Thanks for the term. I’ve Googled it and had a very brief look over evidential versus causal decision theory (will look more later). And I accept that the evidential theory is flawed, and the causal theory is better, and that EDT gives the wrong result on the smoking problem, and so on.
But keeping the original premise that it’s known that out of everyone who’s ever lived in all of history, every single virtuous Calvinist has ended up in Heaven and every single sinful Calvinist end has ended up damned—I still choose to be a virtuous Calvinist. And if the decision theorists don’t like that, they can go to hell.
If God models your actions and pre-decides to save or damn you accordingly, sure. But if, as you say, behavior is caused by elect status, it seems to me that’s exactly the same as the Smoking Lesion (where CDT is in the right).
I see a difference between the two cases as follows:
The Smoking Lesion, as I interpret it when I say it seems to be self-evidently a correct counterexample to EDT, only affects your craving to smoke. Your probability of getting lung cancer is therefore affected only by your craving to smoke. Once you note whether or not you have a strong craving for cigarettes, your choice whether or not to smoke provides no further evidence for your possession of the lesion than the craving alone did.
As I interpret Calvinism, your status as elect or reprobate doesn’t just cause a craving to live virtuously. It actually causes the entire decision. Unlike in the smoking lesion, my internal argument about which decision theory I use to decide the Calvinism problem is itself part of the evidence for whether I am among the elect or not. This gives trying to work out a decision theory about Calvinism a self-referential element that the smoking lesion completely lacks.
As I said to Eliezer, I would be much more confident that living a virtuous life is the correct decision if I noted that many Calvinists before me had engaged in this exact argument about decision theories, and every single one who eventually came to hold my opinion went to Heaven and every single one who rejected it went to Hell.
If we used the same constraint as in the Augustine’s Paradox thread—that God is interested in your decision-making algorithm and rejects anyone who lives a virtuous life for reasons other than their native disposition, then I agree with you and I would live a life of sin.
I don’t find Cleonid’s comparison with run-of-the-mill determinism inappropriate. The position of particles during the infancy of the universe completely determined whether or not I will get out of bed tomorrow morning. In that sense, I am a get-out-of-bed Calvinist: my fate was already decided at the moment of Creation, and I am either one of the elect whose genes/environment/etc predispose him to rising, or one of the reprobate whose genes/environment/etc doom him to stay in bed all day.
And yet I wake up tomorrow morning, and I find myself presented with what seems like a choice to get out of bed or not.
I am happy to say I “choose” to get out of bed, and I am happy to say I would “choose” to be a virtuous Calvinist. Even though I’m imbuing the word “choose” with less mystical force than a believer in free will might.
(here I am assuming a relationship between spiritual causes and the soul that’s similar to the relationship between physical causes and the brain you explained in Thou Art Physics. As far as I know, Calvinists believe that God’s choice is mediated through a change in the effect of original sin on the soul of the person He has chosen, which seems to fit the bill.)
How can I make this clearer...okay. Let’s say there have been a trillion Calvinists throughout history. They’ve all been rationalists and they’ve all engaged in this same argument. Some of them have been pro-sin for the same reasons you are, others have been pro-virtue for the same reasons I am. Some on each side have changed their minds after having listened to the arguments. And of all of these trillion Calvinists, every single one who after all the arguments decides to live a life of virtue—has gone to Heaven. And every single one who, after all the arguments, decides to live a life of sin—has gone to Hell.
To say that you have no reason to change your mind here seems to be suggesting that there’s a pretty good chance you will be the exception to a rule that has held 100% of the time in previous cases: the sinful Calvinist who goes to Heaven, or the virtuous Calvinist who goes to Hell. If this never worked for a trillion people in your exact position, why do you think it will work for you now?
Can one say that the God of the Calvinists as well as the alien of Newcomb’s Problem have the ability to perfectly predict (at least specific things about) the future?
And isn’t having that ability exactly the same as having a crystal ball that actually can look into the future? Isn’t then being able to predict the future with 100 percent certainty the same as having the ability to actually look into the future? Then, I think, it might be possible to say that the God or alien cannot be outwitted. Anything you do – no matter what – has been correctly predicted or is actually seen in God’s or the alien’s crystal ball. If you two box the alien has predicted just that, if you are not virtuos God has predicted just that. Your brain cannot change that. Your brain cannot escape perfect prediction. All escape attempts will trigger God to throw you in hell and the alien to leave you with just $1000.
I think this is to an extent even true if God or the alien are wrong some of the time – if they are only able to predict the future accurately 99 percent of the time. One would only be able to outwit the alien if one where to know under which circumstances the predictions of the alien break down. And as long as we are talking about random failures to predict correctly, a God or alien that is 99 percent accurate has still an almost perfect ability to look into the future. Has this implications for our decisions if we know that someone else can predict our own decisions with some accuracy? I think we should strive to know others models of our decision making and know when these models break down. That could be useful.
And isn’t having that ability exactly the same as having a crystal ball that actually can look into the future? Isn’t then being able to predict the future with 100 percent certainty the same as having the ability to actually look into the future? Then, I think, it might be possible to say that the God or alien cannot be outwitted.
Not so. While you can never surprise a God that can perfectly predict the future, you may certainly outwit him. If the God limits his ‘rules’ to that which could be written on say, a few stone tablets then it is certainly possible to manouver around them. Alternatively, the God in question may be omniscient yet stupid. I’ve seen plenty of humans who will not change their strategy when confonted with absolute and trasparent proof that their strategy is flawed and will not work. I see no reason why a God could not do the same thing.
Yes, certainly, but that is besides the point. This problem here is about actually violating the rules and the question as to whether you can get away with it.
And it still depends on the nature of the the omniscient being in question. Having enough information does not make you make wise choices.
If you happen to know how the God thinks you can answer for sure whether or not you can outwit him. As you said yourself, “I think we should strive to know others models of our decision making and know when these models break down. That could be useful.” This applies to omniscient yet fallible Gods as well. Heck, you can know the entire state of the universe at all times and still not understand the Calvinist problem. In that case, dumbfounding the idiot god is trivial.
But the Calvinist who decides to live a sinful life visibly violates the rules. Even a dumb God who only sets up a simple list of rules and who parses your behaviour only according to that list would notice that. I have certainly no doubts about the possibility of finding holes in the rules of God so that you would certify as virtuous even if other humans would most likely not see you that way.
But as long as prediction is concerned there is no outwitting. If you find holes in the rules that is not the same as finding out where prediction models break down (and if God can predict perfectly there is no point where his model will break down). I think that’s an important distinction to make. You can certainly outwit the rules (if they have holes), you cannot outwit prediction (if it is perfect).
Hrm… you seem to have better google-fu than I. So, what is “evidential decision theory” then? or at least, can you point me to whichever link you found that actually had basic description on it? Thanks.
But then, couldn’t one just say that one has to be more precise and ask: to what extent does the common cause for both induce the correlation in the context of mental algorithms that actually execute that sort of reasoning?
ie, simply include that as part of it. Wouldn’t doing that sort of thing automatically fix these sorts of errors?
I may be misunderstanding, but if it’s what I think it is, doesn’t one have to take a position like that simply by virtue of “I am part of reality. My decisions are not some transcendent divine free will from beyond all, but are part of the web of causality. They are not just causes, but also effects. The particular fact that I’m reasoning in this way is ALSO an effect, not just a cause. I am not an intervention, but a part of reality.”
Well, seems to me at least one of the following is true: I am somewhat misunderstanding what EDT is, or my reasoning above is flawed, or some flavor of EDT is actually Right Way.
See also “evidential decision theory”, the wrong reason to one-box on Newcomb’s Problem.
Thanks for the term. I’ve Googled it and had a very brief look over evidential versus causal decision theory (will look more later). And I accept that the evidential theory is flawed, and the causal theory is better, and that EDT gives the wrong result on the smoking problem, and so on.
But keeping the original premise that it’s known that out of everyone who’s ever lived in all of history, every single virtuous Calvinist has ended up in Heaven and every single sinful Calvinist end has ended up damned—I still choose to be a virtuous Calvinist. And if the decision theorists don’t like that, they can go to hell.
If God models your actions and pre-decides to save or damn you accordingly, sure. But if, as you say, behavior is caused by elect status, it seems to me that’s exactly the same as the Smoking Lesion (where CDT is in the right).
I see a difference between the two cases as follows:
The Smoking Lesion, as I interpret it when I say it seems to be self-evidently a correct counterexample to EDT, only affects your craving to smoke. Your probability of getting lung cancer is therefore affected only by your craving to smoke. Once you note whether or not you have a strong craving for cigarettes, your choice whether or not to smoke provides no further evidence for your possession of the lesion than the craving alone did.
As I interpret Calvinism, your status as elect or reprobate doesn’t just cause a craving to live virtuously. It actually causes the entire decision. Unlike in the smoking lesion, my internal argument about which decision theory I use to decide the Calvinism problem is itself part of the evidence for whether I am among the elect or not. This gives trying to work out a decision theory about Calvinism a self-referential element that the smoking lesion completely lacks.
As I said to Eliezer, I would be much more confident that living a virtuous life is the correct decision if I noted that many Calvinists before me had engaged in this exact argument about decision theories, and every single one who eventually came to hold my opinion went to Heaven and every single one who rejected it went to Hell.
If we used the same constraint as in the Augustine’s Paradox thread—that God is interested in your decision-making algorithm and rejects anyone who lives a virtuous life for reasons other than their native disposition, then I agree with you and I would live a life of sin.
“Choose”?
Yes, choose!
I don’t find Cleonid’s comparison with run-of-the-mill determinism inappropriate. The position of particles during the infancy of the universe completely determined whether or not I will get out of bed tomorrow morning. In that sense, I am a get-out-of-bed Calvinist: my fate was already decided at the moment of Creation, and I am either one of the elect whose genes/environment/etc predispose him to rising, or one of the reprobate whose genes/environment/etc doom him to stay in bed all day.
And yet I wake up tomorrow morning, and I find myself presented with what seems like a choice to get out of bed or not.
I am happy to say I “choose” to get out of bed, and I am happy to say I would “choose” to be a virtuous Calvinist. Even though I’m imbuing the word “choose” with less mystical force than a believer in free will might.
(here I am assuming a relationship between spiritual causes and the soul that’s similar to the relationship between physical causes and the brain you explained in Thou Art Physics. As far as I know, Calvinists believe that God’s choice is mediated through a change in the effect of original sin on the soul of the person He has chosen, which seems to fit the bill.)
How can I make this clearer...okay. Let’s say there have been a trillion Calvinists throughout history. They’ve all been rationalists and they’ve all engaged in this same argument. Some of them have been pro-sin for the same reasons you are, others have been pro-virtue for the same reasons I am. Some on each side have changed their minds after having listened to the arguments. And of all of these trillion Calvinists, every single one who after all the arguments decides to live a life of virtue—has gone to Heaven. And every single one who, after all the arguments, decides to live a life of sin—has gone to Hell.
To say that you have no reason to change your mind here seems to be suggesting that there’s a pretty good chance you will be the exception to a rule that has held 100% of the time in previous cases: the sinful Calvinist who goes to Heaven, or the virtuous Calvinist who goes to Hell. If this never worked for a trillion people in your exact position, why do you think it will work for you now?
Can one say that the God of the Calvinists as well as the alien of Newcomb’s Problem have the ability to perfectly predict (at least specific things about) the future?
And isn’t having that ability exactly the same as having a crystal ball that actually can look into the future? Isn’t then being able to predict the future with 100 percent certainty the same as having the ability to actually look into the future? Then, I think, it might be possible to say that the God or alien cannot be outwitted. Anything you do – no matter what – has been correctly predicted or is actually seen in God’s or the alien’s crystal ball. If you two box the alien has predicted just that, if you are not virtuos God has predicted just that. Your brain cannot change that. Your brain cannot escape perfect prediction. All escape attempts will trigger God to throw you in hell and the alien to leave you with just $1000.
I think this is to an extent even true if God or the alien are wrong some of the time – if they are only able to predict the future accurately 99 percent of the time. One would only be able to outwit the alien if one where to know under which circumstances the predictions of the alien break down. And as long as we are talking about random failures to predict correctly, a God or alien that is 99 percent accurate has still an almost perfect ability to look into the future. Has this implications for our decisions if we know that someone else can predict our own decisions with some accuracy? I think we should strive to know others models of our decision making and know when these models break down. That could be useful.
Not so. While you can never surprise a God that can perfectly predict the future, you may certainly outwit him. If the God limits his ‘rules’ to that which could be written on say, a few stone tablets then it is certainly possible to manouver around them. Alternatively, the God in question may be omniscient yet stupid. I’ve seen plenty of humans who will not change their strategy when confonted with absolute and trasparent proof that their strategy is flawed and will not work. I see no reason why a God could not do the same thing.
Yes, certainly, but that is besides the point. This problem here is about actually violating the rules and the question as to whether you can get away with it.
And it still depends on the nature of the the omniscient being in question. Having enough information does not make you make wise choices.
If you happen to know how the God thinks you can answer for sure whether or not you can outwit him. As you said yourself, “I think we should strive to know others models of our decision making and know when these models break down. That could be useful.” This applies to omniscient yet fallible Gods as well. Heck, you can know the entire state of the universe at all times and still not understand the Calvinist problem. In that case, dumbfounding the idiot god is trivial.
Again, that is certainly true.
But the Calvinist who decides to live a sinful life visibly violates the rules. Even a dumb God who only sets up a simple list of rules and who parses your behaviour only according to that list would notice that. I have certainly no doubts about the possibility of finding holes in the rules of God so that you would certify as virtuous even if other humans would most likely not see you that way.
But as long as prediction is concerned there is no outwitting. If you find holes in the rules that is not the same as finding out where prediction models break down (and if God can predict perfectly there is no point where his model will break down). I think that’s an important distinction to make. You can certainly outwit the rules (if they have holes), you cannot outwit prediction (if it is perfect).
Hrm… you seem to have better google-fu than I. So, what is “evidential decision theory” then? or at least, can you point me to whichever link you found that actually had basic description on it? Thanks.
Nick Tarleton’s “Smoking Lesion” link just below is the best introduction I’ve found so far.
Oooh, okay, I see, thanks.
But then, couldn’t one just say that one has to be more precise and ask: to what extent does the common cause for both induce the correlation in the context of mental algorithms that actually execute that sort of reasoning?
ie, simply include that as part of it. Wouldn’t doing that sort of thing automatically fix these sorts of errors?
I may be misunderstanding, but if it’s what I think it is, doesn’t one have to take a position like that simply by virtue of “I am part of reality. My decisions are not some transcendent divine free will from beyond all, but are part of the web of causality. They are not just causes, but also effects. The particular fact that I’m reasoning in this way is ALSO an effect, not just a cause. I am not an intervention, but a part of reality.”
Well, seems to me at least one of the following is true: I am somewhat misunderstanding what EDT is, or my reasoning above is flawed, or some flavor of EDT is actually Right Way.
My google-fu is weak. I could find stuff mentioning it, but not a basic “what is evidential decision theory”
So, what is it?