Julian Sanchez header image 2

photos by Lara Shipley

What Sort of God Would Allow This?

December 14th, 2007 · 17 Comments

Why, New York Times, why, if you’re going to give op-ed space to an atheist to discuss his views, would you decide to print two grafs that come off as whining followed by a sub-sophomoric attempt at refuting Pascal’s Wager? The Wager is an appallingly bad argument that should take about two sentences to dispense with, and the author still blows it: Once you accept the posit of a lottery with infinite expected value, the size of the finite stake is irrelevant. (It does matter for monetary wagers because of diminishing marginal utility, but the payoff here is imagined to be directly in utility.) The only thing that needs to be said about the Wager is that once you’ve discarded consideration of the probabilities—that’s why the payoff has to be infinite, and the only way the Wager works—then it works for everything. Maybe I’ll be rewarded with infinite bliss if I walk outside in my underwear right now. Vanishingly small probability, infinite payoff, better brace myself for that chill!

Tags: Religion



17 responses so far ↓

  • 1 R. Totale // Dec 14, 2007 at 1:12 pm

    Someone in the comments brings up Stalin, Mao, etc. What a devastating rejoinder.

  • 2 dan // Dec 14, 2007 at 2:10 pm

    I dunno. Mathematically you’re right, giving anything infinite value always screws equations up, but I think the columnist is correct that if you are tempted by Pascal’s wager, you also have to consider whether belief is really costless. Go outside in the cold naked and you have a certainty of being uncomfortable and embarrassed (well, I guess I shouldn’t speak for you on the latter), a non-negligible chance of being arrested, and a negligible chance of something going terribly wrong and causing you to die (you get arrested and hauled off to jail and your cellmate murders you, e.g.), which I assume would be infinitely bad, which as a practical matter I think offsets the possibility of an infinite good. As whosis at the times said, Pascal’s wager only works if he’s right that belief has no cost (or to be more precise, can never have an infinite cost, i.e. a cost commensurate with the benefit of eternal salvation). I don’t know that I agree with whosis that belief itself has costs, but I certainly disagree with you tbat Pascal’s wager can be used to justify anything.

    Of course, that assumes that death (in a universe without God) is as bad as salvation is good. I have no idea how to make that argument, although I’m pretty sure angels dancing on pins come into the equation pretty quickly.

  • 3 Julian Sanchez // Dec 14, 2007 at 2:27 pm

    OK, fine, make it knocking three times on my desk instead of going out in the cold. Or whatever else.

  • 4 dan // Dec 14, 2007 at 3:09 pm

    hmm. I don’t think that’s a trivial point, and certainly don’t mean to suggest I was merely picking on your example. To me, it’s a reminder that Pascal’s wager really does rely on the absolute costlessness of the alternative, and that’s very debateable. It takes a lot of activities off the table, but more to the point it offers a pretty decent weakness for an already weak argument.

    upon reflection, though, maybe your new example just further shows how silly it is to make any arguments based on discussions of infinite value. You knock on the table, get a splinter, it gets infected…

  • 5 dan // Dec 14, 2007 at 3:16 pm

    I have no idea why I said “decent” instead of “significant.”

  • 6 Daniel // Dec 14, 2007 at 3:36 pm

    I’m no philosophy major, but say someone made you the following offer: if you agree to be locked in jail for the rest of your life, you’ll have done the pious thing and go to heaven for your infinite reward. (Some people would say going all your life accepting what you’re told approximates the jail cell, etc…)

    Julian’s point about the fundamental non-sequitur of the argument is perhaps sufficient; but I agree with dan; I think the fallacy of costlessness damns the wager argument whereas pointing out the insensibility just sidelines it. Both are good though…

  • 7 Pablo Stafforini // Dec 14, 2007 at 4:10 pm

    Once you accept the posit of a lottery with infinite expected value, the size of the finite stake is irrelevant.

    This is not strictly true. If you are a pure time discounter and if, moreover, you discount time exponentially rather than hyperbolically, then the present value of a finite period in the near future may outweigh the value of an infinite period in the further future. Whether this is so will depend, among other things, on the size of the finite stake.

  • 8 Rue Des Quatre Vents // Dec 15, 2007 at 5:57 pm

    Not so fast, oh philosophe. If your argument dissolved the riddle of induction (and, obviously, it doesn’t), then you’d have Pascal’s number. You also overlook an important distinction between Pascal’s choice and your own involving knocking on the desk. For both these reasons, your argument fails. You assume that dying is comparable to knocking on a desk. Why is this misguided?

    (1) It is assumed that death could be the only access to knowledge and experience of the divine. Think in terms of cause and effect. If dying is the cause, and its effect is divine knowledge, then it is a process for which we have no evidence, either way. The dead’s silence should affect our priors. (This clearly distinguishes it from your examples involving walking outside, knocking on tables and so on. We obviously have priors about those processes. Since the experience of death is unique to the person dying, I would like to see an example that exhibits the same properties.)

    (2) Since there’s no bond of necessity uniting cause and effect, our knowledge of the effects of dying (for the dead) are even less certain. It may be that there is no infinite payoff after dying a believer…now. This is not to say that won’t change.

  • 9 Julian Sanchez // Dec 15, 2007 at 9:07 pm

    Neither of these is really relevant

  • 10 Rue Des Quatre Vents // Dec 15, 2007 at 10:20 pm

    Dude, great counter argument.

  • 11 Fred S. // Dec 16, 2007 at 1:50 am

    There is a question-begging aspect to the Wager. As J.S. indicates, there are potentially an infinite number of costless activities which we could convince ourselves might lead to eternal life (of which knocking on a table would be an arbitrary example). There is no intrinsic reason to privilege the Christian doctrine of salvation by faith. To abide by the wager, we must implicitly accept that Christianity has a superior claim to theological truth than any other arbitrary, self-imposed demand (and thus we’ve begged the question).

    More damnigly, though, it makes the psychologically-dubious claim that faith can be turned on like a tap.

  • 12 dan // Dec 16, 2007 at 1:14 pm

    Pascal is sort of getting a bad rap here, I think. He wrote a very short, quite interesting argument for belief in god in the course of a much longer, much less interesting argument for adherence to vaguely christian theology. it’s not his fault that his wager doesn’t get you to christianity; he never claimed it would, it just asks whether, in light of the absence of evidence, one ought to believe in god (without too much attention, yet, to what that god would be like). Of course, I guess it’s his fault that the rest of the argument, where he tries to go further, is ponderous and forgettable, but hey, it’s not like anything I’ve ever written is going to be remembered for more than a week or two, so i don’t want to cast stones.

    I’m with fred s.’s last claim, though–I’ve always thought that the silliest part of the wager is thinking that there’s any relationship between concluding that it would be best to believe in god and believing in god. not a philosophically interesting objection, but still…

  • 13 Biff // Dec 16, 2007 at 3:10 pm

    Dude, great counter argument.

    If you’d made any kind of coherent argument in the first place, Julian might have had something to say about it.

    You assume that dying is comparable to knocking on a desk.

    No, the analogy is between spending a lifetime worshipping a god, and knocking on a desk. Now try again.

  • 14 David // Dec 17, 2007 at 8:55 am

    The Stanford Encyclopedia has a nice piece on the Wager:


  • 15 Lee // Dec 17, 2007 at 1:18 pm

    Also check out Ian Hacking’s take on Pascal’s Wager in his “Introduction to Probability and Inductive Logic” and the more historically-oriented “Emergence of Probability”. Pascal’s argument involves some psychological subtleties about how belief works, that make the argument seem less silly.

    Also, Hacking asserts that Pascal was the very first person to make a decision theoretical argument of this type; no one had every made an argument quite like it before. So don’t beat up on Pascal too much. The guy sort of invented the rigorous notion of probability we have today, ya know.

  • 16 Lee // Dec 17, 2007 at 1:29 pm

    Oh, and don’t count on smart philosophers to easily dispense with Pascal’s wager. I was really startled to hear Dan Dennett screw it up during a debate with Dinesh D’Souza. (video series here). He said something to the effect of “Well, if that’s right, why not bet on all the religions at once?” As D’Souza rightly observed, religions have some mutually exlusive claims and rules, so that’s not a very good refutation.

    I do think Julian’s one sentence reductio does the job, however. The infinite utility prospect leads to absurdity.

  • 17 washerdreyer // Dec 18, 2007 at 1:09 pm

    The best response, which I stole from some blogger (I’m thinking dsquared, but not positive), is that since a slight chance of an infinite payoff is of infinite value, and there’s some chance that my intercession with God n your behalf will cause you to receive this infinite payoff, you should give me all of your money right now in exchange for my interceding with God.