.

      There is a difference between the libertarian view I was outlining and compatibilism as you discuss in (a). Compatibilism says that I'm free even if there's no possible world consistent with the actual causes of my choice where I do otherwise. The libertarianism that allows for pasts that wouldn't be present in the alternative possibility would still insist that I'm not free if there's no possible world consistent with the actual causes of my choice where I do otherwise. So that libertarian approach still isn't giving compatibilists what they want.

      The same goes for (b). What I'm saying is that libertarianism can allow for possibilities to serve as relevant alternatives even if there's something in the actual past that would be different in the relevant possibility, as long as that something isn't part of the circumstances that would influence the choice. So it depends on how you delineate what it means to be the exact same circumstances. I was saying what should matter for libertarianism is not the entire past but just the past that is relevant to the choice in question.

      But it turns out I don't think I need to say any of that, because the supposed hard facts turn out not to be. I hadn't thought through the implications of Lewis' paper on time travel (as applied to this case) as fully as I should have. (Or rather, I was trying to take something from that paper that I didn't need because there was already something else in it that did it in a much easier way.)

      Delete
    12. I would certainly deny (FP), because I'm not a libertarian. And the kind of libertarianism I was outlining would deny (FP) and replace it with something about the parts of the hard past that are relevant to the choice, which is what would distinguish it from compatibilism. But I don't think libertarianism needs to do that, because the punishment can take place, and be caused by God as punishment for the future act, without any of those elements being part of the hard past. The only thing that's part of the hard past is the fact that the effects of the punishment occur. And those, for all the intrinsic facts tell us, might not be caused by a choice of the atemporal God to punish for a future act but by something else.

      No, I certainly don't like conceiving of libertarianism as a power to add to the given past. That captures what growing block libertarians would want to say, but it doesn't fit well with libertarianism in any other view of time. It doesn't even work well with presentism, because there is no past to add to if presentism is true. It certainly doesn't work well if the future is ontologically real. And since there are plenty of libertarians who think the future is real, it's not a good idea to try to define libertarianism in a way that precludes those views if you can define it without doing so. That would be question-begging in any debate among libertarians about the nature of time.

      Delete
    13. "a choice is determined if it is caused in a way that is not compatible with any other choice."

      What? A choice might be indeterministically caused in a way that is not compatible with any other choice, yet not determined (by definition). Indeterminism can be a threat to free will too.

      "It occurs to me that the time travel case, the pre-punishment case, and the prophecy case do not involve hard facts if you define those the way you are."

      Wrong. Take pre-punishment for example. Linda sits in jail for 10 hours 10 days ago days. She was flustered. Missed a doctor appointment. Was scared. Had to eat slop, etc. These are as hard a fact as any! So you'd have to say Linda can make it the case at t10 that she never sat in jail—and all that involved—at for 10 hours at a time prior to t10! But that's absurd.

      So perhaps you'll want to say that Linda makes it the case that, though she sat in jail,she wasn't really *punished* because she didn't sit at t10. So, whatever she was doing in jail, she wasn't undergoing *punishment*, since she's innocent. But this commits you to a view that no one is ever punished for a crime they didn't commit, which is a hard saying.

      So perhaps you'll want to reconsider the first answer. It's not so absurd after all. You may then want to say,you have a choice about whether you spent 10 hours in jail 10 days ago. For, if you underwent such activities, your having undergone them depends on what you do at t. Todd says this in response:

      "I think something is clearly amiss here. The proponent of the Third Answer here contends that whether Jones spent 10 hours in jail 10 days ago depends on whether Jones sits at t. But my contention is this. In whatever sense it might be true that whether Jones spent 10 hours in jail 10 days ago “depends on” whether he sits at t, this sense is obviously irrelevant to the question of what is within Jones’ control at t. Imagine that we have just witnessed Jones spending 10 hours in jail. And imagine saying to someone, “Jones just spent 10 hours in jail.” And imagine him replying, “Well, it depends. Whether he was in jail for those 10 hours depends on whether he sits at t.” The proper reply to such a suggestion is this:

      No, whether Jones was in jail just now obviously doesn’t depend on whether he sits at t – it clearly doesn’t depend on anything at all that happens in the future, since, as you can see, Jones already has spent 10 hours in jail. What’s still left to be decided about his having been in jail? There’s nothing here that still ‘depends’. Are you perhaps thinking that whether his having been in jail was just still depends? I happen to know that it was just, since I happen to know that God was the one punishing."

      Cont. below

      Delete
    14. Cont. from above:

      "The libertarianism that allows for pasts that wouldn't be present in the alternative possibility would still insist that I'm not free if there's no possible world consistent with the actual causes of my choice where I do otherwise."

      This points up another problem in (what seems to me as) your idiosyncratic view of LFW. You're focusing too much on *causes*. You seem to think certain *causes* are the only things the libertarian needs to be scared of. Put things in terms of "control." Nefarious (or, deterministic) causes show that you are *controlled*, and thus not free. Freedom crucially rests on the notion of control. It seems intuitive that if you are not in control of your actions, then you are not free. Theological compatibilists (here: those who believe libertarian free will is compatible with God's foreknowledge) seem to think that the only way you can not be in control of your actions is if someone or something else controls you. Thus, you reason (as we see above) that since God's foreknowledge doesn't "control" (nefariously cause) humans to do what they do, then the foreknowledge argument doesn't show that humans don't have libertarian freedom. But this is to assume that the only way to lack control is if someone or something else controls you. However, this isn't so. Consider the case of being **out of** control. This could occur even, or especially (!), in a totally random, accidental world. In this case, nothing outside you would be controlling you; however, you would be out of control. The foreknowledge argument intends to show that your future actions are **out of** your control. This doesn't logically imply that foreknowledge argument is trying to show that your actions are controlled (caused). See the difference?

      And this leads into a response to your response to (b). When you say libertarians can exclude things "as long as [those things aren't] part of the circumstances that would influence the choice," the scope of things you focus on is clearly problematic as we can get threats to LFW from things that don't "influence" (whatever that means, anyway) a choice! The foreknowledge argument intends to show that you don't have the relevant control over your future actions, but that's *not* because foreknowledge "influences" your actions. Even Plantinga recognizes this in Ockham's Way Out, when he notes that, going back to Edwards, the FK argument wasn't intended to show the FK *makes* future actions necessary, it's intended to *show* that they are *given* the prior premises (just like *given* the prior premises in the consequence argument, for example).

      Cont. below

      Delete
    15. Cont. from above

      "And the kind of libertarianism I was outlining would deny (FP) and replace it with something about the parts of the hard past that are relevant to the choice, which is what would distinguish it from compatibilism."

      That's simply an odd and idiosyncratic understanding of libertarianism. Who holds to that? Moreover, the comment about holding fixed only the "relevant" parts of the past is unhelpfully ambiguous and arguably question begging. If we're going to go this route, then what's the problem with compatibilists—like, e.g., Vihvelin, i.e., new dispositionalists—denying that laws or facts about the past are "relevant"? Moreover, how can you say what's "relevant"? Perhaps your prior beliefs determine future actions. Perhaps your genetics do. Perhaps inscriptions by God on the sides of houses do? The only way to see if these don't rule out LFW is to hold them fixed.

      Here's what Ginet says:
      Given any truth entirely about the past, b, if I now have it open to me to make true a certain proposition about the future, a, then I now have it open to me to make true the conjunction of b and a. If I have it open to me now to make the world contain a certain event after now, then I have it open to me now to make the world contain everything that has happened before now plus that event after now. We might call this the principle that freedom is freedom to add to the given past or the principle of the fixity of the given past.

      I don't see how this would be problematic with presentism. It's not saying the past has to "exist". In any case, we can, it seems, run the same argument off presentism's "quasi-truths" about the past. There are past-tensed truths consistent with presentism. But you need to pick a goalpost to defend. You want to act as if timelessness is irrelevant, but then you crucially appeal to it. You want to defend Ockhamism. but then you conjoin it with presentism. On this latter unholy mixture, you may want to read (if you haven't already, then you may want to offer a counter argument) Presentism and Ockham’s Way Out, Oxford Studies in Philosophy of Religion 1 (2008): 1 - 17 (Alicia Finch and Michael Rea), where they authors argue that one can't be both.

      The "power to add to the past" thesis is a principle held to by many libertarians. Indeed, that's who I got it from. I have libertarians who hold different positions regarding time who endorse it. In any event, though it's more than I want to get into now, I find the moves you seem forced to have to make to defend compatibilism bout F&FK to simply illustrate the *costs* of defending libertarianism in light of foreknowledge. You seem to be advancing positions about libertarianism that no libertarian has posited.

      Delete
  5. Listening to Craig explain and defend Molinism, he uses language that suggests that he thinks of it as a likely explanation, but that he really doesn't know for sure if it's true. To be fair, Helm doesn't seem much more about his alternative.

    What seems to me to be missing in Craig's defense of Molinism is an answer to total depravity. Molinism has merit to some degree. But it breaks down where we understand that our decisions are mitigated by the work of the Holy Spirit restraining us.

    ReplyDelete
  6. Jeremy says:

    "On the metaphysical possibility issue, the point is that it is metaphysically possible for someone with LFW to do otherwise. In the world where they do otherwise, God knows that the other thing is what they do. What God knows depends on what they do. That's what it means for a truthmaker to be the act and not God's knowledge of it. I'm not sure why you think God's knowledge that p in the actual world means God would know something false in a world where p is false. Then it wouldn't be knowledge."

    A friend of mine responded as follows:

    Jeremy makes the obvious point that an infallible God's beliefs track truth across possible worlds. But the real issue is how we specify the surrounding circumstances of the agent's choice. At t1, at the moment just prior to Jones's choice to mow the lawn, it is true to say that God timelessly and immutably believes that Jones will mow the lawn. Of course, a timeless God doesn't have these beliefs at times. But nevertheless the divine belief is as much a part of Jones's 'environment of choice' at t1 as the reality of numbers, uninstantiated properties, and possible worlds themselves. It would be really weird to say that at t1 Jones has the power to make it the case that the number 2 never existed. So why think that at t1 Jones has the power to make it the case that a timeless divine belief never existed? This is the rebuttal to the Boethian solution at its strongest. It appeals to our intuition about immutable, timeless reality being both actual and fixed, even as the temporal past is actual and fixed. Against such a reality, Jones's freedom doesn't have a chance.

    To respond to this rebuttal, Jeremy must somehow eat away at the analogy between eternity and the temporal past. As I hinted at in yesterday's email, his best shot here is to somehow make the divinely doxastic aspect of the eternal realm to be 'dependent' upon Jones's temporal choice, in a way that the actual past is never dependent upon the future. OK. But while this is a past/eternity disanalogy, is it disanalogous in the relevant respect, such that it salvages the Boethian solution?

    Let's distinguish between two aspects of belief: the event of belief, and the truth-value of belief. Let's stipulate that the truth-value of God's eternal beliefs depends upon Jones's choice being what it is. Call this 'semantic dependence'. Such dependence seems clear. But in what sense is the event of God's belief dependent upon Jones's choice? Does Jones's temporal choice bring about or cause a timeless event in the eternal realm? That seems inconceivable. Talk of 'truth-makers' here is otiose, since it is propositions that have truth-makers, not events, and we're talking about infallible divine beliefs as events. Events are either caused or they are not.

    ReplyDelete
  7. Cont.

    In his reply to Gibson, Jeremy says that, "What God knows depends on what they do." But this is just semantic dependence, since 'what God knows' are propositions, and on a correspondence theory of truth of course the truth of a proposition depends upon the obtaining of the state of affairs the proposition is about (or upon factors that entail or determine that obtaining). But what about God's belief itself? That is, for lack of a better term, a psychological event in the doxastic life of God. What brings that about? Little old Jones in time? In this respect, eternity seems as causally inaccessible to Jones as the temporal past, and talk of semantic dependence misses this point.

    So one can construe the 'dependence' of the eternal realm on the temporal realm in a number of ways. The more plausible ways seem irrelevant to solving the actual freewill/foreknowledge problem, whereas the less plausible ways may have a shot, except that they're hard to understand once you clarify that divine beliefs are events. The best shot for Boethians, in my present understanding, is to go for something like counterfactual power over the past as the relevant kind of dependence. But since this option is equally open to temporalists about God, going that route seems to make Boethianism irrelevant.

    ReplyDelete
  8. Jeremy Pierce said:

    "But the answer given centuries earlier by Carneades is all you need, and it simply requires two premises: (a) the B-theory of time and (b) truthmakers about future contingent acts are at the time they take place, not necessarily at the time you know them."

    i), it does seem that to the extent you rely on the B-theory to get out of the problem, your solution becomes less relevant to those who reject the B-theory for various reasons. For instance, since Craig thinks the B-theory involves an implicit denial of creation ex nihilo, he would think the price for adopting the theory is too high.

    ii), shoving the truthmakers into the future just seems to push the problem back (err, forward? ;-) If we're wondering how God can infallibly know propositions about future events, it won't help to be told that he simply has to know the future events that make the propositions true.

    ReplyDelete
  9. It's not part of God's necessary existence what God's contingent relations to the world are. That God is the creator at all is a contingent fact, since God didn't have to create. If God hadn't created, God wouldn't have known that George Washington is the first U.S. president, since he wouldn't have been. Your argument against contingent divine knowledge would make all facts God knows necessary, and thus either every fact is necessary or God isn't omniscient. The standard solution to that problem is to allow that God's existence and nature are necessary, but God's relations to the world, including God's knowledge of contingent truths, are contingent matters.

    As I've said, Boethian eternality is irrelevant to the compatibility problem. It's better an one possible answer (though not the only one) to the problem of how God has access to the information.

    Craig says the B-theory is inconsistent with creation ex nihilo, but that's a patently uncharitable accusation. According to the B-theory, creation ex nihilo would simply mean that there wasn't anything before creation. So call the first moment of time the Big Bang. There wasn't anything before that. God's act of creation caused that. Therefore, it's creation ex nihilo. Only if you assume the B-theory is saying something it doesn't say (as virtually every argument against it that I've ever heard does) can you say such a silly thing as Craig does. Since I have no patience for silly and uncharitable arguments, I don't worry about such things. It's more of an excuse not to accept the deliverances of science and to recognize the incoherence of his own position than an actual argument. I'm not trying to convince Craig, just to put forward a sound argument.

    Your final point strikes me as again mixing up the two separate foreknowledge problems. Truthmakers in the future does answer the compatibility problem. It doesn't explain how God knows the future. You need to say more to explain that. Atemporality with immediate access to every moment in time would do it. A temporal God would have to have some other kind of access to times not yet, but I'd be reluctant to imagine a spaceless God looking through wormholes or something like that. But it might be some natural law that governs the entire universe the somehow involves infallible transfer of information backward through time for God to see. As I said, I do think the Boethian answer is the best one. The ones compatible with a temporal God involve strange things. But it's not fair to the solution to the first problem to complain that it doesn't also solve the second problem, which is a problem I never put it to solve, since I don't think it does so.

    ReplyDelete
    Replies
    1. Jeremy Pierce:

      "Your argument against contingent divine knowledge would make all facts God knows necessary, and thus either every fact is necessary or God isn't omniscient. The standard solution to that problem is to allow that God's existence and nature are necessary, but God's relations to the world, including God's knowledge of contingent truths, are contingent matters."

      To which my friend replied:

      I'm not making an "argument against contingent divine knowledge." Why would I do a think like that? :-) It seems clear to me that Jeremy just doesn't have conceptual space for the category of accidental necessity, much less for an analogy to this in the eternal realm. Of course many of God's beliefs are *metaphysically* contingent in their existence and truth-value. The question is whether *Jones* has power to prevent an eternally actualized and immutable realm. Accidental necessity is now-necessity (as Zagzebski puts it), or powerlessness-to-prevent (as Fischer puts it). It's a temporally relative notion indexed to the agent's moment of choice. What is now-necessary at some times is not now-necessary at other times.

      I agree that Craig's accusation against the B-theory is uncharitable.

      "Your final point strikes me as again mixing up the two separate foreknowledge problems."

      I'm not mixing up the problems. I'm explicitly distinguishing between two things: the semantic dependence of the truth-value of God's beliefs upon a (future) truthmaker, and the causal dependence of an eternally actualized event upon a temporal choice. Given that the temporal choice is made in circumstances that *include* this eternally actualized event, is it open to the temporal agent at the time of choice to *prevent* the actualization of the eternally actualized event? That seems quite implausible, even if the fact of semantic dependence is as plausible as could be.

      Delete