Monday, December 5, 2016

A Groundhog Day argument before Christmas

Last Sunday the advent sermon topic was how God’s foreknowledge might partly
explain how he could reliably speak about the future, which raises the question: Is divine
foreknowledge consistent with human freedom?


Here is what I will call a Groundhog Day Argument For the Compatibility of Divine
Foreknowledge and Creaturely Freedom:

1. By the end of the movie Groundhog Day, it is possible for Bill Murray’s character to foreknow some of the free actions of others, without his foreknowledge in any way compromising the freedom of those actions. 
2. What’s possible for Bill Murray’s character is possible for God.
Therefore, 
3. It is possible for God to foreknow some of the free actions of others,
without his foreknowledge in any way compromising the freedom of
those actions.
A longer treatment would argue that the two premises are true and that the most
promising objections to these premises are unsatisfactory. 

This post will focus entirely on supporting the first premise. 

Groundhog Day is a comedy about a TV weatherman, played by actor Bill Murray, who
has to visit a small town to cover their annual morning Groundhog Day festival. He finds
himself unable to leave the town that evening because of a snowstorm. He spends the
night in a local bed and breakfast. He wakes up the next morning only to discover that it
is Groundhog Day again. However, there is a catch: whereas Bill Murray’s character
remembers living through Groundhog Day yesterday, no one else in the town
remembers this. Although it seems to Bill Murray’s character that he lived through a
“first” Groundhog Day yesterday, and is living through a “second” Groundhog Day
today, everyone else in the town perceives the present day to be just the regular old
Groundhog Day, which comes around once—and only once—each year. 

(This is not just an epistemological breakdown of the memory of the other characters.
Rather, the sober but strange metaphysical fact of the matter is that everything in the
town, with the exception of Bill Murray’s character’s memory, is, at the beginning of the
“second” Groundhog Day, precisely as it was at the beginning of the “first” Groundhog
Day. For example, if Bill Murray’s character and another character both had a huge
greasy breakfast omelet on the “first” Groundhog Day, neither of them would have an
increased cholesterol number on the “second” Groundhog Day, but Bill Murray’s
character would remember having eating the breakfast on the “first” Groundhog Day.) 

Bill Murray’s character again tries to leave the town, the snowstorm again prevents him
from leaving, he again spends the night in the bed and breakfast, and, to his dismay, he
wakes up again the next morning only to discover that it is, once again, Groundhog Day.
The rest of the movie chronicles his attempts to escape this predicament, to
communicate it to other characters, and to cope with it in a number of different ways.
The connection between this movie and the first premise is seen in one of the strategies
Bill Murray’s character uses to cope with his predicament of being trapped in a cycle of
re-living the same day over and over again.

He begins to learn what each of the other characters in the movie would do in a certain
situation, and he begins to use what he learns in subsequent live-throughs of the same
day. 

For example, on the first “live-through” of Groundhog Day, he is greeted by an old
acquaintance on the street corner, and this old acquaintance tries to sell him some life
insurance. The exact same thing happens on the second live-through. On different live-
throughs, Bill Murray’s character responds differently to the greeting of the life
insurance salesman. The life insurance salesman, in turn, has a different comeback for
each of Bill Murray’s different responses to his greeting. If Bill Murray’s character does
A in response to the salesman’s greeting, the salesman comes back with action B. If Bill
Murray’s character does C in response to the salesman’s greeting, the salesman comes
back with action D. And so on. Bill Murray’s character, however, can remember how
the life insurance salesman comes back to different responses to his original greeting. 

In this way, Bill Murray’s character gains a very detailed knowledge of what different
characters in the movie will do in different situations. 

After a while, he uses this knowledge to take advantage of the other characters: for
example, he knows exactly when and where a certain security guard is going to look
away from a bagful of money, so he can easily slip in and steal the bag undetected. But
by the end of the movie, he has learned to use his newfound knowledge to benefit
others as well: for example, he knows exactly when and where a certain girl is going to
fall from a tree in her neighborhood, so he can catch the girl and save her from injury.
It seems, then, that by a certain point in the movie, Bill Murray’s character has
foreknowledge of some of the actions of others. For example, on what seems to him to
be the 8th live-through of Groundhog Day, Bill Murray’s character foreknows what
another character is going to do at 10:00 that morning. Bill Murray’s character also
foreknows what that other character is going to do at 10:00 in the morning on the next
live-through of Groundhog Day—what will seem to him to be the 9th live-through of
Groundhog Day. And so on. 

The peculiar features of the movie entail that the same types of actions done on the
“first” Groundhog Day will be repeated on “subsequent” Groundhog Days. 

What makes this argument for the compatibility of foreknowledge and freedom
interesting is that it simply collapses the difference that usually exists between
knowledge of the past and knowledge of the future. Bill Murray can foreknow the
future free actions of others because, in a sense, he’s already seen the other characters
perform these actions in the past.


Russell DiSilvestro
Department of Philosophy
Sacramento State

17 comments:

  1. I guess I'm more inclined to think that that's what makes the argument uninteresting. Murray's (the character's) foreknowledge of things that have already happened and that he's already observed isn't noteworthy precisely because it's not foreknowledge.

    But how does this work in God's case? Does the argument want to suggest that the mechanism by which God comes by his foreknowledge resembles the mechanism by which Murray's character comes by his "foreknowledge"?

    ReplyDelete
    Replies
    1. What? How can it be uninteresting to be suggesting that God has perfect memory of the future?

      Delete
    2. Kyle, apologies for the delay in this reply--it's not my practice to wait a week before responding. The short answer is that it really is still foreknowledge in Murray's case, even though it's partly based on a memorial knowledge of things that have already happened. There's a difference between the question of whether something is foreknowledge or not, and whether something that's foreknowledge is based partly on the past or not.

      I'm not claiming to show how this works in God's case. The argument is less ambitious than you suggest; it doesn't suggest how the mechanisms of divine foreknowledge and Murray foreknowledge resemble one another; just that the latter's compatibility with human freedom shows there's no in-principle problem in the former's compatibility with human freedom. However, I have had it suggested to me by others that this style of argument is pushing for so-called "middle knowledge" in both the Murray case and the divine case: roughly, that the knowledge of what free creatures would do in various counterfactual situations is the basis for foreknowledge of what these free creatures will do in a specific future situation. Maybe that's right.

      Delete
  2. I'm glad to see that I'm not the only one who uses the plot of Groundhog Day to try to answer philosophical questions!

    I have nothing to say against your point in this post--premise 1 seems clearly true to me, for just the reasons you give. 2, though, seems clearly false. It is possible for Bill Murray's character to do evil, cease to exist, become more powerful, etc. But if God has some divine properties necessarily, then it is not possible for God to do these things.

    More relevantly to free will concerns, it is possible for the character, but not God, to have a false belief. So, Bill Murray's character believing in advance that the insurance salesman will do B does not metaphysically necessitate that he will do B--if he failed to do B, the character would just have a false belief. Since God cannot have a false belief, we cannot say the same thing about divine foreknowledge. If God believes that the insurance salesman will do B, then he cannot possibly do otherwise, which some will think makes him unfree. What's the response to this?

    ReplyDelete
    Replies
    1. Brandon, let me echo my apologies above (to Kyle) for the delay in this reply. And let me echo part of what I say in a paper draft in response to a concern like yours here. Let's explicitly restrict the scope of the second premise so that it only applies to actions that correspond to what might be called perfections, and so it does not apply to actions that correspond to imperfections. So, for example, the second premise would apply to actions like remembering, but not apply to actions like forgetting. It would apply to actions like correctly inferring one proposition from another, but not apply to actions like incorrectly inferring one proposition from another. (That's kind of how I meant premise 2 to be taken.)

      The response to your last question, though, I think, is the standard one: God's foreknowledge of the insurance salesman doing B is harmlessly metaphysically supervenient on the salesman doing B, even freely doing B. But if the insurance salesman had failed to do B, then God would not have believed that he would…indeed, to borrow your language, God "cannot possibly do otherwise" than believe that the insurance salesman will fail to do B. The salesman's (freely) doing B or (freely) failing to do B is thus what I think is sometimes called a back-tracking counterfactual here.

      Delete
    2. Brandon, let me echo my apologies above (to Kyle) for the delay in this reply. And let me echo part of what I say in a paper draft in response to a concern like yours here. Let's explicitly restrict the scope of the second premise so that it only applies to actions that correspond to what might be called perfections, and so it does not apply to actions that correspond to imperfections. So, for example, the second premise would apply to actions like remembering, but not apply to actions like forgetting. It would apply to actions like correctly inferring one proposition from another, but not apply to actions like incorrectly inferring one proposition from another. (That's kind of how I meant premise 2 to be taken.)

      The response to your last question, though, I think, is the standard one: God's foreknowledge of the insurance salesman doing B is harmlessly metaphysically supervenient on the salesman doing B, even freely doing B. But if the insurance salesman had failed to do B, then God would not have believed that he would…indeed, to borrow your language, God "cannot possibly do otherwise" than believe that the insurance salesman will fail to do B. The salesman's (freely) doing B or (freely) failing to do B is thus what I think is sometimes called a back-tracking counterfactual here.

      Delete
  3. Russell, I agree with you that the alleged incompatibility of foreknowledge and free will is an error. However, I would not want to base the argument for this on what can happen to movie characters.

    Your entry begins by saying “By the end of the movie Groundhog Day, it is possible for Bill Murray’s character to” do X. Yes, in a movie a person might even be able to do both X and not-X. However, in real life a person cannot do this.

    Do you have a better argument for the compatibility of foreknowledge and free will?

    ReplyDelete
    Replies
    1. Brad, see my apologies (and cheating strategy) with Kyle and Brandon above—they are each for you too, now.

      I am willing to admit that merely describing a state of affairs in a work of fiction does not guarantee that the state of affairs, so described, is genuinely possible. I could write a short story in which the transitivity of identity does not hold: perhaps it helps the plot along that character A is identical to character B, character B is identical to character C, and yet character A is not identical to character C. Or I could produce a film about a universe in which 7 plus 5 does not equal 12, or write a book in which abstract geometrical shapes have mental states and converse with one another. But is there any reason for thinking that Bill Murray’s character is metaphysically offensive in these ways? I do not think there is. There is no logical contradiction involved in what happens to Bill Murray’s character. Nor is there any obvious metaphysical impossibility involved. So then, unless there is some reason for thinking otherwise, I believe we may be fairly safe to conclude, tentatively of course, that what’s possible for Bill Murray’s character is possible for God.

      Delete
  4. There’s a dialogue from one of my favorite comedies (Futurama) where one of the characters is talking to God. It sums up the foreknowledge problem nicely.

    Bender: So, do you know what I'm gonna do before I do it?
    God: Yes.
    Bender: What if I do something different?
    God: Then I don't know that.

    This illustrates Brandon’s concern above. There is not a tension between foreknowledge and freedom per se. There is a tension between infallible foreknowledge and freedom.

    I’m sympathetic to construe God’s foreknowledge as being similar to Murray’s foreknowledge. But I’m also sympathetic to compatibilism about free will (which says that deterministic laws of nature are compatible with human free will). And God’s infallible foreknowledge seems no more of a threat to freedom than deterministic laws of nature.

    However, libertarians about free will (who say that our choices are free and not fully determined by the laws of nature) might accept that Bill Murray justifiably believes (and perhaps even knows) that the salesman will do B in some particular circumstance, but they would still insist that there is some physically possible chance that the insurance salesman will not do B. And if the salesman does not do B (while Murray still believes he will do B), then Murray’s belief would be false in that situation. If Libertarianism is true, Murray’s foreknowledge is clearly fallible. So if libertarianism is true and God’s foreknowledge is akin to Murray’s, then God’s foreknowledge would be fallible, too.

    ReplyDelete
    Replies
    1. Tim, apologies for the delay, and the reliance on a paper draft here in my reply. I actually was thinking (though I did not say it in the original post) that the sort of freedom that this Groundhog Day Argument relies on is what you might simply describe as strong libertarian freedom. I think of this freedom in terms of what is sometimes called agent causation: an individual agent has this sort of freedom when she has an active power to bring about some state of affairs, where the exercise of this active power is not completely determined by past states of affairs inside or outside of the agent.

      Now, I think that when people sometimes doubt the freedom (let's just call it strong libertarian freedom!) of the other characters, what makes them doubt it is something like the following line of thought. In the movie, an individual always performs the very same action whenever they are confronted with a certain situation, no matter what live-through Bill Murray’s character is living. But if an individual were really (strong libertarian!) free, then there has to be some live-through where she freely (strong libertarian freely!) performs some other action when confronted with the very same situation. The different “live-throughs” of Bill Murray’s character thus play roughly the same role as different “possible worlds” play in a possible-worlds metaphysic. Some people think that a person in the actual world freely (strong libertarian freely!) performs an action in a given situation only if there is some possible world where that person in that situation does not perform that action. Likewise, this objection thinks that a character c freely performs an action a in a situation s in one live-through only if there is some other live-through where c in s does not perform a. Since there is no such live-through where this is the case, it follows (according to this objection) that there was no (strong libertarian!) freedom to begin with.

      Now, I disagree with this line of thought, but the reasons are complicated and I won't get into them here...

      Delete
    2. Russell,

      Thanks for your thoughtful reply as well.

      With regard to point (1), I was actually thinking that God - assuming existence - wouldn't have a counterpart. But that would not affect argument. If God knows at t in W1 that E will happen and knows at t in W2 and E will not happen, then God's state of mind at t in W1 is different from God's state of mind at t in W2, which contradicts the hypothesis that W1 and W2 have the same past (i.e., the same temporal initial segment) at least up to and including t.
      What I'm arguing is that if two worlds have the same past, they also have the same future, assuming that God exists and has foreknowledge (and regardless of whether any interpretation of QM is true or approximately true; QM was meant to be an example only). I grant that that might not be a problem for the type of libertarian freedom you defend, given your reply to Tim.

      Regarding point (2), I'm not a theist, and I tend to disagree that considerations about God might (in practice) constitute non-negligible evidence against interpretations of QM, but discussing that would be beyond the scope of the thread, so I'll leave it at that.

      Delete
  5. Russell,

    As a compatibilist, I don't think foreknowledge is incompatible with free will.
    However, I don't think arguments based on human knowledge of the future work in this context, because claims of incompatibility between God's foreknowlegde and human free will are usually based on beliefs that determinism is incompatible with free will (which I think is false), and a belief that God's foreknowledge entails determinism (which I think is true, in what appears to be a common sense of "determinism"; more below).

    On the other hand, the same is not true of human knowledge of the future, as I think the following example illustrates:

    Let's say that determinism is false, and our universe works more or less as an indeterministic interpretation of QM holds (with some adjustment for gravity and potentially other forces, etc., but roughly like that).
    It's actually true that Trump is not going to quantum tunnel through a wall before 2020.
    There is a possible world W2 with the same initial segment up to today as our world in which he will do so.
    So, I know today that Trump is not going to quantum tunnel through a wall before 2020, but in W2, I (or my counterpart) falsely believes today (rather than knowing) that Trump isn't going to quantum tunnel, etc.

    In my assessment, that example - or similar ones - shows that when we stipulate that two worlds have the same initial segment (or more informally, that the past is the same up to some point, etc.), we don't mean to include (by that stipulation alone) in "the past" the epistemic status of any agent's beliefs (i.e., whether they constitute knowledge or not).

    So, in particular, human knowledge of the future does not entail determinism.

    But what about God's knowledge, stipulating he's essentially infallible and omniscient?
    While the stipulation that two worlds have the same past does not entail that the epistemic status of the beliefs of the agents in that world are the same, it does entail that all concrete states of the worlds (including the states of mind of the agents in the two worlds) are the same.

    So, let's say that God exists, and in W1, at t+s, event E happens. Then (by God's essential omniscience and infallibility), at t, in W1 God knows that at t+s, event E will happen.
    If W2 is a world with the same initial segment as W2 up to and including t, then God's state of mind at t in W2 is the same as his state of mind at t in W1. In particular, in W2 God believes that E will happen at t+s (if God doesn't have beliefs, then he has an intuitive apprehension that E will happen, or however one properly construes God's state of mind other than belief).
    But God - unlike Murray's character, or me - isn't possibly mistaken, so E will happen in W2 at t+s, so W2=W1 at every time.

    For that reason, I think God's foreknowledge entails determinism whereas our knowledge of the future doesn't, and so if it's true that determinism precludes free will, then God's foreknowledge of the future precludes free will, but our knowledge of the future doesn't. As I said, I don't think determinism precludes free will, but the argument based on Murray's character doesn't seem to consider that distinction.

    ReplyDelete
    Replies
    1. Angra, thanks for this thoughtful reply. I definitely need to educate myself more on QM, but my initial naive reaction is to wonder whether we might think either

      (1) that God fully exists in W1 and W2, so he doesn't have a 'counterpart' like I (supposedly) do as a full-blooded denizen of W1 but not W2, and so therefore God knows in W1 that E will happen and knows in W2 and E will not happen; or

      (2) that God's foreknowledge indeed might constitute an argument against an indeterministic interpretation of QM, but not in a way that causes trouble for libertarian freedom (see my response to Tim on why this might be so)…

      Still, you've given me a lot more to puzzle over than I bargained on here. Thanks.

      Delete
  6. I have never had a problem with divine foreknowledge and human freedom, but the more interesting problem, I think, is the apparent incompatibility of divine providence and human freedom. This better parallels the problem of the apparent incompatibility of determinism and human freedom. The problem arises because, for many theists, God not only knows the counterfactuals of human freedom, but also determines some for salvation and others for eternal damnation (or, on a lesser scale, determines human events or actions to bring about a particular outcome).

    It gets tricky when a person starts using knowledge to control outcomes, particularly when we’re talking about an omniscient being who knows how every human being would act in every possible scenario and is able to consider this information to effectuate his own plans.

    I’m not a Bill Murray fan and only have a vague recollection of the movie, but, using that example, the real problem arises when Bill Murray’s character begins to use his knowledge to take advantage of the other characters. Does this manipulation undermine human freedom?

    Substitute God for Bill Murray’s character and there’s even more to worry about. God could control every feature of every circumstance (everything that exists in the world, the laws of nature, human rationality) and, knowing the counterfactuals of human freedom, can at least influence every thought and control every action. A person may think he is free to will what he wants to will consistently with his personality, interests, goals, values, and beliefs, but all of these factors may have been previously shaped by God’s control of every feature of every circumstance. I think the presence of such powerful and pervasive manipulation, objectively, would undermine human freedom altogether (just as the determinist may object to Frankfurt by pointing out how the factors--biology, upbringing, experiences, etc.--that are determined by antecedent conditions not only shape our beliefs and desires but also shape our second-order volitions, the divine determinist may point out how God could manipulate the conditions that shape our personality, interests, goals, values, and beliefs).

    But, of course, God would never use his knowledge to take advantage of anyone, so all my worrying here is for nothing.

    ReplyDelete
    Replies
    1. Chong, sorry for the weeklong delay here. As someone whose views on this topic were shaped by reading Thomas Flint's Divine Providence: The Molinist Account (Cornell UP), I confess to a strong desire to make human freedom compatible both with divine foreknowledge and divine providence. And indeed, apart my my desires on the topic, there seems to me no reason not to go all the way if you've learned how not to worry but to love the former compatibility. If you know I am going to "freely" do X in situation Y, what difference does it make whether you are the one to put me into situation Y or not? And what difference does it make whether your designs in the situation are entirely maleficient (say, to see me come to a terrible end), entirely beneficent (say, to see me come to a wonderful end), or somewhat in between (say, to win a bet with another philosopher who said you surely couldn't predict what I would freely do in situation Y)? In all cases we have (partial) providence in action.

      Delete
  7. Refreshing post and one of my favorite topics. I've never understood the alleged tension between foreknowledge and free will. And as for compatibilism, it doesn't seem compatible. That's not to say free will isn't limited by God, but that determinism cannot coexist with free will.

    I would never go so far as the Open Theists and say that God doesn't know, determine, or guide the future, however. He is sovereign, and gives us freedom. I guess you could say I'm "pro-choice" in that regard.

    ReplyDelete
    Replies
    1. Ryan, thanks for responding here. I don't think I would count myself as an Open Theist either, but I think their position has a basis both in philosophical concerns (like the alleged incompatibility of foreknowledge and freedom) and in biblical concerns (like the best way of understanding passages where God almost seems to change his mind when humans grapple with him in prayer). The Groundhog Day argument only helps with the former.

      Delete