The usual analyses of Pascal’s Wager, like many lab experiments, privileges the hypothesis and doesn’t look for alternative hypotheses.
Why would anyone assume that the Mugger will do as he says? What do we know about the character of all powerful beings? Why should they be truthful to us? If he knows he could save that many people, but refrains from doing so because you won’t give him five dollars, he is by human standards a psycho. If he’s a psycho, maybe he’ll kill all those people if I give him 5 dollars. That actually seems more likely behavior from such a dick.
The situation you are in isn’t the experimental hypothetical of knowing what the mugger will do depending on what your actions are. It’s a situation where you observer X,Y, and Z, and are free to make inferences from them. If he has the power, I infer the mugger is a sadistic dick who likes toying with creatures. I expect him to renege on the bet, and likely invert it. “Ha Ha! Yes, I saved those beings, knowing that each would go on to torture a zillion zillion others.”
This is a mistake theists make all the time. They think hypothesizing an all powerful being allows them to account for all mysteries, and assume that once the power is there, the privileged hypothesis will be fulfilled. But you get no increased probability of any event from hypothesizing power unless you also establish a prior on behavior. From the little I’ve seen of the mugger, if he has the power to do what he claims, he is malevolent. If he doesn’t have the power, he is impotent to deliver and deluded or dishonest besides. Either way, I have no expectation of gain by appealing to such a person.
The usual analyses of Pascal’s Wager, like many lab experiments, privileges the hypothesis and doesn’t look for alternative hypotheses.
Yes, privileging a hypothesis isn’t discussed in great detail, but the alternatives you mention in your post don’t resolve the dilemma. Even if you think that that the probabilities of a “good” and “bad” alternatives balance each other out to the quadrillionth decimal point, the utilities you get in your calculation are astronomical. If you think there’s a 0.0000quadrillion zeros1 greater chance that the beggar will do good than harm, the expected utility of your $5 donation is inconceivably greater than than a trillion years of happiness. If you think there’s at least a 0.0000quadrillion zeros1 chance that $5 will cause the mugger to act malevolently, your $5 donation is inconceivably worse than a trillion years of torture. Both of theses expectations seem off.
You can’t just say “the probabilities balance out”. You have to explain why the probabilities balance out to a bignum number of decimal points.
You have to explain why the probabilities balance out to a bignum number of decimal points.
Actually, I don’t. I say the probabilities are within my margin of error, which is a lot larger than “0.0000quadrillion zeros1”. I can’t discern differences of “0.0000quadrillion zeros1″.
OK, but now decreasing your margin of error until you can make a determination is the most important ethical mission in history. Governments should spend billions of dollars to assemble to brightest teams to calculate which of your two options is better—more lives hang in the balance (on expectation) than would ever live if we colonized the universe with people the size of atoms.
Suppose a trustworthy Omega tells you “This is a once in a lifetime opportunity. I’m going to cure all residence of country from all diseases in benevolent way (no ironic or evil catches). I’ll leave the country up to you. Give me $5 and the country will be Zimbabwe, or give me nothing and the country will be Tanzania. I’ll give you a couple of minutes to come up with a decision.” You would not think to yourself “Well, I’m not sure which is bigger. My estimates don’t differ by more than my margin of error, so I might as well save the $5 and go with Tanzania”. At least I hope that’s not how you’d make the decision.
The usual analyses of Pascal’s Wager, like many lab experiments, privileges the hypothesis and doesn’t look for alternative hypotheses.
Why would anyone assume that the Mugger will do as he says? What do we know about the character of all powerful beings? Why should they be truthful to us? If he knows he could save that many people, but refrains from doing so because you won’t give him five dollars, he is by human standards a psycho. If he’s a psycho, maybe he’ll kill all those people if I give him 5 dollars. That actually seems more likely behavior from such a dick.
The situation you are in isn’t the experimental hypothetical of knowing what the mugger will do depending on what your actions are. It’s a situation where you observer X,Y, and Z, and are free to make inferences from them. If he has the power, I infer the mugger is a sadistic dick who likes toying with creatures. I expect him to renege on the bet, and likely invert it. “Ha Ha! Yes, I saved those beings, knowing that each would go on to torture a zillion zillion others.”
This is a mistake theists make all the time. They think hypothesizing an all powerful being allows them to account for all mysteries, and assume that once the power is there, the privileged hypothesis will be fulfilled. But you get no increased probability of any event from hypothesizing power unless you also establish a prior on behavior. From the little I’ve seen of the mugger, if he has the power to do what he claims, he is malevolent. If he doesn’t have the power, he is impotent to deliver and deluded or dishonest besides. Either way, I have no expectation of gain by appealing to such a person.
Yes, privileging a hypothesis isn’t discussed in great detail, but the alternatives you mention in your post don’t resolve the dilemma. Even if you think that that the probabilities of a “good” and “bad” alternatives balance each other out to the quadrillionth decimal point, the utilities you get in your calculation are astronomical. If you think there’s a 0.0000quadrillion zeros1 greater chance that the beggar will do good than harm, the expected utility of your $5 donation is inconceivably greater than than a trillion years of happiness. If you think there’s at least a 0.0000quadrillion zeros1 chance that $5 will cause the mugger to act malevolently, your $5 donation is inconceivably worse than a trillion years of torture. Both of theses expectations seem off.
You can’t just say “the probabilities balance out”. You have to explain why the probabilities balance out to a bignum number of decimal points.
Actually, I don’t. I say the probabilities are within my margin of error, which is a lot larger than “0.0000quadrillion zeros1”. I can’t discern differences of “0.0000quadrillion zeros1″.
OK, but now decreasing your margin of error until you can make a determination is the most important ethical mission in history. Governments should spend billions of dollars to assemble to brightest teams to calculate which of your two options is better—more lives hang in the balance (on expectation) than would ever live if we colonized the universe with people the size of atoms.
Suppose a trustworthy Omega tells you “This is a once in a lifetime opportunity. I’m going to cure all residence of country from all diseases in benevolent way (no ironic or evil catches). I’ll leave the country up to you. Give me $5 and the country will be Zimbabwe, or give me nothing and the country will be Tanzania. I’ll give you a couple of minutes to come up with a decision.” You would not think to yourself “Well, I’m not sure which is bigger. My estimates don’t differ by more than my margin of error, so I might as well save the $5 and go with Tanzania”. At least I hope that’s not how you’d make the decision.