The Fine-Tuning Argument (FTA) is one of those standard arguments for the existence of God. The argument goes that humans can only arise when the parameters of the universe are tuned exactly right. And while it’s possible that we just got lucky, the argument goes that it’s far more likely that God did the tuning.
The standard way to talk about the FTA is delve into a bunch of math equations. Not that there’s anything wrong with math, but here I wanted to write an in-depth overview that doesn’t talk about the math. There will, however, be a lot of physics. The goal here is not to refute the FTA (although refutations will occur incidentally), but to explore it, and to understand how we test hypotheses about the universe.
Outline
(Links to be added later)
1. The Fine-Tuning Argument: A walkthrough
2. Prediction distributions and inflation
3. Ignorant hypotheses
4. Anthropic reasoning
The parameters of the universe
The core premise of the FTA is that the universe is fine-tuned. Which is to say, the probability of life looks like this:
The probability of life is sharply peaked. x0 marks the parameters of the universe that we have in the real world.
The “universal parameters” on the horizontal axis can represent multiple things. There are the 26 fundamental constants of nature, such as the masses of elementary particles, and strength of the fundamental forces. There are derived constants which could in principle be calculated from fundamental constants. Finally, there are historical facts about the universe, such as the energy density at the largest scales.
The claim is that for at least some of these universal parameters, the probability of life is sharply peaked. Which is to say that if you change the parameter by a small amount, then the probability of life drops significantly.
What does it mean to change a parameter by a “small” amount? By one standard, the change in a parameter is “small” if it’s a small percentage of the parameter’s value. For example, if we change the inverse fine structure constant (~137) by 1, that’s small; if we change the cosmological constant (~10-122) by 1, that’s large. It’s not clear that this standard is always the most appropriate one, but it’s the one we use in most cases.1
What if it’s not fine-tuned?
As a former physicist, my impression of fine-tuning is, “reasonable hypothesis, but dreadfully difficult to prove”. In the actual world, we know that we have one set of parameters for the universe, and we know that it’s fairly difficult to make predictions from those parameters. In fact, most physicists dedicate their careers to making such predictions. There are far fewer physicists who work on counterfactual parameters of the universe, and those physicists can’t even rely on experiments for help. Thus, you might expect that the scholarly work proving that our universe is fine-tuned is relatively superficial. I say “relatively”, because perhaps the work is still deep enough to get lost in, deep enough that neither you nor I have enough time to review and understand the scientific literature. But we need to remember that this is an inherently difficult problem full of pitfalls, and we may never know the true answers.
One version of the FTA claims that life is very sensitive to the amount of energy released by hydrogen-to-helium fusion. If this were to change by 1 part in 7, then we’d either have all hydrogen or no hydrogen, and life would be impossible.2 But other physicists argue that if we were to change this parameter by tweaking the strong coupling constant, this would also change the binding energy of deuteron, which also affects the density of hydrogen–and before long we’re in the weeds.
I think it’s very easy to tweak one universal parameter by a little bit, and see that it negatively impacts some particular process that was important to life as we know it. But if the behavior of the universe is so sensitive to small changes, doesn’t that just demonstrate how hard it is to make these predictions? What if we tweak the parameter even more, and some new substitute process appears? What if we tweak multiple parameters at once? What if life is still possible, but it occurs on different kinds of planets near different kinds of stars, at different times in the history the universe? What if there were entirely different forms of life, unrecognizable to us? In short, maybe the graph looks more like this:
The probability of life is sharply peaked, but perhaps ours is not the only peak. (Image generated by adding a bunch of random peaks.)
It doesn’t particularly matter how sharp the peak is, if you don’t know how many peaks there are. In fact, the sharper the peak, the more it demonstrates the difficulty of the problem, the more it demonstrates our ignorance of the landscape.
Granted, if the cosmological constant is so large that everything gets pulled to pieces, then it’s hard to imagine life existing in any form. So it’s still possible that the landscape is as flat as we had initially presumed.
Of course, when we talk about the FTA, usually nobody wants to dig into the physicists’ arguments. So we just take fine-tuning for granted and move on. And that’s what I’m going to do too in this series. But let’s take this moment between posts to dwell on the humility of our knowledge.
1. One common exception is the cosmological constant. Since the cosmological constant is on the order of 10-122, you’d think a change of 10-120 would be very large. However, some theories of physics predict that the cosmological constant should be on the order of 1, and it’s actually very surprising that the cosmological constant should be so small. Thus, any change less than 1 is considered small, and any change larger than 1 is considered large. This standard is used in fine-tuning arguments, and it’s used to argue in favor of alternative physics theories. (return)
2. This argument comes from astrophysicist Martin Rees, in his book Just Six Numbers. The energy release of hydrogen fusion is 0.7% of the mass of the initial hydrogen atoms. Rees claims that if it were as low as 0.6%, everything would be hydrogen; if it were as high as 0.8%, hydrogen would all disappear. This is an example of a “derived constant”, which can in principle be calculated from the fundamental constants. (return)
I don’t know if you care, but you can use latex in the FTB version of WordPress, so rather than using
instead.
10^-122, you could change it to $latex 10^{-122} $ and getYou can also use <sup> </sup> (superscript) and <sub> </sub>(subscript) tags (commenters cannot, so I can’t demo how it would look). 10<sup>-122</sup>
You might eventually get to some of this, so I’m sorry if this is jumping the gun. But I’ll put it out there anyway while it’s on my mind.
Some people just seem to take it for granted, then move on to criticizing a later part of the argument. (There are lots of bits to criticize, so I don’t really blame them.) But I don’t think it’s obvious that having fine-tuned parameters is good evidence for god(s) anyway. I mean, if a god wanted to make life and part of this process involves picking a set of physical parameters (or laws, initial conditions, etc.), there’s no reason why it should’ve selected from finely-tuned ones rather than coarsely-tuned ones. There’s just no logic behind it, and the claim is that there’s some kind of logic behind it. The end. That’s the short and sweet version….
Suppose we had found that small (or large) changes to our actual parameters would’ve also been life-permitting (what I’m calling a “coarsely-tuned” set, but I put basically no limits on how coarse it ought to be as long as it’s not “fine”). That’s not evidence against the claim that a supernatural entity put things together that way. So it wouldn’t be evidence for this entity, if/when we see fine-tuning. The one simply doesn’t depend logically on the other, or at least I don’t see it and have never seen anyone else explain how that’s supposed to work. So how is this argument even a thing?
Maybe it wanted the choice to be taken from some very tightly constrained set of parameters. Maybe it was showing off or whatever. But then again, maybe that’s not what it was like at all. We just don’t know. On the contrary, it very well could have wanted a large palette of colors to pick from, so to speak, any one of them good enough to paint the right sort of picture (i.e. make life). If its goal is just to make life, then there isn’t an issue either way, because it ends up making life either way. Note that in the “coarsely-tuned” case, it seems natural to say that this god was making it much easier on himself, not harder, since less precision is needed to hit a valid target. So you might be able to convince yourself of that, if you wanted to, then be disappointed that we found finely-tuned ones in the real world.
But the real point is that we have no information about it and no apparent reason to draw such conclusions from one set of observations or the other. We can invent all sorts of stories about it, like I just did, which could rather trivially tell you contradictory things, and that obviously leads nowhere. I’m sure some see it differently, but to me it just doesn’t feel like one of the options (fine-tuning) has any sort of logical force behind it that’s lacking in the alternative(s). Why exactly is this making it more likely or more believable that there’s a god? You simply have more arbitrary assumptions that some people are tempted to make (maybe without even realizing they’re doing it), about what some mystery entity must have been thinking or how it must have gone about the process of choosing the parameters from the entire space of possibilities. (We also don’t really know much about what that space is like, but that’s another matter.) How is that sort of thing, if that’s what these fine-tuning proponents have, supposed to move the argument any closer to a valid conclusion? It doesn’t seem like it could. If they wanted to say it’s not a baseless assumption but they have special access to god’s intentions or personality or whatever, in the form of revelations, mystical visions, etc., then they can say that, but there would be no point in making fine tuning arguments in that case. They’d be making a very different argument about very different things, and that’s what we’d have to think about instead. The fine-tuning shit would just a bunch of superfluous noise.
On that note, speaking of superfluous noise…. I’m sure nearly everybody who thinks fine-tuning arguments are compelling were already believers to begin with. It’s convincing for some people, pretty much as long as they were already convinced, but not so much otherwise. So this is definitely more of an academic exercise than anything else. But it does seem confused in various ways, and if what you wanted out of this deal was to give your beliefs a veneer of intellectual credibility, I would think you’d want it to not be a confused mess. So it’s probably still worth challenging it for that sort of reason.
(or even: $latex \frac{1}{10^{122}} $ to get
)
@Owlmirror,
Yeah I know about the sup/sub tags and latex. I just didn’t do it. There’s no rationale for that so I just added them in.
consciousness razor @2,
Yeah, the coarse-tuning objection. I recently read about that on Use of Reason, who was responding to philosopher Hans Halvorson. I mention this objection later in the series, but it’s only a mention, so this is as good a place as any to talk about it.
The problem with the coarse-tuning objection is it’s kind of hard to formalize. I’m not sure how you would state its premises and conclusions as Bayesian equations. You could say that if God exists, he would prefer a coarse-tuned universe (ie P(~F|G) >> P(F|G)), and this implies that F is evidence against G (ie P(G|F) > P(G)). But depending on who you ask, P(G|F) isn’t the right number to be looking at, and you should really be looking at P(G|F,L,R).
(I’m using G = “God exists”, L = “life exists”, R = “the universe is right for life”, and F = “P(R) is sharply peaked at specific values of universal parameters”.)
I mean, it’s not really the rebuttal’s fault for being so messy. The fine-tuning argument itself is very handwavy, and people can’t even agree whether the “evidence” is L, R, F, or some combination. Of course the rebuttals are going to be handwavy too. I don’t have any strong reason to think that God would strongly prefer a coarse-tuned universe, but also, proponents don’t have any strong reasons to think otherwise, so maybe the whole argument is inconclusive, which is exactly what the critics are trying to say.
Still reading, but I came across this bit:
that happened to be separated by an awkward line break. Is there a way to get wordpress to treat certain text as a single word, and so not subject to auto-breaks? It might be helpful in a couple places in your text. Mathematical expressions are rarely as easy to interpret when spread over more than one line.
Okay, done. And now I see that others have brought up that particular mathematical expression as well, for reasons other than line breaks. interesting. And now I feel foolish for having commented before reading the rest of the comments.
Next:
Again, this may be in your plan, but something new occurred to me while reading your post. I’ve never been a fan of the argument, mostly because I know how ignorant I am of physics. I can’t even list the fundamental constants, much less explain why we expect that these are fundamental and will not be soon shown to be dependent on other, more fundamental physics. Therefore, I’m not in a position to judge the fine-tuning and therefore not in a position to say whether fine-tuning exists, and therefore saying it supports any particular god hypothesis is just weird. Likewise, I think it’s weird for anyone to say it without understanding all the physics, and I doubt there’s a single theologian out there who does understand it all. So I don’t take the arguers themselves seriously because I have good reason to believe that they don’t understand their own argument.
This, of course, was my thinking even before you made your “multiple peaks” argument.
But the something new is this:
Let’s say that there are universes with single-peak parameter sets. Each universe can be judged to be within the peak where a significant chance of life exists or outside that peak, and thus the universe will be dramatically unlikely to have life.
If you’re an omnipotent god, you don’t give a fuck about whether life is unlikely in a particular universe or not, and, since the peak is supposedly narrow, nearly every universe in which you might create life is going to be outside the peak. So you grab the nearest universe, create some life in it, and BOOM! now you have life in a universe that is almost guaranteed not to be fine-tuned for life.
BUT, if you’ve got no god and life will only appear in a universe if it has a parameter set conducive to life, then nearly every universe with life will be a universe where the parameters are located under the peak, giving a much higher than average chance for life.
In other words, given the god hypothesis and the no-god hypothesis, and adding in the information about fine tuning theologians assert to be true (which information we don’t have, but which some people suppose we do), which hypothesis is more likely? Obviously the no-god hypothesis is more likely because the vast majority of universes in which life arises without the assistance of a god are going to be “fine tuned” universes, but the vast majority of universes in which god poofs life into existence are going to be non-“fine tuned” universes. So our odds of being in a fine-tuned universe are actually greater in a no-god universe than in a god-universe.
Logic, how does that god again?
Well, no.
It’s hard to imagine life existing for very long. There’s still going to be a period of time, brief though it may be, where stars and planets form.
But assuming the the CC is fundamental and not dependent on anything else about the universe may allow us to say that this time would appear short to us, subjectively, compared to the amount of time it took complex life to evolve on earth.
But suppose Planck time is related to the CC (doesn’t matter which is dependent on the other) in such a way that for the CC to be much larger, tP would have to be much shorter. In this way, the number of time quanta that pass during the period of star and planet formation might be the same (or nearly so) regardless of how much the CC is stretched. From the subjective perspective of an observer in any of a number of various universes in which the CC varies relative to other universes, there might be the same amount of time available to form and evolve life.
I’m not saying this is true, of course, I’m just saying that when you say it’s difficult to “imagine” a universe with high-CC and simultaneously large chance of intelligent, multi-cellular life, I think you’re dramatically underestimating the possibilities available to the human imagination.
Crip Dyke @6-8
I can’t reproduce the line-break problem so I think it might be browser-dependent. Might be fixed if I switched to latex instead of sup tags.
I first heard this argument from a paper by Ikeda and Jefferys where they go into all the Bayesian equations, and argue that this defeats the Fine-Tuning argument. Basically, the argument is that if God exists, then L would be true regardless of whether R is true (L & R being defined in comment #5). Thus R is not evidence in favor of God, it’s evidence against God.
My issue with their argument is similar to the issue I brought up with the coarse-tuning argument in #5. I agree that R isn’t good evidence, but what about F and L? Ikeda and Jefferys argue that L doesn’t count as evidence because reasons. But again, the FTA is handwavy, the rebuttals are handwavy, nobody knows anything.
I was thinking of a scenario where the cosmological constant is ~1. I’m not sure if this is right, but as a rough estimate, I think an object a meter away from me, would accelerate away at a rate of 1 meter per planck time, per planck time. I don’t think planets or stars could form ever, and probably even atoms are in trouble. I don’t think this could be fixed by changing the timescale under consideration–by what, 120 orders of magnitude? But I haven’t thought about it too much.
I’m not clear on the distinction between L and R. Do you mean that R is potential for L, even if not actual L? I could understand that it’s permitting life in some sense, even if it just so happens that life doesn’t actually come to exist. That would be different from a set that doesn’t even have that kind of potential. Something like that? No matter.
Let me try to go step-by-step, in a fairly realistic chain of reasoning that some person might make, rather than only a bunch of abstract symbols, since it’s sort of easy to get lost. We first discover L (and/or R). That’s presumably one of the first things we’re going to learn: there’s life. Then we discover F, if that’s what the relevant physics tells us.
L is evidence for G, which is to say it raises P(G). The basic reasoning is that G is basically defined to be powerful enough to do whatever it wants, and along with that comes the stipulation that one of the things G supposedly wants is L. We just take that as a given about G. We see L, so there you have it: P(G) is bigger, even if not much bigger — no logical problems in that and no complaints from me.
But how does F raise it any more when conjoined with L? It would be incoherent to say that ~F also increases P(G), more than L all by itself, just like F does. We’re stuck with one or the other (or neither since these may be irrelevant, which is my contention), definitely not both. So now we need an explanation of why F is the correct and logically necessary choice, rather than ~F.
Another way to phrase it: what is it about F that implies P(G) increases, over and above the increase from L that we already accounted for, something which isn’t also satisfied by ~F? I don’t see why it should do that. I think people are just confused by the structure of the argument somehow: they think they’ve got a good line of reasoning, but it’s illusory.
It’s definitely handwavy. But I wouldn’t say it’s a fine-tuning argument if you only had L (or R, if there’s a difference). There’s not necessarily any fine-tuning when you just look around and see life exists. You’d need something that suggests fine-tuning for life, something along the lines of these physical parameters. (And there may be other things that would work just as well, like a special cosmological structure or whatever. Think of a geocentric universe, for example: that by itself would look “tuned” for us, independently of fundamental physical laws, because we’d be in a special place and stuff could all be coordinated to allow for us to live in there.)
There’d be nothing to discuss if it were F all by itself, without L. I mean, one could imagine a lifeless universe with parameters that were nevertheless finely-tuned for life. It would just be the case that when god threw that dart he missed the small targets that would’ve produced life, yet the dartboard still has the structure that it has. Of course, we wouldn’t be here to talk about it, but that’s conceivable anyway. Maybe god wanted to make rocks instead, and he managed that with these parameters (maybe they’re fined-tuned for rocks or maybe not), so he’s not concerned about it. Anyway, that’s not the situation we’re worried about.
One more way to put it, using the dartboard analogy some more, since that’s a helpful picture for me right now. You’ve got this dartboard, which could have many different patterns. It could be one with a giant target that’s easy to hit, or many mid-sized targets or whatever you like. There could be lots of tiny targets that are easy to miss, with a huge amount of space where no points are scored. Any possible pattern at all. Life existing would be a dart which has landed inside a target zone – we know we’ve got that, so we can move on and start looking at other stuff. Now you find out something about how the dartboard looks, which could be any logically possible pattern whatsoever as I said, and the actual one we find will be some specific pattern or another.
The fact that you see a certain type of dartboard (as opposed to some other type) doesn’t seem like it should boost the odds that there is a god throwing darts at it. When you see the dart in a target zone instead of elsewhere, you could think to yourself, “well, how about that? there’s life; maybe that’s a clue that a god was aiming for it, since the god I thought existed would do that sort of thing.” But looking at how the rest of the dartboard is structured isn’t making that case any stronger, is it? Plus, that tells us nothing about other possible ways it could have been structured — those are still hanging out there as our “coarsely-tuned” ones, which would (by hypothesis) have given the same results if this god had any role to play in the story. What people are noticing is that the targets are small…. If there were a god aiming for them, he’d need to have good aim in order to hit it. And maybe he would have good aim, if there is such a god. But that doesn’t imply that there is a god aiming at it. Right? We just know that there’s this dart on a board. We know it could’ve been somewhere else, and we know the board could’ve looked differently. There’s still a big leap to make, because fine-tuning arguments just don’t look like they’re actually doing that.
The fine-structure constant is 1/137, not 137. That’s why we can do perturbation theory in QED. 😉
CR @10,
The distinction between L and R is there because it comes up in other objections. Crip Dyke’s objection in #7 was basically raising the possibility of L & ~R. But it’s not relevant for your argument, you can safely ignore it.
I would say that the “evidence” used by the FTA is the conjunction of F and L. This can be confusing because proponents of the FTA will often take F for granted and only consider L as evidence; or else they might take L for granted, and only consider F as evidence. And you can look at either F or L individually and say that neither of these seem like good evidence. You have to consider them both. And it shouldn’t really matter whether you take F first or L first.
So, in your intuitive way of thinking about it…
Could you try thinking of it in the opposite order? Consider the hypothetical where you first discover F, and then discover L. Does this lead to the same conclusions? I mean, just as a consistency check.
Or perhaps, do you have difficulty accepting the hypothetical? I’m sensing hints of the anthropic principle in your reasoning. And I’m not saying that’s wrong, I’m just trying to clarify your/our thoughts about it. Is your objection really related to coarse-tuning, or does it have more to do with the anthropic principle? If you took only the anthropic principle, or only the coarse-tuning objection, would either of those be enough on its own to defeat the FTA?
In the dartboard analogy, there isn’t really an anthropic principle since obviously we can observe darts whether or not they hit any targets. So, suppose we go to a bar with two friends, one who likes throwing darts at targets, and the other who likes throwing darts randomly. If we observe a dart on a target, we might guess that it was thrown by our first friend–unless the bar is covered wall-to-wall with targets. So, the simple statement that the bar is not covered wall-to-wall with targets seems relevant, does it not?
And the “coarse-tuning” objection is kind of like saying, well if this guy planned to throw darts tonight, and if he likes hitting targets so much, maybe he would have dragged us to a different bar where the targets are easier to hit. I mean, that might be true, or maybe there are other factors involved.
Rob Grigjanis, @11
Oops!
I’ve heard somewhere a very different sort of argument. The mathematics of probability has a certain structure, and there are certain requirements before you can argue that particular events are “likely” or “unlikely”. The fine-tuning argument essentially says that if we imagine that physical constants are actually variables, and we twiddle the dials to imagine (in full mathematical detail) alternate universes that have different physical constants, very few of them look hospitable for life. Therefore, the argument goes, the Universe we observe is “very unlikely” and must have been created deliberately to allow life. The counter argument that I heard, protests that the fact that we can IMAGINE alternate universes, EVEN imagine them in full mathematical detail, is not EVIDENCE that such alternate universes actually exist. It is not even evidence that such alternate Universes are possible. Imagined universes ARE NOT OBSERVED universes. We only observe ONE universe, and you cannot say anything about probability or “likelihood” with a sample size of one. All you can infer is that the universe we observe is “likely enough to have happened once”. The fine-tuning argument is a category error…. it is an attempt to use a particular concept, probability, at a place where it simply does not apply.
johnhodges @14,
That’s touching on frequentism vs Bayesianism. The view you expressed is the frequentist point of view, while the Fine-Tuning argument requires that you take more of a Bayesian point of view. I’m more of a Bayesian myself, I’m pretty much taking it for granted in this series, and I suspect that Bayesianism is the more popular view hereabouts. Bayesianism vs frequentism has implications that reach far beyond the FTA, so you wouldn’t really want to make a decision about it based only on how it impacts the FTA.
Anyway, “Bayesianism” and “frequentism”, those are your search terms if you’d like to learn more.
Siggy at #15,
Thank you. You are the first person who has given me any sort of reply to that objection.
Is the Bayesian “fine tuning” argument equivalent to the “Argument from personal incredulity”? I.E. “I can’t imagine how such a thing could have happened naturally, therefore Yahveh must have done it”?
At any rate, thanks. I’ll probably (!) investigate B versus F at some time.
From the Hitchhiker’s Guide to the Galaxy:
“There is a theory that says, if ever anyone figures out what’s really going on in this Universe, it will immediately vanish and be replaced by something even more bizarre and inexplicable. There is another theory that says that this has already happened.”
Siggy:
That’s not the real sequence of events we’re trying to represent, of course, but I think it should be the same. I do understand that logically and probabilistically there’s no difference. Anyway, to be clear from the beginning, my claim is not that F decreases P(G), or that ~F is the evidence theists should hope to find instead, so this is different from the Halvorson argument you linked to earlier. Instead, it’s that F doesn’t increase it, and only L has a (perhaps very small) positive contribution. I don’t need to say L came first, but that is going to be the story that you’d tell about a person finding these pieces of evidence.
Neither, I guess….? The claim is that F (as well as ~F) is not good evidence for or against theism. They’re not relevant to the conclusion; the FTA commits a non sequitur. We are using the evidence we can gather about the world, when we see (L + F) to figure out if we can conclude anything about G or ~G; but that’s not invoking the anthropic principle (weak or strong), just routine conditionalization. That’s pretty much what everybody is supposed to be doing, when dealing with the FTA.
I think some people want to say that life (L) is unlikely if naturalism is true (~G) and we see F, which of course is not saying F increases P(G) although it may sound almost the same. They think the space of (L + ~G) is mostly filled with ~F, while only a very small proportion of it is F, due to the characteristics of F and ~F. The remaining naturalistic part (~L + ~G) will definitely drop out, and you may do that whenever you like, because we certainly have L as some of our evidence. I won’t contest that we also have F as evidence, although we’re much less certain about it; I simply don’t think it gives theists what they want even if it is conceded to them.
The G side of the picture is somewhat different: there’s basically no part of G with ~L, because the theistic proposal is that God could/would make L either way (with F or ~F) and making L was his goal. So L is evidence for G, since the nonliving part of ~G will be knocked out of the picture when we learn L, but the entirety of ~G will remain as it contained no nonliving part. It’s easier to imagine learning L first and to express things this way, but in the end, doing it earlier or later doesn’t affect the logic or the resulting probabilities.
Now we can look at F and not ~F, for both the G and ~G sides. Why would F take up a larger fraction of G than of ~G? That’s the question. FTA proponents could try to argue that somehow, but I think they should be the same. No?
Well, I admit that this is where things can get a little confusing. I wanted an example that would make it easy to imagine a dart-thrower, just to give theists something they could work with. But the question is not which type of dart-thrower it was. The question is whether or not we have any type of dart-thrower at all. I think I get what you’re trying to suggest with the random dart-thrower, but it’s worth pointing out. All we actually know is that the dart is located in some target or other, and the board has a certain pattern. We don’t know that we came to the bar with a friend (or two or three) who likes doing things a certain way, and we may not have any friends at all.
The trouble is that this friend who “likes throwing darts at targets” isn’t merely attempting to hit a target, like a real-world dart game, with some chance that they miss in any given attempt. It makes no difference what the board looks like when this friend plays, because we’re supposed to assume that they can hit a target no matter how “difficult” it may be for any other imaginable person.
Here’s where we get to the key point. You don’t merely find a dart inside a target zone instead of outside (i.e. discover that there’s life). What you also do is discover that F is true: you look at the whole board and see it has lots of scattered/tiny/finely-tuned target zones or something of that sort. I agree that the chances for your target-loving friend go up, when you find the dart inside a target. Again, maybe not by much, but they do go up. But why should it add anything more to the chances that your friend is playing, when you learn the board has one sort of layout, if it’s stipulated from the beginning that they can always hit a target no matter what?
This stipulation seems like a fairly decent representation of the various claims being made. We’re supposed to understand that (1) god is very powerful and capable of making life independently of naturalistic constraints, hence why we’re told theism is a better solution to a problem that naturalism supposedly can’t resolve very well. However, it seems to imply (2) the chances of G aren’t increased when you learn F. With just the evidence provided, it looks like those chances only went up because of L, not because of F, since (G+~F) is just as good for L as (G+F) is. Yet for no apparent reason, the FTA asserts that F does something more than L all by itself, and it doesn’t depend only on L. That’s where they’re apparently being inconsistent or incoherent. Maybe there’s still a way around it; I’d want to know how theists think the argument does work, if perhaps I’ve misconstrued their claims or whatever.
Sorry, an important correction:
I meant the entirety of G will remain. For me, that wasn’t large to begin with, but all of that tiny thing is still there.
johnhodges @16,
Yes, that’s another rebuttal to the FTA. The way I would put it is: it’s certainly possible to come up with naturalistic explanations why a parameter of the universe should look exactly so. But the FTA doesn’t give us a chance to do that, because science is too hard and we’re impatient. I talk about this in part 3.
Consciousness Razor @18,
For the dartboard analogy, I think in order for it to be similar to the FTA, you really do need to consider both friends. The premise of the FTA is that we’re looking at exactly two hypotheses: naturalism, which leads to a random, broad distribution of universal parameters; and theism, which leads to a very narrow distribution of parameters. These premises, are of course, contestable. But if you’re trying to understand how the FTA even works at all, it seems like we should begin by granting the premises.
In the dartboard analogy, the two observations are: there’s only one dartboard with a small target (F), and there’s exactly one dart in the room, and it’s on the target (L). Now, this is not quite the same as the FTA, since in reality, we can’t know how many darts in the room, and all we really know is that there’s at least one dart on a small target. That is, in the FTA we have to apply the anthropic principle. But first, I’m trying to understand what the situation looks like without the anthropic principle, since you seem to be saying that your objection is not based on the anthropic principle (?).
Intuitively, it seems to me that if you observe F and L together, that increases the probability that it was my target-loving friend (G) rather than the other friend (~G, or sometimes we call it N for naturalism). Do you agree with this intuition?
I’m going to pull out a couple quotes from your comment that seem wrong to me.
F doesn’t really matter to the God hypothesis because God can create life either way, that’s correct. However, F does matter to the naturalism hypothesis, because naturalism can’t create life as easily under F. That’s how F can qualify as evidence.
I’m trying not to go into math , since that seems to be your preference. But in math terms, you are saying saying P(L|G&F) = P(L|G&~F), and therefore F isn’t evidence of G. I’m saying that F could still be evidence of G if we have P(L|~G&F) < P(L|~G&~F).
You’re right, I don’t really know how much of a fraction F takes up of either G, or ~G. But you also need to condition this evidence on L.
Again, in math terms… You seem to be saying P(G|L) ~ P(G) and also P(G|F) ~ P(G). But that’s not the relevant comparison. If we take L first, and want to consider the power of evidence F, then the relevant comparison is P(G|F&L) to P(G|L). If we take F first, and want to consider the power of evidence L, then we compare P(G|F&L) to P(G|F).
@siggy #9:
yeah, I guess when bringing it up it’s not so much that it’s a convincing defeater argument, but that it defeats the premise that high F is necessarily evidence for god.
Absolutely no one knows and it’s all handwavy, but if they can’t rely on their premise because it’s not proven, no longer intuitively obvious, and now at least significantly in doubt, then however valid the argument may be, it’s still far from proven sound.
In other words, it’s not an interesting argument because it proves anything. It’s an interesting argument because the mere possibility that F is bad for the god hypothesis sends them back to the drawing board and we get another decade of peace while the god-botherers argue amongst themselves about what their next great argument will be.
…ah! if only that were true.
Sure, but I still don’t think it follows that F by itself increases the probability. Simply by adding in L, it goes up. And I don’t think F makes it go back down or anything, since I’m saying it’s just irrelevant. Does F do anything to make it go up, or is this increase all a result of L? When you say F does this, that’s where I get off the train.
I wasn’t focusing on something before, so let me try another analogy, since I think it’s a big part of the issue.
Suppose you find my arguments reasonable. Tell me what you think the odds are, if you find my arguments reasonable, that you’ll say “wow, you’re a genius, CR; let me present you with a very prestigious award and a big prize, for having such a reasonable argument.” I think the odds of that happening are extremely low. It’s unlikely, even if you do think my arguments are reasonable (and unlikely if you don’t). It’s not like that would be evidence against the claim that you find my arguments reasonable, due to the fact that it was unlikely on that hypothesis. Being more excessive about it (or whatever) also doesn’t increase the odds that you find them reasonable, compared to anything else that would’ve sufficed just as well. What was actually observed (if you gave me the quoted statement above) simply was not one of the more likely possibilities; but it, like the others, would do the job just the same.
That’s the kind of mistake people can easily make with fine-tuning. It’s unlikely that there’s life, if there’s fine tuning and no god. Chances are slim that we’d have life, what with all that fine-tuning and no god to fix it up somehow and make life anyway, since it could use magic or whatever. You see life and fine-tuning, then think god. Right? Because this random-loving player who’s doing it probably wouldn’t hit the target. They’d probably miss.
But this doesn’t raise the chances of god, just like lavishing me with praise (and giving awards, prizes, etc.) doesn’t raise the chances that you don’t like my arguments, not on account of the fact that it’s improbable on the hypothesis that you do like them. Indeed, in this case it’s pretty straightforward how they raise (not lower) the chances of you liking my arguments, although of course it’s not always or generally an either/or situation, because many things are just plain irrelevant as evidence.
Life by itself does provide some help for theism, as we both agree. That would be like saying my arguments are reasonable, because it’s clear how this ought to affect my beliefs about the situation and doesn’t do anything more than what is required. But getting life in some very specific way, rather than in some other very specific way? Why was that supposed to be built into the idea of theism, first of all? And secondly, wouldn’t that be the same in either case? How exactly is this supposed to help theism, beyond what we already got with the observation that there’s life?
The R-loving player could have hit a target, just like the T-loving one, and it’s irrelevant how likely it was for the R-lover to do that. There’s just a question of why a certain dartboard (instead of any other) would lead you to think one rather than the other was playing. For that, you don’t ask whether a dart is likely to land inside or outside a target, because all you’re doing is looking at the whole board. Their playing-style, or how capable they are of hitting a target, or what have you, doesn’t affect that, right? So then what is looking at the whole board actually supposed to indicate, separately from asking where the dart is located on it?
I mean, is the claim not only that the T-lover wants to hit targets, not only can it always hit targets irrespective of the board design, but we can also be sure T-lover would’ve designed the board itself to be very difficult for the R-loving player to score (even though R-lover doesn’t exist and seems to have no bearing on T-lover’s actions)? I don’t understand why I should have to accept all of that.
separately, about the CC:
You’re imagining a universe where expansion = λ/tP. Okay, that does sound like a defeater to gravitic aggregation of mass, so long as the relationship between expansion rate and plank time stays constant.
Are we sure here that we (that is, everyone discussing the argument, atheists and theists) is on the same page as to the definitions of all of the terms used?
For example, “Life”. Would those who favor the argument agree that a universe that had nothing but bacterial mats growing and dying on any planet that life developed on, for billions of years until the stars of the planets they revolved around expanded and killed them all, would argue for a God that had created such a universe?
Or is there an unstated implication of intelligent life in there? Does the God being argued for just want metabolizing and mitosising cells, or a collective of cells that is self-aware and can communicate?
Another question that comes to mind is, how is God defined?
The God that seems to be argued for is something like, “an agent that has sufficient knowledge and power to create a universe [that can eventually develop [intelligent] life]”, as the weakest definition. A stronger definition would be “an agent that has sufficient knowledge and power to create a universe [that has [intelligent] life in it [regardless of whether that universe would develop it or not]]”. The strongest definition would be “an agent that knows any nonimpossible knowledge, and can do any nonimpossible thing [which includes as a subset the ability to create a universe which harbors intelligent life]”.
Theists want for the strongest definition to be true, but the fine-tuning argument only really supports the weakest definition, and I think the FTA could be said to argue against the strongest definition.
There’s a novel that develops some of these concepts called Calculating God, by Robert Sawyer. It uses many of the arguments favored by theists, including the fine-tuning argument, intelligent design arguments of various kinds, and so on, and blocks counterarguments such as the multiverse by having intelligent aliens state bluntly that there have only ever been a small number of universes (9, IIRC).
The god that the book ultimately presents is not the god of any religion at all. It is completely natural rather than supernatural — that is, while it can traverse higher dimensions, it has an origin from the matter of the first universe coalescing; basically some sort of Boltzmann Brain. It is neither omnipotent nor omniscient, but it has learned to create universes with different fundamental constants, and our universe is its most recent experiment. It wants intelligent life to exist because it has some sort of agenda with regard to said life, and has been guiding the development of life on various planets such that life will eventually develop intelligence.
CR @23,
I’ve said a couple times that I think the evidence under consideration is F&L, not just F and you keep on saying stuff like this:
So I’ll just say again, that’s not what I was saying.
At this point I should just ask you to formalize your argument.
Owlmirror @25-26,
Yeah, it’s usually implied that we’re talking about intelligent life. But we don’t talk about that because it’s not to either side’s advantage. Intelligent life is obviously strictly less likely than life, so if anything it helps the theist side. But we don’t know how much less likely, so the theist doesn’t want to bring it up lest they waste time trying to prove an irrelevant point.
If we accepted the FTA, of course it doesn’t really prove the god of any particular religion. I suppose at minimum it only proves some goal-oriented process, something that sets the parameters of the universe based on the predicted consequences of those parameters.
But this already a lot better than many arguments for the existence of god. The cosmological and ontological arguments generally only prove a thing, which could be nearly anything.
There are plenty of explanations for a fine-tuned universe besides a “god” — though a god is certainly one of them.
1. Of course there’s always the infinite multiverse theory.
2. In the future we invent time travel, then go back to the beginning of time and start up the big bang ourselves.
3. Perhaps, in a universe with some higher number of dimensions than ours, the formation and destruction of universes is a natural process that comes with its own set of rules and laws. Maybe they grow on bushes, like berries. With a sample of one, there’s no way for us to deduce what these rules are.
4. Even if you believe that the universe is created by some conscious being, there is no particular reason to be certain that the creator is the same as the person currently running the show. Perhaps universes are mass-produced somewhere, and some kid bought ours as a science project and is currently keeping it in a box under his bed.
5. This one is my favorite. Someday, if we don’t wipe ourselves out, we may be able to create universes ourselves. If the ability to create a new universe is an emergent quality of life, then a universe becomes equivalent to an organism, with intelligent life as its gonads. So the universe is fine-tuned to support life for the same reason our bodies are fine-tuned to reproduce — because of evolution.
I think you’re more than capable of understanding it in context, which you’re not doing here. I’ve made it very clear, multiple times, that I’m asking whether it does anything more, in addition to life.
– You have some amount that the god hypothesis increases, because we discovered life: P(G|L). This is not meant to represent the FTA, which involves additional premises, but it is already easy to see that P(G) goes up due to L. When you say “If we observe a dart on a target, we might guess that it was thrown by our first friend” this is what you must be talking about, since being on a target is merely being in a life-permitting region of parameter space, as a shorthand for life actually existing. That does not tell us about fine-tuning, since it tells us nothing about the overall pattern, the relative size of the target zone, etc. It’s only where the dart is found, in a target which is L. Are we clear?
– You have some amount that the god hypothesis increases, because we discovered both life and fine-tuning: P(G|L+F).
– Are those the two same amount, or did the latter give G an extra boost, beyond what the former already did for G? In other words, did we get some good evidence for G when learned F, or did we not? Why would you say we did? Especially when you also say “You’re right, I don’t really know how much of a fraction F takes up of either G, or ~G.” That seems to me to be conceding the point. If there’s no such comparison being made, then how is (F+L) doing anything for G, instead of doing something for ~G? If we had told you (~F+L) is true, would the story be any different? Why?
CR @30,
You seem unwilling to formalize your argument, which means you have to rely on me formalizing it, and I’m telling you that it doesn’t work.
The usual way to argue the FTA is to prove P(G|F&L)/P(~G|F&L) >> P(G|F)/P(~G|F). This can be proven formally from a certain set of premises. And it’s presumed that P(G|F)/P(~G|F) ≈ P(G)/P(~G).
If you insist on going the other way, taking L as evidence, and then F, then we have two possibilities:
(1) P(G|F&L)/P(~G|F&L) >> P(G|L)/P(~G|L) ≈ P(G)/P(~G)
(2) P(G|F&L)/P(~G|F&L) ≈ P(G|L)/P(~G|L) >> P(G)/P(~G)
It’s actually ambiguous whether (1) or (2) is correct. If P(F) ≈ 1, then we have case (2). If P(F) << 1, then we have case (1). You keep on saying L isn’t very strong evidence, which commits you to case (1), and that F isn’t very strong evidence, which is correct in case (2). I have been forced to agree that both of these statements are possible, but they are not simultaneously possible.
And you won’t formalize your argument, so all I can say is either you’re doing your math wrong or you’re rejecting one of the premises without saying so.
(I will provide proof of any step here on request)
No, I’ve been saying L is the only decent evidence in this mess. My problem has been with F. Is that just a typo or a very big misunderstanding? Did you mean to write “F isn’t very strong evidence”?
It may not matter, but what do you mean by “very strong”? Are you an ID creationist now, because life exists? I’m certainly not, and I assume you’re not. For me, given lots of other considerations, I had already thought P(G) is very small. Starting from there, L doesn’t do enough to make P(G)>P(~G). So I wouldn’t want to say L is that strong as evidence for G. But it’s not nothing, and I’ll happily admit that. Or maybe I’m being irrational or don’t understand just how strong it really is. That is possible.
Regarding the alleged significance of F, for me, it boils down to “I just don’t grok that.” I don’t see a reason to think F tells us anything useful about the question of G or ~G, in the way that L does: no coherent arguments or empirical data indicating a bias one way or the other, nothing.
I think the burden’s on you to explain (formally or informally, with words, pictures, numbers, whatever you like) how there is a rational way to convince yourself that fine-tuned physical parameters are good/probability-increasing evidence for the existence of a god. Or if not you, someone who likes the FTA. Such things had better help (at least a bit) to make the case for theism, if that’s what you think. Otherwise, we’re not really disagreeing about that, and it’s not clear to me where specifically you think the problem is. I know you have a few more posts lined up to discuss it, and perhaps some of that will give me a better idea of where you’re coming from. So I won’t make a big deal of it. I’ll just drop out for now and let you do your thing.
Sorry, that was a typo of my own. Your other option was already F, obviously. You might have meant “(F&L) isn’t very strong evidence,” rather than L, since I could definitely agree that it isn’t very strong, in the sense that I’m unconvinced and still an atheist. But the way you’ve set it up, you do seem to think I’ve been arguing that L is no good…. I don’t know where you got that idea.
To be extra clear, although this is repeating myself some more, I’m not questioning that we have observed F, which could then be used as evidence for this or that or the other thing. Physicists should decide that, and I’ll just take their word for it. I just don’t think it’s adding anything relevant in the argument for G. So, I have no issue with (2). How much greater P(G|L)/P(~G|L) is compared to P(G)/P(~G) is harder to say, but whatever.
CR @34,
I thought you were saying L was evidence, but not very strong evidence. For example, in #10 you say:
If you agree that P(G|L) is a lot bigger than P(L), then you’ve basically already conceded the FTA before we even got around to talking about F, and of course F is superfluous in that situation. If you think that P(G|L) is only a little bigger than P(L), then F is very strong evidence, as I said in #31.
You can wait for the rest of the series, but this is definitely not one of the things I write about. It is not one of the common objections to the FTA, and the objection doesn’t make sense, and the best way to explain why it’s wrong is to go through math, which I’m trying to avoid. I’d only ever really talk about it for your benefit.