Dr. Retard

Some people look at 'fine-tuning' and infer God. Some people look at it and instead infer many worlds. It seems to me that there is no good reason for either inference -- there's nothing 'special' about life, there's no good way to gauge the probabilities at play, there's no reason for God to play Rube Goldberg, etc. But I want to leave all such objections aside. Assume that life-permission is special, terribly improbable, and that the best way for an omnipotent God to get it would be to play Rube Goldberg.

I want to examine this question: Assuming all the above, is it reasonable to infer many worlds from 'fine-tuning'? Roger White's article "Fine-Tuning and Multiple Universes" (<a href="http://www.nyu.edu/gsas/dept/philo/faculty/white/papers/ftmu.pdf" target="_blank">here</a>) argues that it's not. He claims that to infer many worlds is to commit the "inverse gambler's fallacy". I agree with him. Here's an illustration of the fallacy:

As in the gambler's fallacy, the flaw can be exposed by the slogan "dice have no memories". Each roll of the dice is stochastically independent from each other roll. One roll does not influence another roll. Consequently, even if the dice were rolled many times, that wouldn't make them a whit more likely to turn up double-sixes for the observed roll. Each roll is on its own.

Here are some similar cases:

There is a random number generator with a range from 1 to 100. You are allowed to see one run, and you don't know whether there have been other runs. 17 is the result. Can you reason like so? "17 is a quite improbable result -- one in a hundred. But if there were many runs of the machine, it would be more likely that 17 would result from some run or another. So there were probably many runs of the machine." No, you cannot. The probability of that run turning up 17 is one in a hundred, regardless of how many runs there were. The 'many runs' hypothesis receives no confirmation from the 17-result.

Here is a parallel example that might tempt you more: Another similar 1-100 random number generator is hooked up so that, if 17 results, it spits out fifty bucks for the observer; otherwise, nothing special happens. You are allowed to see one run, and you still don't know whether there have been other runs. Lo and behold, 17 is the result. Can you reason like so? "Wow! What do you know! I got 17 -- the $50 result! That's one in a hundred odds, how do you like that! I'll bet they've run that machine a lot of times; otherwise, it's too incredible that I happened to get the money". No, you still cannot. The hypothesis of many runs makes it not a whit more likely that 17 would result for that run. It's still one in a hundred.

Now, suppose that you walk up and run the machine, knowing full well that it has been run many times. You get a 17 and the cash. You should probably be quite unsurprised. After all, the machine has been run many times before and someone was bound to get the 17 eventually. It just happened to turn out that you got it.

Here are the results of the thought-experiments again:

1. Dull-6-6 Story: Unremarkable. 6-6 doesn't confirm 'many-rolls' hypothesis
2. Dull-17 Story: Unremarkable. 17-result doesn't confirm 'many-runs' hypothesis
3. Special-17 Story (# of runs unknown): Surprising! 17-result doesn't confirm 'many-runs' hypothesis
4. Special-17 Story (# of runs high): Unremarkable. 17-result made unremarkable by 'many-runs' hypothesis

There is no relevant difference between 1 and 2. Between 2 and 3, the difference is that the improbable result is 'special'; that's why we are more tempted to commit the fallacy. What's the difference between 3 and 4? In 4, we know that there have been many runs, whereas in 3, we don't know that. But why is one surprising and the other unremarkable?

The answer is that in 3, and not 4, we can explain the result by resort to a speculative hypothesis that denies our background assumptions. How? We can speculate that a benevolent programmer has messed with the machine to help you out. Then the result would not be improbable, since it wasn't random like you thought. Why can't we do this in 4? Because in 4, with so many runs, the benevolent-programmer hypothesis explains nothing. After all, there is no reason for him to help out with your run instead of all the previous runs.

The moral is this: We cannot infer many runs of a chance process from the fact that an improbable result happened; the many-runs hypothesis cannot explain the improbable result. But sometimes, if we know that there are indeed many runs, then an improbable result is thereby made unremarkable; this because the many-runs knowledge screens off the efficacy of possible speculative hypotheses.

The moral for the fine-tuning argument (maybe): If we are surprised and astonished by the improbable life-permitting result, then we cannot explain it satisfactorily by invoking the possibility of the many-worlds hypothesis. However, if we find out that there are indeed many worlds, then we should lose our surprise at the improbable result. After all, if there are many worlds, why should God choose this world instead of any of the others?

Now to some cases I'm having trouble with. Consider coinflipping examples where you just flipped heads 30 times in a row.

5. Suppose you are, so far as you know, the only coinflipper in existence. Can you infer from your all-heads results that there are many other people flipping coins? Is that many-coinflipper hypothesis confirmed by your all-heads results? I say No.

6. Now suppose that you are surrounded by many other coinflippers. Should you lose your surprise at your all-heads result? I say No.

In 5 and 6, it looks like the many-coinflippers hypothesis is worthless. In 5, the all-heads result does not confirm the many-coinflippers hypothesis. In 6, where you know that the many-coinflippers hypothesis is indeed true, it does nothing to remove your surprise. So what's the difference between 3-and-4 and 5-and-6?

You might say this is the difference: The salient speculative hypothesis in 5-and-6 is that your coin is loaded. And knowledge of many-coinflippers does not screen off the efficacy of the loaded-coin hypothesis. The loaded-coin hypothesis still explains your results well. In contrast, knowledge of many-runs does screen off the efficacy of the benevolent-programmer hypothesis. The benevolent-programmer hypothesis doesn't explain your results, because there is no known reason for him to favor you over everyone else.

But is this really the difference? Is it really true that the loaded-coin hypothesis explains your results? After all, there's no reason why your coin would be loaded and not that of one of the other coinflippers. In that way, perhaps, the many-coinflippers hypothesis really does screen off the efficacy of the hypothesis.

But this conjecture seems completely fishy to me. If we know of many runs, the benevolent-programmer hypothesis seems useless; but even if we know of many coinflippers, the loaded-coin hypothesis seems useful. The question "why would the programmer help you and not someone else?" seems to neutralize the benevolent-programmer hypothesis. But the question "why would your coin be loaded and not someone else's?" seems silly.

Can anyone help? Any thoughts?

P.S. I'm 'stealing' a lot of this from a Wikipedia article I wrote on the inverse gambler's fallacy.

[ October 12, 2002: Message edited by: Dr. Retard ]</p>

Clutch

The "this-universe" objection to MW strikes me as confused.

Universes are individuated by their contents/descriptions/conditions. On MW, there is no surprise in the fact that, out of all the universes, this one contains life. That's what makes this universe this universe, and not some other one. It's not as if, having individuated this universe by all its properties, we should then be surprised that among those properties is this universe's life-aptitude. That something satisfies the law of identity seems rather shallow grounds for surprise.

excreationist

The main evidence for the Many Worlds Interpretation is quantum physics experiments though...
See <a href="http://www.newscientist.com/hottopics/quantum/quantum.jsp?id=22994400" target="_blank">New Scientist - Taming the multiverse</a>
That article also has this quote:
That doesn't necessarily mean the Copenhagen interpretation is wrong though...

Richard Feynman has the same idea with a different name - he calls it <a href="http://www.ma.org/sciences/Lanik/Time_Travel/page6.html" target="_blank">Multiple Histories</a> rather than multiple or many worlds....

Anyway, the MWI was only first proposed in 1957 - by Hugh Everett III. It probably would have been used to explain quantum physics.... fine-tuning is just an extra thing that the theory/hypothesis can explain - fine-tuning isn't the main evidence for MWI.

Thomas Metcalf

Originally posted by Dr. Retard:

"But this conjecture seems completely fishy to me. If we know of many runs, the benevolent-programmer hypothesis seems useless; but even if we know of many coinflippers, the loaded-coin hypothesis seems useful. The question 'why would the programmer help you and not someone else?' seems to neutralize the benevolent-programmer hypothesis. But the question 'why would your coin be loaded and not someone else's?' seems silly."

Suppose that instead of a room full of coin-flippers with their own coins, there's one coin and everyone passes it around. Every time anyone receives heads, she gains $1, and every time she receives tails, she loses $1. Every time you flip it, heads comes up, and other people's results are predictably unpredictable. If this were the situation, I think most people would ask, "Why should the coin work this way for me and not for anyone else?" To me, we are safer in avoiding the "Gamblor zaps the coin every time I flip it" hypothesis here than we are if everyone has a different coin and no one gains anything from the heads result.

"Why should only my coin be loaded?" doesn't seem to screen off the hypothesis of a loaded coin, but "Why should the coin only give me $1 every time?" works better, I think. This (one coin, passed around) scenario seems to work just as well as an analogy with the formation of these multiple universes.

Think of Craig's sharpshooter analogy. Suppose I'm part of a large number of executions. If there are one million executions scheduled for that day, and the sharpshooters' guns all happen to jam for me, I'm still not really justified in concluding that someone tampered with them. Why would the Unnamed Jammer choose the guns of my executors and no one else's?

Also: Does "many worlds" result in an easy, intuitive way around the sharpshooter analogy? Anyone in the "guns jammed" shooting range would think "Gosh! Somebody Up There likes me." So why didn't God reach down and save the people in the other shooting ranges?

Jobar

Interesting topic, but I don't think EoG is really the proper place. Let's try Science & Skepticism.