The "fine tuning" argument: A "meta" look

[Rumraket (Mikkel R.)]

@ThomasTrebilco

Ok, well this is just deeply uncharitable.

No, it’s true.

[Faizal_Ali]

@ThomasTrebilco

This was followed up by a two person interview in which he was joined by Luke Barnes and a philosopher called Wade Tisthammer on the subject. It’s very thorough, but it’s about an hour and a half long.

2+ hrs of videos in response to Adams’s puddle story? FFS.

[ThomasTrebilco]

@Rumraket

No, it’s true.

Keep in mind the context of the post!

Wade sets out to identify and fix an issue with Paulogia’s lottery parody. The low prior was deliberately made for the purposes of illustrating a particular point, it was not Wade making a case for God.

@Faizal_Ali

2+ hrs of videos in response to Adams’s puddle story? FFS.

Yes, it is deeply unfortunate that people think it is a good rebuttal! There would be no need for people to make responses to it otherwise.

People should look to the relevant experts instead

[Faizal_Ali]

@ThomasTrebilco

I think the substitutions are fine. In the absence of anymore posterior evidence, the conclusions that socrates being able to dunk the ball is evidence he is 15’ is justified.

The conclusion is not warranted because of other posterior evidence - but the parody breaks down for the fine tuning argument because it’s not clear that there is any other posterior evidence.

That’s exactly the point. Without the “posterior evidence” the Argument from Slam Dunking Socrates arrives at a conclusion that we know to be wrong.

Which means the burden is on anyone who tries to argue in the absence of such posterior evidence to demonstrate his conclusions are not similarly worthless.

[ThomasTrebilco]

@Faizal_Ali

That’s exactly the point. Without the “posterior evidence” the Argument from Slam Dunking Socrates arrives at a conclusion that we know to be wrong.

Which means the burden is on anyone who tries to argue in the absence of such posterior evidence to demonstrate his conclusions are not similarly worthless.

Unlike frequentist stats, Bayesian probability works on a principle of belief revision. It is a perfect encapsulation of sciences provisional knowledge; accepting the conclusion of our hypothesis tests until new evidence arrives to say otherwise. Because it’s in a Bayesian calculus, the FTA typically presented is logically equivalent to a standard scientific hypothesis test.

As an atheist you don’t need to be made uncomfortable by this though. You could take a view of the multiverse like most of the new atheists (E.g. dawkins) and Sean Carroll.

But even if you think the multiverse fails, you’re still not compelled to accept theism simply on account of the FTA succeeding. Like I said above, most theists accept that the FTA is but one piece of evidence among many.

We have accounts of false-positive evidence elsewhere for other topics; in Bayesian terms, you could simply hold that - all things considered - there are other factors which cause the evidence to be weighed towards naturalism, even though the FTA succeeds and is good evidence for theism.

You just need to be prepared to tell people what you think that other evidence is

[John_Harshman]

@ThomasTrebilco

Can you point to me where he says either that:

1. We cannot calculate probabilities for the cosmological constants
2. That the fine-tuning of life for individual universes are very small

Before I point you, did you at least watch the video? It’s only 6 minutes long, and you expect other people to rebut a two and a half hour video.

Your first proposition resembles what we’ve been talking about, but your second here seems incoherent. Please rephrase.

What makes you think that a multiverse theory entails or requires some particular distribution of possible constant values?

[Faizal_Ali]

@ThomasTrebilco

Unlike frequentist stats, Bayesian probability works on a principle of belief revision. It is a perfect encapsulation of sciences provisional knowledge; accepting the conclusion of our hypothesis tests until new evidence arrives to say otherwise. Because it’s in a Bayesian calculus, the FTA typically presented is logically equivalent to a standard scientific hypothesis test.

I don’t think so. A proper scientific hypothesis would not lead us to conclude Socrates was 15 ft tall because he could dunk.

It seems to me apologists often see Bayesian probability as some Magic Special Sauce that can salvage a bad argument. But the principle of GIGO overrules any formula.

[ThomasTrebilco]

@John_Harshman

Before I point you, did you at least watch the video? It’s only 6 minutes long, and you expect other people to rebut a two and a half hour video.

Your first proposition resembles what we’ve been talking about, but your second here seems incoherent. Please rephrase.

What makes you think that a multiverse theory entails or requires some particular distribution of possible constant values?

I thought it was the full debate which I have seen before - my bad. I’ll give it a watch

I don’t necessarily expect people to watch the long video. I mainly just posted it as a resource, but also the paper that @Rumraket has interacted with for ease.

Yes, bad wording for 2.

Rephrased:

Can you point me to where he says either that:

1. We cannot calculate probabilities for the cosmological constants.
2. That the probability that any single universe is fine tuned for life is not incredibly small

Bolded bit there because I needed to specify more with #2, and in the right direction.

@Faizal_Ali

I don’t think so. A proper scientific hypothesis would not lead us to conclude Socrates was 15 ft tall because he could dunk.

Sure it could. It doesn’t though, because the evidence doesn’t stop at the considerations you put in the syllogism.

@Faizal_Ali

It seems to me apologists often see Bayesian probability as some Magic Special Sauce that can salvage a bad argument. But the principle of GIGO overrules any formula.

I’m not sure people think Bayes can salvage a bad argument. There is discussion of whether it can, or should be used for non-scientific hypothesis, such as in history or metaphysics, but I think that’s a different issue.

[Faizal_Ali]

@ThomasTrebilco

I’m not sure people think Bayes can salvage a bad argument.

You are correct, and I did not mean to suggest that. I should, rather, say that apologists often do not know how to use Bayesian probability and so end up arriving at incorrect conclusions.

Another example: William Lane Craig tried to demonstrate that Jesus’s resurrection really happened with Bayesia probability. I can’t recall the details but it was along these lines: If Jesus was not resurrected, then the disciples died for something they knew was not true. It is very unlikely that someone would die for something they knew was not true, therefore Jesus was resurrected.

You could just turn that around: If the disciples did not die for something they knew was not true, then Jesus was resurrected. Resurrections are very unlikely, therefore the disciples died for something they knew was not true.

[ThomasTrebilco]

@Faizal_Ali

You are correct, and I did not mean to suggest that. I should, rather, say that apologists often do not know how to use Bayesian probability and so end up arriving at incorrect conclusions.

Yeah - a lot of people don’t seem to use it properly. I am not trained well enough in it to feel confident using it, so I don’t - but others still do

From what I can tell, I think Richard Swinburne uses it quite well in his book ‘the existence of God’. IIRC he has some other publications on this too. He’s considered a maverick by most who review his work (including many of his atheist peers).

I’m not familiar with craigs work on this, but I’m not sure that it is as reversible as this.

I need to head to bed now, but I’ll finish watching the video @John_Harshman later and think more about your reversal here then too @Faizal_Ali.

Have a good day/night guys

[John_Harshman]

@ThomasTrebilco

Can you point me to where he says either that:

1. We cannot calculate probabilities for the cosmological constants.
2. That the probability that any single universe is fine tuned for life is not incredibly small

He says the first. He doesn’t say the second; what he says is that we have no way of calculating the probability, and thus can’t say whether it’s large or small.

[Rumraket (Mikkel R.)]

@ThomasTrebilco

Maybe you could look at what they actually say instead of dismissing the response merely because of the concluding sentence of the piece.

I did. I read the whole article, and it does not point out any flaw in Paulogia’s video.

@ThomasTrebilco

Or maybe there’s not really any interest in taking fine-tuning seriously? I’m not being condescending here when I say I can understand that and respect it, @Rumraket.

Or maybe you are just mistaken. Could that be it?

@ThomasTrebilco

Unlike frequentist stats, Bayesian probability works on a principle of belief revision. It is a perfect encapsulation of sciences provisional knowledge; accepting the conclusion of our hypothesis tests until new evidence arrives to say otherwise. Because it’s in a Bayesian calculus, the FTA typically presented is logically equivalent to a standard scientific hypothesis test.

Yes, and in a Bayesian calculus, you have a hypothesis that is supposed to account for the evidence. The hypothesis has a prior probability if you have relevant background knowledge(we don’t for Gods creating universes, nor for them arising naturally, it is wholly outside our experience).
Otherwise you might use what is called intrinsic probability. In considering the intrinsic probability of a hypothesis before evidence is included in the Bayesian calculus, you must be making some sort of estimate of the probability of the hypothesis in question.

We don’t really know what the intrinsic probability of theism or naturalism is, but one common method is to just take the space of all “allowed” possibilities.

With respect to the values of the physical constants and the initial conditions of the universe on one hypothesis of naturalism, the values of the physical constants and initial conditions we see is then drawn at random from that space of all possible values for these. So in effect, the naturalistic hypothesis of “random chance” has an incredibly low prior probability, because it is assumed there is an enormous range of allowed values for the physical constants and initial conditions of the universe. So it must be 1 in the number of all possible combinations for these values. Which must be an incredibly small number.

Now here’s the problem with the fine-tuning argument for theism: It is simply assumed, but never actually shown, that theism has a much greater intrinsic probability. But does it?

Well we now have to consider the case of all allowed values for theism. The space of all possible Gods. Gods that don’t want universes with life but more Gods, Gods that want universes with life but not like ours, Gods that want coca-cola cans made of ectoplasm, Gods that want nothing. The “allowed” range for Gods is infinitely large(and you’re welcome to insist until your head explodes that Gods by definition want our universe, good luck with that), hence the intrinsic probability of the theistic hypothesis of a God that wants our exact universe has an infinitely low probability. Doesn’t look good, does it?

This is known as the fine-tuning argument against the existence of God. No multiverse needed.

[T_aquaticus]

@ThomasTrebilco

The FTA is purported by many atheists to be the best argument in natural theology precisely because it causes a lot of tension for the naturalist position.

I was reminded yet again of a Steven Weinberg quote, so I did a google search and it found an old post of mine at Biologos on a very similar topic. Go figure:

For an analogy, suppose that there is a planet called Earthprime, in every respect identical to our own, except that on this planet mankind developed the science of physics without knowing anything about astronomy. (E.g., one might imagine that Earthprime’s surface is perpetually covered by clouds.) Just as on earth, students on Earthprime would find tables of fundamental constants at the back of their physics textbooks. These tables would list the speed of light, the mass of the electron, and so on, and also another “fundamental” constant having the value 1.99 calories of energy per minute per square centimeter, which gives the energy reaching Earthprime’s surface from some unknown source outside. On earth this is called the solar constant because we know that this energy comes from the sun, but no one on Earthprime would have any way of knowing where this energy comes from or why this constant takes this particular value. Some physicist on Earthprime might note that the observed value of this constant is remarkably well suited to the appearance of life. If Earthprime received much more or much less than 2 calories per minute per square centimeter the water of the oceans would instead be vapor or ice, leaving Earthprime with no liquid water or reasonable substitute in which life could have evolved. The physicist might conclude that this constant of 1.99 calories per minute per square centimeter had been finetuned by God for man’s benefit. More skeptical physicists on Earthprime might argue that such constants are eventually going to be explained by the final laws of physics, and that it is just a lucky accident that they have values favorable for life. In fact, both would be wrong. When the inhabitants of Earthprime finally develop a knowledge of astronomy, they learn that their planet receives 1.99 calories per minute per square centimeter because, like earth, it happens to be about 93 million miles away from a sun that produces 5,600 million million million million calories per minute, but they also see that there are other planets closer to their sun that are too hot for life and more planets farther from their sun that are too cold for life and doubtless countless other planets orbiting other stars of which only a small proportion are suitable for life. When they learn something about astronomy, the arguing physicists on Earthprime finally understand that the reason why they live on a world that receives roughly 2 calories per minute per square centimeter is just that there is no other kind of world where they could live. We in our part of the universe may be like the inhabitants of Earthprime before they learn about astronomy, but with other parts of the universe instead of other planets hidden from our view .(Weinberg, S., “Dreams of a Final Theory,” Pantheon: New York NY, 1992, pp.252-253. Emphasis original)