Monday, August 24, 2009

The Fine-tuning argument (part 1)

One of my favorite apologetic arguments is the Fine-Tuning argument. It goes like this:
"If God exists, there is a high probability that he would create sentient beings such as ourselves. You know, because we humans are so cool, and any old god would want us to exist. If God does not exist, then there is an extremely low probability that sentient beings like ourselves would come to exist. We know this because even if the laws of physics are necessarily configured in the way they are, there are still a bunch of fundamental physical constants whose values are arbitrary. A small change in any of these physical constants is unlikely to produce humans.

"So, given that humans exist, the probability of God is high."
Let me explain first why I like this argument, and then, later, why I disagree with it.

I like this argument, because it involves physics, and gets people interested in physics. Now everyone is asking what a "fundamental physical constant" is. Well, I'll tell you! Most people think of things like "the speed of light" or "the planck constant" as fundamental physical constants, but in fact they're not. Those constants depend on our choice of units; they depend on our definition of a "meter", or a "second", or a "kilogram". So they are not really fundamental. All fundamental physical constants are essentially a ratio between two things. For example, the ratio of the mass of an up quark to the mass of a down quark. Or another example, the ratio of an electron's classical orbit speed to the speed of light (this is known as the fine structure constant). By John Baez's count, there are 26 such fundamental physical constants. Most of them are related to mass in some way. Incidentally, gravity is the least-understood physical force in quantum field theory--perhaps some of the physical constants will be eliminated when we figure quantum gravity out?

Many of the counterarguments, too, discuss physics. For instance, it can be argued that many of the fundamental constants are absolutely irrelevant to life. For instance, the mass of the top quark has absolutely no effect on us. The top quark has a lifetime of 5 x 10-25 seconds, and is not a component of any known object. Creating and detecting a top quark was a task that required some giant particle accelerators and many very clever physicists. Other arguably "useless" constants include the masses of neutrinos, muons, tauons, bottom quarks, charm and strange quarks, the CKM and MNS matrices, and the cosmological constant. But that's not all of them. In any case, even if all the fundamental constants were completely irrelevant, there's still the matter of the equations which relate the constants to physical laws.

Another physics counterargument is, how can we know that there would be no life without these particular fundamental constants? If you think we can just take a new set of fundamental constants and predict the prospects of life, then you have sorely overestimated the power of physicists. A prediction so complex must be informed by experiments and empirical observations. But we don't have any of those; we can only observe and experiment on our own universe. What we can do, however, is look at a specific process which led to life. For instance, the triple-alpha process allows helium nuclei to fuse, forming carbon, which is, as far as we know, essential to life. The triple alpha process seemingly depends on a lucky coincidence: carbon-12 has a quantum energy level which happens to be at 7.65 MeV. If we changed the fundamental constants, would this lucky coincidence disappear? Would there appear any new lucky coincidences to circumvent the process? I wouldn't claim to know.

A third physics counterargument is the idea of a multiverse. If we live in a multiverse, then our universe is simply one of many universes with different fundamental physical constant. We can restate the Fine-Tuning argument as follows:
"If there is a multiverse, there is a high probability that at least one of the universes would allow for our existence. If there is only one universe, then there is an extremely low probability that sentient beings like ourselves would come to exist. We know this because even if the laws of physics are necessarily configured in the way they are, there are still a bunch of fundamental physical constants whose values are arbitrary. A small change in any of these physical constants is unlikely to produce humans.

"So, given that humans exist, the probability of a multiverse is high."
We can use the same argument for the existence of a god, or for the existence of a multiverse. So why should we prefer god over the multiverse?

I must say, my physics intuition tells me that this is a terrible argument for the multiverse. The reason physicists are entertaining the possibility isn't because they just thought up the multiverse and said, "Wouldn't that be philosophically satisfying?" At least, I hope that's not how it happened. I would hope that physicists thought it up because serious cosmological theories required or suggested it. And if we wish to support the idea, we must support it with real scientific experiments or deductions, not just some rhetorical argument. A lot of this cosmological theory is still up in the air, so I wouldn't jump the gun at this time.

If I find the Fine-Tuning argument for the multiverse so unconvincing, one wonders if the Fine-Tuning argument for a god is really any better.

However, none of these are my primary disagreement with the Fine-Tuning argument. I'm a physicist in training, so I would prefer not to commit to the idea that most constants are irrelevant, or that life could exist given different constants, or that the multiverse does or does not exist. Therefore, my primary response to the Fine-Tuning argument has nearly nothing to do with physics, and instead relates to logic.

...To be continued!


smijer said...

Isn't the multiverse originally an alternative to the Copenhagen interpretation and only incidentally related to the fine tuning argument?

Of all the arguments for God, the fine tuning argument gets the best marks in my book - though I still think it is poorly reasoned.

miller said...

Nope, what you're thinking of is the Many Worlds interpretation of quantum mechanics, which is completely unrelated to the multiverse. The multiverse is the idea that in different parts of the universe, there are different physical constants. The Many Worlds interpretation is the idea that there is no wavefunction collapse. People get them confused frequently.

But I agree that the multiverse is only incidentally related to the fine-tuning argument. I believe that we should reject or accept the idea of a multiverse on the basis scientific evidence, not questionable philosophical arguments.

smijer said...

I hadn't thought about it that way, and you're right. But I think there is room for overlap. Depending on what constrains constants (and/or how symmetry is broken historically) - a multiverse could have the same properties as Many Worlds, couldn't it?

miller said...

Yes, I'd think so.

DeralterChemiker said...

As a chemist, I always heard that the most fortuitous constant for the origin of life was the ratio of the density of ice at 0° to the density of liquid water at 0°. This causes ice to form on the top of water instead of at the bottom, the way solids form in most other liquids when they freeze. If water solidified from the bottom up, life would never have formed. Or so the argument goes. And if there were no life, who would care if there were a god or not?

DarkSapiens said...

I think that sometimes the possibility of any self-replicating entity to emerge is too related to the only examples we have data about— Earth's lifeforms.

Perhaps we could use an argument similar to the Fine-Tuning one that would be like this:
"If there's a lot of different planetary systems in the Universe, each one with slightly different conditions and properties, there is a high probability that at least one of the planets would allow for our existence. If there is only one planetary system, then there is an extremely low probability that sentient beings like ourselves would come to exist. We know this because even if the laws of planetary formation are necessarily configured in the way they are, there are still a bunch of events and parameters that are arbitrary to happen. A small change in the history of the planet is unlikely to produce humans.
So, given that humans exist, the probability of a universe with multiple planetary systems is high."

We know that in fact there are a lot of planets out there, and then we can use this argument to the universe itself and the physical constants to make a similar reasoning about carbon production and such. If I had to bet I'd say that if there's life in other systems it will be probably based on carbon, because this element is relatively abundant to perhaps give time to life based in whatever "less efficient" element to appear. But we don't know. As you state in the second counterargument, I can imagine sentient life forms in another universe asking the same questions and saying that if carbon-12 had a different quantum energy level perhaps there would be produced in such large quantities that they would have never appeared.

So my point is that our existence doesn't have to imply there's a God or a multiverse. We exist because in this universe there are the suitable conditions, but the reverse thing seems unlikely to me.

Perhaps this problem would be solved reaching a Theory of Everything that predicts the values of those constants. I'll wait to read the second part to see what you have to say :)

Sorry for the large comment, by the way…


miller said...


The second part will not be proposing any theories of everything. Sorry! :)

Yes, there's a parallel Fine-Tuning argument which relates to our planet. Our planet seems so perfect for life, it had to be made. This argument completely fails now that we know about the sheer number of galaxies, stars, and planetary systems. Of course, it's still an issue worth thinking about, from an astrobiology perspective, since it would be nice to know the density of intelligent civilizations in the universe.


I am wondering if that is really the best example of a cosmic accident. Does the precise ratio really matter, or would it be fine as long as ice is less dense than liquid water? Would it be fine if, instead of floating ice, we had a warmer planet?

Also, you could say that in chemistry, we already have found a "theory of everything" so to speak. We already know about the deeper laws which lead to the ice to liquid water density ratio. Theoretically predicting that ratio is probably an intractable problem, but who knows? The ratio may be insensitive to the parameters of the deeper laws. It may not be much more of a cosmic accident than the fact that pi is about 3.14.

DarkSapiens said...

Yes, I knew that wouldn't be the point. I suppose you will explain your response related to logic. And yes, I want to read it ;)

I hope Kepler (the mission) discovers something interesting in the next three years about how rare is our own planet. It would be nice to discover other Earth-like planets, even if they don't harbor life. But I've never been convinced by the arguments of the Rare Earth hypothesis, as I let you see in my other comment :P

The case of water is interesting, because it's the same property what makes it such a good ingredient for life and at the same time causes (ordinary) ice to have less density than its liquid form —hydrogen bonds. I think they could be actually predicted theoretically… you would just need to simulate the properties of its molecules and the thermal movement in a computer, isn't it? Would be a numerical approach, but I suppose a valid one.

By the way, I've been reading your blog backwards since the day before yesterday, and I like it. I'm a friend of Mireia, who recommended it to me ;)

DarkSapiens said...

I think you'll enjoy this:

It's a very good webcomic, by the way ;)


Chris said...

Interesting post, but I disagree with your physics intuition. I think the fine-tuning argument is an extremely strong argument for multiple universes.

I'm not entirely sure I get your first objection, but the fact that some constants are not fine-tuned doesn't seem (to me, at least) to punch any holes in the argument. If there's a 10^-60 chance of life being permitted because of one parameter, the fact that another set of parameters has no effect is irrelevant, isn't it? There's still only a 10^-60 chance of a universe fostering life.

Your second objection, that we can't know that the parameters would not permit life were they different, is, I think, true to an extent. We can't be absolutely certain, but in the case of at least one of the parameters we can be fairly sure. The force of the big bang was such that, had it been 10^-60 stronger, the universe would have collapsed on itself within a few seconds; had it been 10^-60 weaker, it would have expanded too quickly for anything but very sparse hydrogen molecules to have existed. Now, we can't be sure that life cannot exist in either of these circumstances, but it at least seems a good bet.

The objection of multiple universes is what I'd plump for against theism. But if indeed the fine-tuning argument does establish [Many universes or Creation], while this isn't quite the conclusion a theist would like, it's certainly more than most atheists would like to concede – particularly if we’ve no reason to prefer the many universes hypothesis over the creation hypothesis.

(Incidentally, some have argued that Ockham's razor dictates us to choose creation over multiple universes. The multiple universes hypothesis posits many more things than the creator hypothesis, with no additional explanatory power. Personally I don't dig this argument.)

On your other blogpost you mentioned that the theist makes the "having it both ways" fallacy. I don't think that's right. You say that if the natural laws are life-friendly, theism is more probable, and if they're not, theism is more probable, so no matter what you observe God is more likely than previously thought. But it's not that we observe that the natural laws are life-friendly and so God probably exists; it’s that we observe that the natural laws are life-friendly and that's extremely unlikely in a singular, uncaring universe. Had we observed that the natural laws were life-frienly and that was extremely likely in a singular, uncaring universe, God would be (slightly) less likely. The fact that the parameters are unlikely is the key in this argument, not that they’re life-friendly.

My biggest gripe about this argument as one for the existence of God is that, even if it concludes that this universe is the result of creation, it doesn't follow that the Christian God exists and the Bible is his infallible word. If the argument does in fact establish creation, it does not establish that the creator was singular, conscious, omnipotent, omnibenevolent, omniscient or called YHWH. It certainly doesn't prove that Genesis is 100% literal, that gay marriage is wrong, that you go to Heaven when you die, or that Barack Obama is the Anti-Christ. The "creator" conclusion is so weak that it's kind of bizarre that so many monotheists put so much stock in this argument.

If you're interested, Nick Bostrom (Oxford philosophy lecturer/researcher) wrote a chapter defending the fine-tuning argument for multiple universes, available online here - chapter 2.

I fully applaud your hatred of Karl Popper.

miller said...

The first two objections I consider rather weak, but I point them out because I have seen them made elsewhere, and I think they're interesting to consider. For instance, one wonders whether having extra "useless" laws and parameters is more in spirit with a guiding god who carefully chose everything to be just so, or more in spirit with a random sample which is filtered through the WAP. But it's more interesting than persuasive, IMO.

On Ikeda and Jefferys' argument... I see your point, but I'm still not convinced that it works. I've tried adding a new statement, Q, such that P(F|Q&N)<<1 (if that makes any sense), but I think it fails to imply anything without several new assumptions which I don't necessarily agree with. This is pretty much what the paper said too, though I wish they had gone over it more in depth. I might contact them about it, perhaps when I have more time.

I'm not sure why I neglected to mention the counterargument that the god argued for does not necessarily have any of the properties we normally attribute to God. It's sort of a general counterargument that applies to most arguments for God. Perhaps I neglected it because I felt I had discussed it a lot already elsewhere.

miller said...

As for the multiverse, I could elaborate more on my intuition. In theoretical physics, you never want to get too attached to a particular philosophy. Though the philosophical argument here may involve very large probability ratios, there's a nonvanishing probability that it's completely invalid. The universe doesn't have to follow our ideas, no matter how elegant. So heuristically, I don't trust an idea like that.

From another perspective, it's not a good argument for the multiverse because it just isn't useful. There are all sorts of possible multiverse theories, and fine-tuning fails to distinguish between most of them. This is very similar to how fine-tuning fails to tell us anything about God.

And another perspective, more technical... The argument is too sensitive to prior probabilities. Usually, we pick uniform priors, but uniform in what variable? Uniform in N, where N is the number of universes in the multiverse? Or uniform in log N, or in 1/N, or what? Without a more specific multiverse theory, it's hard to tell. (And then we get into even tougher statistical problems, ie see here.)