Followup to: Simultaneously Right and Wrong

Related to: Augustine's Paradox of Optimal Repentance

"When they inquire into predestination, they are penetrating the sacred precincts of divine wisdom. If anyone with carefree assurance breaks into this place, he will not succeed in satisfying his curiosity and he will enter a labyrinth from which he can find no exit."

            -- John Calvin

John Calvin preached the doctrine of predestination: that God irreversibly decreed each man's eternal fate at the moment of Creation. Calvinists separate mankind into two groups: the elect, whom God predestined for Heaven, and the reprobate, whom God predestined for eternal punishment in Hell.

If you had the bad luck to be born a sinner, there is nothing you can do about it. You are too corrupted by original sin to even have the slightest urge to seek out the true faith. Conversely, if you were born one of the elect, you've got it pretty good; no matter what your actions on Earth, it is impossible for God to revoke your birthright to eternal bliss.

However, it is believed that the elect always live pious, virtuous lives full of faith and hard work. Also, the reprobate always commit heinous sins like greed and sloth and commenting on anti-theist blogs. This isn't what causes God to damn them. It's just what happens to them after they've been damned: their soul has no connection with God and so it tends in the opposite direction.

Consider two Calvinists, Aaron and Zachary, both interested only in maximizing his own happiness. Aaron thinks to himself "Whether or not I go to Heaven has already been decided, regardless of my actions on Earth. Therefore, I might as well try to have as much fun as possible, knowing it won't effect the afterlife either way." He spends his days in sex, debauchery, and anti-theist blog comments.

Zachary sees Aaron and thinks "That sinful man is thus proven one of the reprobate, and damned to Hell. I will avoid his fate by living a pious life." Zachary becomes a great minister, famous for his virtue, and when he dies his entire congregation concludes he must have been one of the elect.

Before the cut: If you were a Calvinist, which path would you take?

Amos Tversky, Stanford psychology professor by day, bias-fighting superhero by night, thinks you should live a life of sin. He bases his analysis of the issue on the famous maxim that correlation is not causation. Your virtue during life is correlated to your eternal reward, but only because they're both correlated to a hidden third variable, your status as one of the elect, which causes both.

Just to make that more concrete: people who own more cars live longer. Why? Rich people buy more cars, and rich people have higher life expectancies. Both cars and long life are caused by a hidden third variable, wealth. Trying to increase your chances of getting into Heaven by being virtuous is as futile as trying to increase your life expectancy by buying another car.

Some people would stop there, but not Amos Tversky, bias-fighting superhero. He and George Quattrone conducted a study that both illuminated a flaw in human reasoning about causation and demonstrated yet another way people can be simultaneously right and wrong.

Subjects came in thinking it was a study on cardiovascular health. First, experimenters tested their pain tolerance by making them stick their hands in a bucket of freezing water until they couldn't bear it any longer. However long they kept it there was their baseline pain tolerance score.

Then experimenters described two supposed types of human heart: Type I hearts, which work poorly and are prone to heart attack and will kill you at a young age, and Type II hearts, which work well and will bless you with a long life. You can tell a Type I heart from a Type II heart because...and here the subjects split into two groups. Group A learned that people with Type II hearts, the good hearts, had higher pain tolerance after exercise. Group B learned that Type II hearts had lower pain tolerance after exercise.

Then the subjects exercised for a while and stuck their hands in the bucket of ice water again. Sure enough, the subjects who thought increased pain tolerance meant a healthier heart kept their hands in longer. And then when the researchers went and asked them, they said they must have a Type II heart because the ice water test went so well!

The subjects seem to have believed on some level that keeping their hand in the water longer could give them a different kind of heart. Dr. Tversky declared that people have a cognitive blind spot to "hidden variable" causation, and this explains the Calvinists who made such an effort to live virtuously.

But this study is also interesting as an example of self-deception. One level of the mind made the (irrational) choice to leave the hand in the ice water longer. Another level of the mind that wasn't consciously aware of this choice interpreted it as evidence for the Type II heart. There are two cognitive flaws here: the subject's choice to try harder on the ice water test, and his lack of realization that he'd done so.

I don't know of any literature explicitly connecting this study to self-handicapping, but the surface similarities are striking. In both, a person takes an action intended to protect his self-image that will work if and only if he doesn't realize this intent. In both, the action is apparently successful, self-image is protected, and the conscious mind remains unaware of the true motives.

Despite all this, and with all due respect to Dr. Tversky I think he might be wrong about the whole predestination issue. If I were a Calvinist, I'd live a life of sin if and only if I would two-box on Newcomb's Problem.

New Comment
37 comments, sorted by Click to highlight new comments since:

We're forgetting signaling. Robin would never forgive us, because he sees it in a lot of things, and I happen to agree with him that it's far more pervasive than people think.

In fact, the Tversky example gives people two opportunities to signal: not only do they get to demonstrate higher pain tolerance [especially important for men], they also get to "demonstrate" a healthier heart. Both should be boosts in status.

The same goes for Calvinists: though, when you think about it, you truly believe in the elect, you don't think about it most of your life [as we know, much of our day to day life is subconsciously guided] and are instead focused on signaling your elect status with a good life.

For good measure, it even works with the car: you buy a new car to signal wealth to signal health.

However, I do believe that we engage in lots of automatic self-deception [making it easier to deceive others into believing we have higher status]: thus, we may actually believe that an extra car/a good life/a higher pain tolerance would improve your life expectancy/grace/heart, but that's merely the proximate cause. Ultimately, we're driven by status-seeking.

This was exactly my thought, and I now wonder whether it's possible to determine via experiment. So how do you give the information to the subjects but not have them think that the researchers know it.

A confederate who's a subject and just happens to gossip about the thing is one way -- if the researchers proceed to deny it, you might be able to split them into groups based on a low status confederate versus a high-status confederate, and a vehement denial vs a "that study hasn't been verified" vs a "that was an urban legend."

Or providing a status signal that it's better to have a "bad" heart -- having a high status researcher who says "sure, we may live less long, but there are all sorts of other benefits they're not telling us about"

It's really hard to separate the information from the humans passing on the information.

[-]roland170

If I were a Calvinist, I'd live a life of sin if and only if I would two-box on Newcomb's Problem.

I disagree. In Newcomb's problem the superintelligence's decision is conditional on your future actions whereas in the calvinist doctrine this isn't the case.

This is true. According to Calvinism, election is explicitly stated to be "unconditional." Election that is conditional on future knowledge of your actions is not Calvinism, but is closer to its theological opposite, Arminianism.

[-]Nebu70

Interestingly enough (to me, at least), I would live a life of sin, and yet one-box. Now I'm trying to come up with justifications for this gut reaction.

The best I can come up with so far is that I can see God having a "causal" relationship towards me. That is, I am exactly what I am because God made me that way. The Alien, however, isn't causal, merely observational. It didn't cause me to choose whether to one-box or two-box, it simply observes me and determines which I would do.

The reason I would sin is because it seems like the way the hypothetical situation is set up, we're assuming predestination, which seems to make me really not worry to hard about putting effort towards anything. If I was supposed to "fight" predestination, then things will turn out so that I do fight predestination. But until that happens, why not simply take the path of least resistance (especially since in a predestined universe, you cannot offer up any resistance at all) and enjoy whatever pleasures you wish? Language is failing me here, because I'm speaking as if I can actually make decisions or control my actions, but obviously I can't.

Also, it might be important to clarify that I would choose "sin" only in this hypothetical world where I somehow know with certainty that the universe was predestined. In reality, I'd be unsure about whether or not the universe was predestined, and thus wouldn't automatically choose "sin".

Contrast this with the Newcomb problem, where there is a possibility that I have "free will" or some sort of control over my actions. In this scenario, I want to be the type of agent whom the alien will reward, and I know the alien rewards one boxers, and so I choose to be a one boxer.

I can sense all sort of inconsistencies in my reasoning, and I can sense ways of making these inconsistencies more salient E.g. the alien is able to predict whether I will one-box or two-box before human life has evolved, and thus plants the two boxes somewhere on Earth during the dinosaur age in a location it knows no one will discover except for me (or perhaps plants it on Neptune, but sets up some chain of events so that the boxes will arrive to me by the time I'm in my mid-20s). Since the alien infallibly knew whether I would one-box or two-box, can I really in any sense say that my universe is not predestined?

Clearly there is confusion in my mind that I have to meditate upon and clear up.

If you know that the elect tend to live lives of virtue for various reasons, then the play is to live a life of virtue. If instead you know that the elect tend to live lives of virtue because they naturally want to, you may as well sin. I'm not familiar enough with Calvinism to say which of these is the case. (Well, neither is the case, of course. To say which they believe.)

[-][anonymous]50

If you were a Calvinist, which path would you take?

Answering as a hypothetical ideal rationalist thrust into a predestination connundrum then I must follow in Yvain's footsteps and live a life of sin if and only if I would two-box on Newcomb's problem. More specifically, I wouldn't live a life of sin because that'd be stupid. Hell bad. Paradise good.

But the real answer to "If you were a Calvinist which path would you take?" would be quite different. A Calvinist would take actions determined by the same cultural influences that made him a Calvinist. In most cases the abstract philosophical positions have very little influence on behavior. He would go ahead and take actions that protect his self image, without realising that this was what he was doing.

Given that I happen to have the mixed fortune of taking my verbally expressed beliefs rather literally, I would be a far more unpredictable Calvinist. I expect my path would follow much the same path as I did as my actual, non-counterfactual self. I would take the path of virtue, for reasoning much the same as Yvain's.

Yet that is hardly the end of the scenario. It turns out that being motivated by rational argument based on a literal belief is not nearly as robust as being motivated by self-image, or peer approval. Social motivation is stable, or at least it is dynamic in a way that it is considered stable. If you go around actually believing things then pesky things like evidence go around getting in the way, along with some confusion as to why other Calvinists don't seem to process this evidence in a vaguely sane 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.

I still choose to be a virtuous Calvinist

"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.

[-][anonymous]00

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.

[-][anonymous]40

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?

In chapter 0.6 of his book Evidence, Decision and Causality, Arif Ahmed also argues that Calvinist predestination is like Newcomb's problem.

One level of the mind made the (irrational) choice to leave the hand in the ice water longer. Another level of the mind that wasn't consciously aware of this choice interpreted it as evidence for the Type II heart.

I'm not sure that the mind isn't consciously aware of this choice why do you think so? The individual could be thinking "I'll try extra hard to keep my hands in longer, after all this is evidence that I have a healthy heart."

I think this is closer to what's really going on than the rest of the thread. Most people do not understand that "Heart type II plus exercise causes pain tolerance change" does not imply that they can cause themselves to have a different heart type by altering their pain tolerance after exercise!

On some level -- conscious or otherwise -- they believe the uncertainty exists in the territory, not the map. Then they seek to change reality by cheating the evidence.

Or, alternately, they know that they will feel better after the test if the test tells them that they have a healthier heart, so they act in such a way as to get that internal reward.

Human reward mechanisms just aren't set up properly; in this case, the internal reward isn't actually for living longer, it's for passing the test, so you try to pass the test. In some ways, it's similar to a standardized test in school; your actual aptitude or intelligence doesn't change based on how hard you try on the test, but you try hard anyway, because your reward mechanisms aren't based on your intelligence per se, they're based on the test score, so that's what you try to maximize.

After posting I realized this could be interpreted as condemning the one-box approach to Newcomb's. I think Newcomb's is a different story but I'm having difficulty identifying why. Perhaps causality under Newcomb's feels different because it involves a zero-uncertainty map.

(I dislike Newcomb's anyway. But I do one-box.)

[-]pwno30

I think the subjects just had a stronger incentive to test out their true tolerance. Without hearing about the two types, they probably just took their hands out to minimize discomfort.

That doesn't explain why subjects who thought a good heart would mean a lower post-exercise pain threshold took their hands out sooner.

Looking at the actual data from the article, since Yvain neglected to actually state the results of the second case. Subjects told that a good heart was correlated with higher pain threshold after exercise showed an 11.84 second increase in mean immersion time, while subjects told that a good heart was correlated with a decrease in pain threshold showed a 7.63 second decrease in mean immersion time.

You don’t have to mix Calvinism into this problem.

Determinism (a rationalist term for pre-destination) follows naturally from the laws of physics. Quantum mechanics throws a tiny element of chance into the play, but since you have no conscious control over the quantum effects anyway, it seems that all your future actions have already been taken out of your control.

With this additional indulgence from science, you can now embark on a life a sin. However, I am afraid that you may find it no more gratifying than the life of self-denial chosen by the Chosen.

Predestination, but not determinism, implies no binding between actions ("freely chosen" or not) and consequences. See Thou Art Physics and Timeless Control.

I believe this is not quite correct. Predestination does not imply that actions do not predetermine the consequences (i.e. “no righteous man can be saved without the grace of God”). Rather it implies that actions themselves are predetermined (i.e. “no man can become righteous without the grace of God”).

To quote Luther: ” God is said to hate men before they are born, because, He foreknows that they WILL DO that which WILL MERIT hatred”.

Quantum mechanics throws a tiny element of chance into the play,

QM's tiny elements of chance can be amplified as much as you like. See Feynman's Bomb

but since you have no conscious control over the quantum effects anyway

Whatever that memans. You cant predetermine an indeterministic event in your brain. but if an indteterministic event does occur in your brain it does not have to lead to action: other neural events could vote it down.

The subjects seem to have believed on some level that keeping their hand in the water longer could give them a different kind of heart.

Or they just made themselves believe they were feeling less pain than they were - or made themselves actually feel less pain - without having any causal hypotheses on any level.

It seems to me that the overwhelming majority of Calvinists are also irrational, in that their attitudes and actions contain contradictions which they make no attempt to explicitly understand or even acknowledge.

Furthermore, it is far from clear that Calvinism is compatible with rationality. If it isn't, in a hypothetical where I am Calvinist, I wouldn't be rational. I would likely do the same things that the majority of Calvinists do: claim to believe in cosmological determinism while acting as though my own thoughts and decisions (and thus their consequences for my behavior) weren't subject to determinism.

Just to make that more concrete: people who own more cars live longer. Why? Rich people buy more cars, and rich people have higher life expectancies. Both cars and long life are caused by a hidden third variable, wealth. Trying to increase your chances of getting into Heaven by being virtuous is as futile as trying to increase your life expectancy by buying another car.

This reasoning only works because buying cars is not perfectly correlated with living longer. If buying cars guaranteed that you would live longer, would you do it?

If buying cars guaranteed that you would live longer, would you do it?

Of course, but then you have made the correlation imply causation, and it doesn't. If number of cars owned and life expectancy was perfectly correlated, with no counter examples, that would be strong evidence for causation in either one direction or the other, but that isn't the case, and also wasn't the point of the example.