Q&A: Pascal’s Wager and Adi Tzemach’s Argument
Pascal’s Wager and Adi Tzemach’s Argument
Question
Hello Rabbi,
In recent days I read the chapter on Pascal’s Wager in the Rabbi’s book. The Rabbi relates there to Pascal’s argument as an argument dealing with expected payoff—if the expected payoff of belief in God is positive, then the rational person will choose to believe in God (and keep the commandments) for the sake of the benefit in that belief. Although the Rabbi goes on to discuss the rational aspect of choosing according to expected value, I would be glad to hear the Rabbi’s opinion regarding the argument Adi Tzemach raises about Pascal’s Wager.
Tzemach argues that Pascal conditions belief on the claim about the benefit that grows out of it, and that this dependence does not allow one to believe the thing at all:
He assumes, like Pascal, that the chances of God’s existence are no greater than the chances that He does not exist.
However, the benefit of believing in Him is greater than the benefit of not believing in Him.
Therefore it pays to believe in God more than not to believe in Him.
Now Tzemach looks at this claim: suppose Pascal is convincing enough for us to believe that the probability that it pays to believe in God (more than not to believe in Him) is indeed higher than the probability that it does not pay to believe in God. Now we must examine the benefit of this claim:
Does it pay to believe that it pays to believe in God rather than not to believe in Him, more than not to believe that it pays to believe in God rather than not to believe in Him?
I apologize for the cumbersome wording, but it illustrates Tzemach’s argument well: the dependence between belief and benefit creates an infinite recursion (in his words) that makes it impossible to believe anything, since we will always have to examine the benefit in the claim we have just proved, and so on.
Tzemach concludes by saying that there should be no dependence between the epistemic value of a claim and its benefit.
What does the Rabbi think about this? Is the distinction Tzemach creates within Pascal’s argument unavoidable? Can it be explained differently?
Answer
I don’t see any recursion here. It seems to me just empty words.
Pascal’s argument deals with belief in God, and that is a claim that is not accessible to scientific and logical tools (on his assumption). Therefore there he uses the statistical argument in order to make a decision. But the statistical claim that leads to belief is itself not a claim of that sort, and therefore there is no point in applying Pascal’s algorithm to it itself.
If, from your perspective, expected payoff really is the criterion for choosing a way of life, and if the expected payoff of belief really is infinitely greater than the expected payoff of non-belief, then one should believe. QED. How are you applying a wagering argument here?
Discussion on Answer
On second thought I don’t understand the answer, and the recursion seems very problematic to me.
I’ll try to formulate the recursion:
1. The statistical claim “it pays to believe in God” is presented to us.
2. Does this claim seem certainly true to us? No. There is a chance it isn’t true. Maybe it really doesn’t pay?
3. It isn’t certainly true, so how do we decide whether to accept it (at least practically, if it’s impossible to really “accept” opinions)?
4. According to the criterion of what is more worthwhile for us.
5. And it seems that it is indeed worthwhile for us to believe that it pays to believe in God (it simply gets swallowed up).
6. So we accept this claim.
7. Does this claim (that it pays to believe that it pays to believe in God) seem certainly true to us? No.
8. But it is probably worthwhile. And so on.
9. That is, we require an infinite number of justifications of “it’s worthwhile” in a regressive attempt to ground the claim that it pays to believe in God.
Correction:
Section 2 should say “Maybe it really doesn’t pay,” and not “Maybe it really does not not pay.”
What is this ridiculous attempted refutation? If you discovered that it pays to believe in God, then that is certainly true. It’s not a statistical matter. It seems the writer didn’t understand simple matters such as expected value. He got confused with the concept of variance.
In any case.
In the past I already pointed out simply what the refutation of Pascal’s argument is. And the Rabbi didn’t really have anything serious to answer.
The simple refutation reveals the falsehood in the argument. A hidden assumption on which the whole argument stands:
The claimant assumes that belief in God brings infinite pleasure in the future.
The point is that, a priori, one could argue with exactly the same degree of certainty that belief in God brings infinite torment in the future.
And if someone asks what the logic is behind this strange argument, one could explain to him that it’s obvious to everyone that God wants people not to know about Him. If He wanted people to know about Him, He would reveal Himself. But He wants to hide. Therefore, whoever discovers the God who wants to hide will be punished with infinite torment.
In any case, there is no a priori reason to assume that God is something that wants people to believe in Him. One could just as well say that He is interested in people not believing in Him.
And Pascal’s whole argument collapses.
It seems to me that the answer can be phrased a bit differently. Even if it doesn’t pay to believe that it pays to believe in God, that still seems to us to be a true claim. Expected-value considerations aren’t relevant when I know what the correct side is; that is, truth also plays a role and not only benefit.