A sound argument against Open Theism
December 5, 2008 — 14:04

Author: Alexander Pruss  Category: Open Theism  Comments: 29

The following argument is sound:

  1. Tomorrow I will freely eat dinner. (Premise: I have good albeit fallible inductive knowledge of this)
  2. God infallibly knows every true proposition that it is (metaphysically) possible to infallibly know. (Premise)
  3. If God infallibly knows p and God infallibly knows q, then God infallibly knows everything entailed by the conjunction of p and q. (Premise)
  4. Possibly, God infallibly knows that I will eat dinner tomorrow. (Premise)
  5. Possibly, God infallibly knows that tomorrow I will eat dinner freely or not eat dinner. (Premise)
  6. I will eat dinner tomorrow. (By (1))
  7. God infallibly knows (6). (By (2), (4) and (6))
  8. Tomorrow I will eat dinner freely or not eat dinner. (By (1))
  9. God infallibly knows (8). (By (2), (5) and (8))
  10. (6) and (8) entails (1). (Conceptual truth)
  11. God infallibly knows (1). (By (3), (7), (9) and (10))
  12. If Open Theism is true, God does not infallibly know anything I will freely do. (Premise)
  13. Open Theism is false. (By (11) and (12))

An Open Theist (OT) who thinks that there are no contingent truths about the future will deny (1). But (1) is something I have very good evidence for. (Could the OT say that probably I will freely eat dinner? Not if the OT thinks she knows that (1) is not true, since if one knows that p is not true, one can’t rationally and sincerely say that p is probably true.)

Perhaps an OT will deny (4) and (5). But there are possible worlds where God now promises me unconditionally that I will eat dinner tomorrow, and, if I choose not to eat dinner tomorrow freely, he forces me to eat it, in fulfillment of that promise. In those worlds, God knows (6). Similarly, there are worlds where I exist where everything turns deterministic, and God unconditionally promises not to intervene in the deterministic stuff, and God knows that I will eat dinner tomorrow–I will eat it unfreely, but (6) does not say I will eat it freely. So (4) is true. And similarly, so is (5). For there are worlds where God unconditionally promises, or otherwise necessitates it to be the case, that if I don’t freely eat dinner, he won’t allow it to happen that I eat dinner unfreely. In those worlds, God knows that I will either freely eat dinner tomorrow or I will not eat dinner at all tomorrow. Hence, (5) is true.

There is a way in which the argument is unfair or even silly. It presupposes a literalistic reading of (2) which is contrary to the spirit in which the OT might have assented to (2). The OT might say that God knows every proposition p that is known in a world where p holds for exactly the same reason as it holds in the actual world. To work out the details of this proposal is not an easy task. One worry is that the account of omniscience embodied in (2) might not be strong enough, once one tightens up the notion of knowability like that.

Comments:
  • Jonathan Jacobs

    Alexander,
    Is it really irrational to assign a probability to a proposition that lacks truth value? Sure enough, the proposition is not true. But it’s also not false.
    If it were irrational, then wouldn’t any view that assigns no truth value to propositions about the future be obviously false?

    December 5, 2008 — 22:25
  • Tom

    Alex: An Open Theist (OT) who thinks that there are no contingent truths about the future will deny (1). But (1) is something I have very good evidence for. (Could the OT say that probably I will freely eat dinner? Not if the OT thinks she knows that (1) is not true, since if one knows that p is not true, one can’t rationally and sincerely say that p is probably true.)
    I’ll bite. ;o)
    I’d be among those who deny (1). On a Piercean semantic, if it’s true that you “will” eat dinner tomorrow, then you don’t do so freely. But if you choose to do so freely, then it’s false at all times prior to say that you “will” do so (and equally as false to say that you “will not”). With respect to future contingent events, both “will” and “will not” props would be false (not ‘neither true nor false’). We prefer not to deny either bivalence or excluded middle. The prop that would truthfully describe the future here would be “I might and might not eat dinner tomorrow.” That would be the prop God knows.
    As I understand it, ‘contingency’ is a property of the objective world of becoming and not of ‘truth’ or ‘propositions’ per se. A proposition (assuming bivalence) is either true or false. So I don’t speak of “contingent truths about the future.” There’s nothing contingent about the proposition or its truth value at the time the claim is made. It is what it is in the actual world. I speak rather of “truths about the contingent future.”
    And an open theist could easily agree that “I ‘will’ eat dinner tomorrow” is, strictly speaking, false, while it’s true that I’m more likely to eat dinner tomorrow than not to eat. There’s room for probability within the open view. We argued this in F&P 23 (2006): 432–459 (http://www.alanrhoda.net/papers/opentheism.pdf). Again, on a Piercean semantic, one could construe as follows:
    ‘will’ = ‘with a probability of 1’
    ‘will not’ = ‘with a probability of 0’
    ‘might’ = ‘with a probability greater than 0’ (thus the contradictory of ‘will not’)
    ‘might not’ = ‘with a probability less than 1’ (thus the contradictory of ‘will’)
    On the standard square, ‘will’ and ‘will not’ occupy the top left and right corners (respectively) as contrary props and ‘might’ and ‘might not’ occupy the bottom left and right corners (respectively) as subcontrary props. As all know, contraries may both be false, in which case the subcontraries are both true. And this is precisely where contingency with respect to the future finds propositional expression, in the conjoining of subcontraries “might and might not” (= ‘with a probability greater than 0 and less than 1’). That just is the truth an omniscient God would know. (Hartshorne followed Pierce in this as well.)
    Open theists disagree among themselves as to whether ‘probability’ ever or sometimes fits in. I think it can. Some indeterminate event (probability greater than 0 and less than 1) maybe have a probability of, say, greater than .5 and less than 1, in which case it’s “likely” to happen but still indeterminate. The question is whether the probability range that define ‘indeterminately’ (viz., greater than 0 and less than 1) represents a single block or whether it can be traversed. I suppose it depends on the nature of the objective world of becoming. I think it’s objectively the case that the world sometimes ‘tends’ to particular outcomes. But the important point is that the ‘contingency’ and ‘probability’ spoken of here are both objective features of the world’s becoming and not limitations on our ability to know what the truth really is.
    Tom

    December 6, 2008 — 0:23
  • Jonathan:
    If p entails q, then P(p) is no bigger than P(q). Now, for all p, p entails true(p). Therefore, P(p) is no bigger than P(true(p)). If the latter is zero, so is the former. 🙂
    But, yes, there is a really quick refutation of any view on which p has a probability and no truth value. I think Richard Gale and I made this point, maybe even in the same context, in our introduction to the Existence of God anthology.
    Tom:
    Consider December 7, 2007. Let’s suppose I don’t remember anything about that day. But I have strong inductive evidence that I ate some form of dinner on that day: I am just the sort of person who eats dinner every day. My evidence that I ate dinner on that day could well be no better than my evidence that I will eat dinner tomorrow. But, nonetheless, I have strong evidence that I ate dinner then, strong enough to justify me in affirming the claim. Likewise, then, my evidence about eating dinner tomorrow would be sufficient to justify me in affirming the claim, since it is no weaker.
    Minor point on probabilities: Probability 0 is not at all the same as “will not”. It is possible that an event has probability 0, but will happen. Suppose a fair coin will be flipped infinitely many times in the next minute. The particular infinite sequence of heads/tails that it will take has probability 0. After the fact, it will be correct to say: It had probability 0 of showing this sequence (HHHTHTTHTHTHHTTTHHTHTHHTHTHTHTHTTTTTTTTTHHHTHTHTH…) but did.

    December 6, 2008 — 9:49
  • Jonathan Jacobs

    I suppose I’ll need to look at the quick refutation. But one question: If V(p) is undefined, why must P(true(p)) be 0? (For the quick argument you gave to work, P(true(p)) must be 0.)
    Prima facie, it seems too easy a refutation of some forms of presentism. Current science seems able to assign probabilities to future events. Hence, if you’re right, then some forms of presentism should be *obviously* false. But, or so it seems to me, they aren’t *obviously* false.

    December 6, 2008 — 11:15
  • Tom

    Alexander: Consider December 7, 2007. Let’s suppose I don’t remember anything about that day. But I have strong inductive evidence that….
    I’ll probably (no pun intended) describe this poorly, but if I follow you, this seems to me to mistake the property of an epistemic state (how ‘certain’ I am, given the evidence, of what a past event ‘was’ or what a future event ‘will’ be) for the property (i.e., truth value) of propositions describing (in this case) the past and the future. You may be justified, given the evidence, in being as certain about the future as you are about the past, but that hardly places the past and the future on equal footing ontologically speaking. And since the truth value of props describing the past and the future are grounded in something other than how certain we are, I don’t see how being as certain of what a future event will be as I am about what a past event was justifies the claim that I can know the outcome of future contingent events the way I know the outcome of past events, for while there is nothing in fact indeterminate about what occurred in the past, there may indeed be much that is indeterminate about which possibilities unfold in the future.
    Alexander: Minor point on probabilities: Probability 0 is not at all the same as “will not”. It is possible that an event has probability 0, but will happen. Suppose a fair coin will be flipped infinitely many times in the next minute.
    One can complete an infinite number of coin tosses in a finite span of time?
    Alexander: The particular infinite sequence of heads/tails that it will take has probability 0. After the fact, it will be correct to say: It had probability 0 of showing this sequence (HHHTHTTHTHTHHTTTHHTHTHHTHTHTHTHTTTTTTTTTHHHTHTHTH…) but did.
    After which fact? How do we get to the other side of having traversed an infinite series to speak of its outcome as a “fact”? Not sure I’m following you. But in any case, this particular outcome (if we may speak of it) would ground the proposition that posits the truth of what ‘occurred’. But it wouldn’t (as some open theists would argue) in turn ground the truth of a proposition positing this particular outcome prior to beginning the coin tosses. That you begin be describing the outcomes as that which ‘will’ occur is the point of debate. That it ‘did’ occur, we’d argue, doesn’t mean it was true before it occurred that it ‘will’ (or ‘would’) occur. All that its being true that it occurred requires prior to its occurring is that it ‘might and might not’ occur.
    I think I can agree that the outcome of an infinite series doesn’t qualify as a candidate for probabilistic attribution. An infinite series wouldn’t have an “outcome” on which we look back “after” it’s traversed, right? In any event, we’re talking about particular (finite) events, like the outcome of a toss of a coin or a finite series of tosses.
    Don’t want to abuse the privilege of posting as a non-member. Thanks all!
    Tom

    December 6, 2008 — 11:40
  • Mike Almeida

    Minor point on probabilities: Probability 0 is not at all the same as “will not”. It is possible that an event has probability 0, but will happen.
    That’s trickier than it sounds. Most people would not assign 0 to anything but a contradiction or, more broadly, an impossible proposition. I certainly wouldn’t. Call a probability function regular just if it assigns 1 only to necessary propositions, 0 only to impossible propositions. Lot’s think that you’re rational only if your probability function is regular. A. Shimony shows somewhere that rejecting regularity makes you susceptible to a dutch book. But Appiah, Carnap, Lewis all endorse something like regularity as a condition of rationality. Suppose you (Alex) endorse regularity. In that case, you’re committed to saying in the infinite case that an p is impossible and could happen. Contradiction. Suppose you reject regularity. In that case you’re assigning 0 to propositions that are not impossible. No surprise that a proposition that is both possible and assigned 0 might happen. What is the best route to take? I say in cases of infinitely many equiprobable outcomes, the only option is to assign each infinitessimal positive probability.

    December 6, 2008 — 12:21
  • Clayton Littlejohn

    I confess that I’m not an open theist, but I don’t yet see what’s wrong with the following sort of response. (That is to say, once you’ve bought into open theism, I can’t see what’s wrong with saying this in response to your argument.) Let’s distinguish:
    (a) Necessarily, if there is some possible world, w, in which p is known infallibly, God knows p infallibly.
    (b) Necessarily, if there is some possible world, w, in which p is known infallibly, God knows p in that world infallibly.
    I don’t see that these are equivalent and while an open theist might grant (b), I don’t see why they’d grant (a). It looks like your argument needs (a) rather than (b).
    Suppose in @ that (1) is true. I suppose an open theist might say that this assumption is inconsistent with the conjunction of two further claims:
    (4′) In @, God infallibly knows that I will eat dinner tomorrow.
    (5′) In @, God infallibly knows that tomorrow I will eat dinner freely or not eat dinner.
    If we work from the supposition that I freely eat dinner in @, then prior to the decision to eat dinner there is nothing in the universe that would constitute infallible grounds for the belief that I shall freely eat dinner in @. If the conjunction fo (4′) and (5′) is true, however, then the infallible grounds available to God ensure that I eat but don’t do so freely.
    So, why can’t the OT just say that in any world in which I freely A, that’s not a world where God has infallible knowledge of my freely A-ing, but in worlds where God has infallible knowledge of my A-ing, I don’t A freely? This gives them the right to assert that (b) is true, that seems enough to uphold the idea that God is omniscient, but they can deny (a) and say that your argument commits a modal fallacy.

    December 6, 2008 — 12:24
  • Mike Almeida

    (a) Necessarily, if there is some possible world, w, in which p is known infallibly, God knows p infallibly.
    Clayton,
    (a) tells us that it is true in every world w that if p is known infallibly in some world w*, then God knows p in w. The problem, of course, is that (unless you assume that God knows infallibly only necessary truths) p might be false in w even if it is known infallibly in w*. So God would not know p in w, contrary to (a). I don’t think (a) can be true. Does Alex’s argument depend on this?

    December 6, 2008 — 12:39
  • Jonathan:
    What is V(p)?
    I was thinking in the following way in my quick argument. Let’s say we know for sure that no future contingents have truth value. Then, if p is a future contingent, the epistemic probability of true(p) is zero. But p entails true(p). Hence, if we are to be consistent, we must assign zero epistemic probability to p.
    This does not make it obvious that presentism is true, but it does make it obvious that the combination of indeterministic probabilistic science, scientific realism and open future is untenable. It is unsurprising that open future views are incompatible with science, since open future views damage induction. (Quick argument: If there is a crucial difference between facts about the future and facts about the past, then induction from past to future is weakened. See this post.)
    Clayton:
    Of course, I want (a), since (b) is trivial. Why should the open theist accept (a)? Because it expresses an intuition about the unsurpassability of God’s knowledge. Also, because one might (incorrectly–Jon Kvanvig has a nice paper on this) think there is a parallel between omniscience and omnipotence, and think that the right account of omnipotence is that God can do whatever is doable.
    Mike:
    Infinitesimals aren’t going to help you in all cases. There is no coherent way of assigning even an infinitesimal probability in infinite lottery cases. (See Section 3 of this paper draft which has appeared in Philosophia Christi.) Moreover, one would need an infinite hierarchy of infinitesimals. For the problem re-appears on the infinitesimal level.

    December 6, 2008 — 12:59
  • Tom:
    Visitors are always welcome, and your thoughtful comments are especially welcome.
    I am untroubled by the possibility of an infinite number of events in finite time. But if it bothers you, imagine a random process whose outcome is a point particle appearing on some point in a square in space, with all points being equally likely. Then the probability that the particle would appear just where it did was zero.
    As for the epistemic and the ontological, I am not sure there is a confusion. If I have very strong evidence for p, then I should believe p. And if I should believe p, I should likewise be willing to say p is true. I could be wrong. The defender of an open future who does not believe propositions about future contingents have truth value has to say that we are always wrong when we believe a proposition affirming that a future event will happen, except in the (rare) case when we’re dealing with a necessary event. This is a seriously problematic bit of anti-realism.

    December 6, 2008 — 13:07
  • Mike Almeida

    Alex,
    (1) I don’t see how (a) could be true for reasons given above, here Mike Almeida | December 6, 2008 12:39 PM | Reply.
    (2) If you do not use infinitessimals and reject regularity, you’re open to a dutch book. So assigning 0 to propositions that are not impossible doesn’t work, either.

    December 6, 2008 — 13:20
  • Clayton Littlejohn

    Hey Mike and Alex,
    Now I think Alex wanted (a) because he said so, but _now_ I think he won’t want (a) because of what Mike just said.
    There might be something in between (a) and (b) that gets things just right. My worry is that things in the neighborhood of (a) have God believing infallibly that p in worlds where it seems the OT will say at best there are fallible grounds for believing p. That leads to something in the neighborhood of Mike’s worry. Maybe (b) isn’t good enough for real omniscience, but like I said, I’m not a defender of OT, I’m just trying to imagine what they’d say in response to the argument.
    Why would God have something less than unsurpassable knowledge if God doesn’t fulfill the conditions specified in (a)? Maybe this is a way of bringing out the idea that (b) doesn’t quite capture the intuition that God’s knowledge in unsurpassable and expresses Alex’s worry. I suppose this might be a worry about denying (a) and going with (b). If Alex knows (1) but God believes only that which is known infallibly, God won’t believe what Alex knows. Thus, Alex knows something God doesn’t. I suppose that’s kind of bad. (I vaguely recall there being a problem of this sort that arose for Hoffman and Rosenkrantz’s account of omniscience, but the details are foggy.)

    December 6, 2008 — 13:30
  • Tom

    Alex: The defender of an open future who does not believe propositions about future contingents have truth value has to say that….
    I’m also unconvinced by those who attempt to defend open theism by denying that propositions about future contingents have truth value. But then open theism doesn’t require such a move.
    Tom

    December 6, 2008 — 13:50
  • Alex,
    Thanks for the interesting and challenging argument. It is, though, somewhat tendentious to call it “sound” since no open theist is going to grant you all of the premises. At any rate, here’s why I’m not persuaded.
    Consider, first, the semantics of the future tense operator “will”. On a Peircean account it simply can’t be true of a future contingent that it “will” happen, hence (1) is self-contradictory and therefore false. On an Ockhamist account, however, the truth of (1) depends solely on whether tomorrow you do eat dinner freely or not. If you don’t, then (1) is false. If you do, then (1) is true, but if future events don’t exist, as open theists and other antirealists about the future believe, then there is no event of your eating dinner tomorrow, and thus nothing to ground the truth of (1). Arguably, then, (1) is not true even if the Ockhamist semantics be granted.
    Second, it is questionable whether your evidence really gives you strong reason to believe (1). On the one hand, the more strongly we take your inductive evidence to support your eating dinner tomorrow, the more habitual or settled we must take your behavior to be, and hence, the less reason we have to think that you are still “free” (in the libertarian sense) with respect to whether you eat dinner tomorrow. On the other hand, the more “free” we take your eating dinner tomorrow to be, then less reason we have for thinking that your past behavior is a reliable guide to your future behavior.
    You consider the suggestion that your inductive evidence might support “Probably, (1)” rather than (1). I submit that “probably” is better placed inside the scope of (1), viz., “I will probably freely eat dinner tomorrow”. The modifier “probably” should qualify the tense operator rather than the whole proposition since it reflects the degree to which we should expect you to freely eat (present tense) dinner tomorrow rather than the degree to which we should expect you to be going to freely eat dinner tomorrow.
    Third, your comments at the end about (2) are on target given the way you use it in your derivations of (7) and (9). (4) and (5), which also figure in those derivations, are claims about what God could infallibly know in some possible world or other. The open theist, however, will not construe the sort of possibility referred to in (2) in that way. Rather, the version of (2) that he would grant takes the relevant sense of possibility to include as a constraint all ‘hard facts’ about the actual past.
    Finally, it’s a bit quick to say that zero probability does not imply impossibility. That result follows if we define probabilities on the real numbers and use limits. On non-standard analysis, however, the probability of an infinite sequence is not zero, but infinitesimal.

    December 6, 2008 — 14:04
  • Mike:
    I suppose the Dutch Book argument might go something like this: If I have an infinite lottery with equiprobable outcomes, and I assign probability 0 to each outcome, I should have no objection to a bet where I get zero if that outcome comes up, and where I pay $1 if that outcome does not come up. But then I’d have a Dutch Book against me.
    If that’s the argument, then one way out of it is to deny the following principle: If every bet is individually rationally acceptable, the bets taken together are also rationally acceptable. Here is what I would say about this: This principle holds for finite collections. It holds for countable infinite collections in contexts where probabilities are countably additive (see my Phil Christi paper for a discussion of countable additivity). But it does not hold for countable infinite collections in contexts where probabilities are not countably additive, and in general does not hold for uncountably additive collections.
    Clayton:
    You’re right–I missed the fact that (a) omitted one condition I had in (2)–namely, that the proposition is true. Just add that condition to the antecedent of the conditional in (a), and you’ll get something I affirm.

    December 6, 2008 — 14:10
  • Correction. The bet would be: I pay $1 if the outcome comes up, and get $0 otherwise.

    December 6, 2008 — 14:11
  • Jonathan Jacobs

    Alexander,
    Sorry about that. By V(p) I just meant the truth value of p.
    I didn’t realize you were using epistemic probabilities.
    But I’m still unsure why, if the truth value of p is neither true nor false, we must therefore assign 0 to the probability of p’s being true. (This may be obvious, I admit.)

    December 6, 2008 — 14:40
  • Jonathan:
    Well, if I am certain that the truth value of p is neither true nor false, I will presumably also be certain that the truth value of p is not true. But then I should assign 0 to the probability of p being true, no? At least if we’re talking of that absolute kind of certainty.
    But what if you’re not so certain of the open future view, but merely think it is more like than not? In that case, you had better assign a probability less than one half to p being true, and hence to p. But if p is the proposition that tomorrow I will freely eat dinner, then P(p) is much more than one half, since I eat dinner almost always.

    December 6, 2008 — 15:10
  • Clayton Littlejohn

    Alex wrote:
    Clayton:
    You’re right–I missed the fact that (a) omitted one condition I had in (2)–namely, that the proposition is true. Just add that condition to the antecedent of the conditional in (a), and you’ll get something I affirm.

    That might help with some cases, but isn’t there still a remaining problem? In any world in which you freely eat dinner at t, (i) it’s true that at t you eat dinner and (ii) there isn’t prior to t infallible grounds for God to believe that you’ll freely eat dinner. How then does it turn out on your view that:
    (iii) In every world in which you eat dinner, God believes it.
    (iv) There’s no world in which God’s beliefs about whether you’ll eat dinner are mistaken.
    In in @, let’s say, God believes that you’ll freely eat dinner correctly and does so on the basis of grounds, G1. If in every world in which God has G1, God believes what God believes in @, there’s going to be some possible world in which God believes you’ll freely eat dinner but you don’t eat dinner freely. If it’s not the case that God forms just the same beliefs in all possible worlds given just these grounds, then it seems that you’re going to have to deny what Roger White calls “Uniqueness”:
    Uniqueness: Given one’s total evidence, there is a unique rational doxastic attitude that one can take to any proposition.
    I think it’s sort of bad if an omniscient being were to respond differently to a given set of grounds in different possible worlds when by that being’s lights, there would be no reason to do this.
    So, I can’t tell from what you’ve said thus far, but is part of your response ultimately going to be that God can have infallible grounds for beliefs about libertarian free choices before these choices are made? Otherwise, it’s going to look like a bizarre modal accident that your modified version of (a) is true and God manages to believe correctly in every possible world in spite of having only fallible grounds for some of those beliefs.

    December 6, 2008 — 16:08
  • Clayton:
    It’s a consequence of (2) that God has foreknowledge of free choices. But (2) is very plausible. Hence, it is very plausible that God foreknowledge of free choices.
    How he manages it is a different question. My own story takes into account divine simplicity and is strange (a sketchy sketch is in my OSPR paper on divine simplicity). Perhaps, interestingly, my story is no more strange in the case of foreknowledge than in that of knowledge of the present, since God is outside of time.
    But let me make some additional remarks that do not make my assumption of timelessness:
    1. If we do not limit the evidence to what occurs at present, the argument you give seems not to work–the argument depends on the evidence all having to be present.
    2. Even if we do limit evidence to what occurs at present, why can’t there be present evidence of a future event? After all, there plainly can be present evidence of a past event. If you respond that past events have present effects, I will ask why future events can’t have present effects?

    December 6, 2008 — 17:17
  • Clayton Littlejohn

    It’s a consequence of (2) that God has foreknowledge of free choices. But (2) is very plausible. Hence, it is very plausible that God foreknowledge of free choices.
    (2) might be very plausible (I’ll say more in a second), but I don’t see why an OT would accept it. On its face, it looks as if the kind of infallibilism you accept gives God infallible knowledge of every contingent truth within a world even if God has less than infallible grounds for those beliefs in those worlds. But, that seems odd. I would have thought that if God had but fallible grounds, G1, for believing p, either there’s a world in which God has G1 and believes p even though ~P or there are pairs of worlds, w1 and w2, such that:
    In w1, God has G1, p is true, God knows p.
    In w2, God has G1, p is false, God refrains from believing p.
    If G1 is just God’s total grounds, it seems that God would respond differently to the same grounds _and_ God would know there was no epistemic reason for this differential response.
    If you deny both, then you have God believing things about free actions on the basis of infallible grounds, and I just can’t see why the OT won’t say that when it comes to free actions there cannot be infallible grounds for judgments about (libertarian) free actions.
    Anyway, I still don’t quite know what your view is. I don’t mean to be difficult, I’d like to figure this stuff out, but I suppose I have these questions.
    Q1: Can there be infallible grounds for beliefs about future (libertarian) free actions?
    Q2: Can there be infallible knowledge of future events without infallible grounds consisting of things that obtain in the here and now?
    Even if we limit evidence to what occurs at present and say that this includes evidence of a future event, I don’t see why it would be entailing evidence. But, then I don’t see how on the basis of such evidence one could infallibly know what happens later.

    December 6, 2008 — 17:46
  • Jesse

    With regards to (2) implying that God has foreknowledge of free choices,
    it seems Greogry Boyd’s third scentence in ‘divine foreknowledge: 4 views’ is appropriate:
    “The debate over God’s foreknowledge is rather a debate over the content of reality that God perfectly knows. It has more to do with the doctrine of creation than it does with the doctrine of God”
    So doesn’t this argument need to provide some reason to think that there exists a true proposition about some (in terms of a tensed fact) future, free choice?
    Also, Prof. Pruss says,
    “2. Even if we do limit evidence to what occurs at present, why can’t there be present evidence of a future event? After all, there plainly can be present evidence of a past event. If you respond that past events have present effects, I will ask why future events can’t have present effects?”
    But, again, isn’t the whole idea behind OT the affirmation that the future is open, i.e. objectively indeterminate (at least with respect to free choices)?

    December 8, 2008 — 0:07
  • Jesse:
    “So doesn’t this argument need to provide some reason to think that there exists a true proposition about some (in terms of a tensed fact) future, free choice?”
    Sure: Whether I eat dinner is a free decision. But I’ve always eaten dinner in the past. Therefore (ampliatively), I will eat dinner tomorrow. Therefore, it is true that I will eat dinner tomorrow.

    December 8, 2008 — 7:26
  • Sure: Whether I eat dinner is a free decision. But I’ve always eaten dinner in the past. Therefore (ampliatively), I will eat dinner tomorrow. Therefore, it is true that I will eat dinner tomorrow.
    I hate to snipe, but it seems this just a case where you are ignoring (intentionally) the possibility of the failure of induction. You are working with a strong, elliptical, principle of uniformity (in the Humean sense) which there is really no cause for anyone to accept. Here is an argument in the same form as your little argument here that I think we can all see a flaw in:
    (1) I’ve always been alive in the past.
    (2) Therefore (ampliatively), I will be alive tomorrow.
    (3)Therefore it is true that I will be alive tomorrow.
    But of course this cannot be the case since I could very well die this evening (or in the next half hour)! the fact of the matter is that the *ampliative* truth of (2) is insufficient to justify the inference to (3), since it could be *ampliatively* true now that I will live tomorrow and it could be false tomorrow that I am alive.
    If (2) were true simpliciter, then (3) would follow, but then OT would just be false since there would per impossible be facts about the future.

    December 8, 2008 — 8:26
  • I could die in the next half hour, but I almost surely won’t, at least if I am careful in crossing the streets. Likewise, I am not claiming that I infallibly know that I will eat dinner tomorrow. I am only claiming that I know it. (Actually, maybe I should withdraw that claim, which I do not need for the argument, and only claim that I have strong evidence for it. I am kind of uncomfortable with saying I don’t know I won’t die in the next half hour. Christians should always be ready for death. Memento mori, and all that. I am inclined to a contextualism here, though.)
    Knowledge does not require an infallible connection between the evidence and the evidenced event. Sure, something could happen which would ensure that, despite all of my evidence, I won’t eat dinner tomorrow. But likewise, something could have happened which would ensure that, despite all of my evidence (such as my apparent memory of a quesadilla), I didn’t eat dinner yesterday. My knowledge of the past is fallible, just as my knowledge of the future is. You might claim that my beliefs about the future are more likely to turn out to have been false (or whatever the Open Futurist wants to say) than my beliefs about that past are to be false, but I am not sure about that, and even if it were so, this could just be a quantitative difference.
    In any case, it seems pretty clear that we can have very strong evidence about future contingents–evidence strong enough for us to stake our lives on it for the sake of very minor goods. Moreover, frequently we not only stake our own lives but the lives of our children. Surely I must have very strong justification in believing that when I leave home with my family, the people on the street won’t suddenly all start murderously attacking us without any provocation. Yet this is a future contingent claim.
    Maybe you could say that the arguments for an Open Future are a defeater for my evidence for the fact that we won’t be murderously attacked when we leave the house. But I don’t think so. The premises in the arguments for an Open Future include claims which, I think, no one can reasonably to know with a degree of certainty sufficient to stake the lives of their loved ones on them. Principles such as there is no backwards causation (if there is backwards causation, divine foreknowledge is surely not a problem in principle) or if an action is necessary given present facts, then it is not free, while somewhat plausible, are surely not nearly certain enough for us to stake the lives of our dependents on them.

    December 8, 2008 — 9:42
  • APT:
    “If (2) were true simpliciter, then (3) would follow, but then OT would just be false since there would per impossible be facts about the future.”
    Not quite, by the way. For some defenders of OT do believe there are facts about the future, but just don’t think God knows them.

    December 8, 2008 — 9:44
  • In any case, it seems pretty clear that we can have very strong evidence about future contingents–evidence strong enough for us to stake our lives on it for the sake of very minor goods. Moreover, frequently we not only stake our own lives but the lives of our children. Surely I must have very strong justification in believing that when I leave home with my family, the people on the street won’t suddenly all start murderously attacking us without any provocation. Yet this is a future contingent claim.
    I think the case is quite to the contrary: I think we never really have any evidence about future contingents, and thus we never have “strong justification” for believing in them. Take the simplest case, I have a two-sided coin, one side is called “heads” the other is called “tales”, I am intend to toss it in the air and let it strike the table and fall on one side or the other. On this bases I from the the following future contingent proposition:
    (FC) It is the case that the coin-toss will come out either heads or tales
    How can I have any better evidence for such a future contingent claim? I AM going to toss the coin, and it only has TWO sides, and if it lands it will land on only ONE side. So this proposition seems clearly true.
    But then suppose the laws of nature are not necessary, and that while the coin is in mid-air the weak nuclear force vanishes and the coin flies apart. Well then there is no outcome, and thus everything that seemed like evidence for (FC) at t1 turns out to not have been evidence at all.
    Here’s an alternate explanation of why you still go out with your family at night: it’s purely habitual. You ignore the possibility of danger (at least somewhat), because doing otherwise would interrupt your personal projects (there is also something to be said, on the this score, in favor of Sartre’s observation that we tend to treat others as background, or to negate their existence, in the course of pursuing our own interests). Moreover, murders happen relatively infrequently (nowadays), thus even though you have chosen to disregard the possibility that you or your loved one could be murdered, you have gotten along fine so far, but surely this is no evidence at all that it will always be so, your next trip out (or mine) could always be your last (or my last).

    December 8, 2008 — 11:11
  • Sure, in the case of knowledge of future contingents, the evidence is not sufficient to guarantee correctness. But the same is true in the case of past contingents. Just as you can give an unlikely story about how the coin might fly apart in the future, I can give an unlikely story how the random causes in the past have produced all the texts about Napoleon and there never was a Napoleon.
    The explanation you give for going out with family–and the problem is not just at night, but at any time and place where there other people–is just an explanation, and not a justification. However, the action is surely justified.

    December 8, 2008 — 11:17
  • I think we in agreement about past contingents, I (for one) never said they were any more believable than future contingents.
    And, to be clear, I was claiming that your going out is unjustified, but explicable in terms of your habit of acting unjustifiably. You seemed to assume strong justification, I deny that you have, or indeed that you even need, any. You go out because that is what you have always done and thus far you have not been murdered, had you been, you’d not be having this discussion.

    December 8, 2008 — 11:28