In Bayesian Counter-Apologetics I outlined why the Fine Tuning Argument actually disproves the existence of God. And I didn’t make that up. What I outline there was independently corroborated twice by the peer-reviewed research of multiple experts (see On the Bayesian Reversal of the Fine Tuning Argument). And I’ve formally laid this all out in my chapter on Design Arguments in The End of Christianity. The Fine Tuning Argument also suffers from a false hidden premise: that the “God” hypothesis requires less luck than competing explanations of the same observation; it doesn’t (see A Hidden Fallacy in the Fine Tuning Argument). But here I want to bring into focus the way Fine Tuning itself turns into a proof of atheism. This will in turn serve as a tutorial on some of the ways Bayes’ Theorem reveals how to employ the logic of probability correctly (as I recently demonstrated more generally in a peer reviewed article for SHERM).

Which God Do We Mean Here?

The conclusion is clearest, of course, with traditional definitions of God (the only kind of God anyone on Earth actually believes in), whereby God is someone who intends to create a universe for life and for that life to know They exist. But one could instead posit a Bizarro God, an all-powerful intelligence who deliberately decided to make the universe look exactly like a universe with no God in it, thereby deceiving us into the conclusion that God does not exist; or who had some other bizarre reasons (compulsory or voluntary) to do essentially that same thing—such as lacking any interest whatever in life and only, let’s say, wanting to make a universe that would generate black holes, and who is actually annoyed or indifferent to the mere accidental byproduct of that effort being life.

Bizarro God still falls to the same Prior Probability Defeater (far simpler godless theories exist that explain all the same observations). But it also falls to an additional Prior Probability Defeater: it commits the gerrymandering fallacy. I explained the logic of that fallacy in Bayesian Counter-Apologetics. Basically, any theory can be “gerrymandered” with multiple (usually bizarre) theoretical ad-ons so that it exactly predicts any evidence. Thus, I can gerrymander the theory that Donald Trump is a one-hundred-foot-tall lizard man from Alpha Centauri so that it exactly predicts all the evidence—including any possible evidence against that theory you could ever propose. I can gerrymander the theory that heirloom tomatoes are sentient angelic superbeings controlling world agricultural policy. I can gerrymander the theory that you are at this moment actually on the fourth moon of Jupiter. And so on.

Because all theories can be gerrymandered, the ability to gerrymander a theory cannot render it any more likely. And indeed, that is exactly what happens in the probability logic: all gerrymandering logically does is move an improbability in our equation from the evidence column to the prior probability column, producing no net gain in the probability of the theory—and often a serious net reduction in its probability. This is why all Cartesian Demon arguments fail to gain any credibility. And “Bizarro God” is just another Cartesian Demon. We can therefore simply dismiss this God out of hand. Lacking any evidence for it, it is simply always too improbable to credit. You can only change this assessment if you find specific evidence for that gerrymandered God—meaning evidence independent of what that gerrymander was created to explain, as for any other Cartesian Demon.

Therefore, I will not include under the term “God” here any gerrymandered and thus unfalsifiable Cartesian Demons. No evidence can ever make them more likely than alternatives, except very specific evidence of a kind that clearly we don’t have any of. Which is why no one believes in Bizarro God; not even the most lunatic or fanatical of theists. But in logical terms, I can ignore Bizarro Gods because any Bizarro God gerrymandered to turn evidence that’s otherwise only, say, 50% likely to be observed into evidence that’s 100% likely to be observed will drop in relative prior probability at least 50% at the same time, producing the same lower posterior probability of that God’s existence as for the Non-Bizarro God. I’ll illustrate this below. But the upshot is: you can’t escape the conclusion of the following argument by positing a Bizarro God. You just end up with an equally improbable theory.

The Odds on God Given the Observed Facts of Fine Tuning

The standard trick pulled by all Christian apologetics is to make an argument by leaving out all the evidence that would, if restored, entirely reverse that argument’s conclusion. I gave ten examples of this in Bayesian Counter-Apologetics. The Fine Tuning Argument does this by only looking at the arrangement of what are, actually, bizarre and logically unnecessary physical constants (everything from the relative strength of gravity and electromagnetism to the relative mass of the quark and electron), and noting that almost any other configuration of them would have prevented any life from arising in the universe (which is then argued to require intelligent design).

One can challenge that claim; but it isn’t necessary to. Because, as we’ll see, it doesn’t matter. So it won’t be challenged here. For the sake of argument, let’s just take it as if it were proved that “almost any other configuration of fundamental physical constants would have prevented any life from arising in the universe.” The real problem here is that this leaves out pertinent evidence. Because we are here testing two competing hypotheses to explain observations: either (A) chance accident produced that alignment of constants or (B) someone or something intelligently selected them. (The third possibility, that other arrangements are actually logically impossible, lacks evidence, but more importantly is ruled out by the premise we are granting so we needn’t consider it here.) So we have two hypotheses, and each makes a number of predictions (not just the one), and therefore to compare them requires looking at all those predictions, not just “cherry picking” the one single prediction we like and ignoring all the others that didn’t go the way we want.

So what different predictions do our two hypotheses make? Theory A predicts the following:

[T]he only way we could [observe ourselves existing] without a God is by an extremely improbable chemical accident, and the only way an extremely improbable chemical accident is likely to occur is in a universe that’s vastly old and vastly large; so atheism predicts a vastly old and large universe; theism does not (without fabricating excuses—a bankrupt procedure, as I already explained … ).

Similarly, the only way we could [observe ourselves existing] without a God is by an extremely long process of evolution by natural selection, beginning from a single molecule, through hundreds of millions of years of single cells, through hundreds of millions of years of cooperating cells, to hundreds of millions of years of multicellular organisms; so atheism predicts essentially that; theism does not …

Likewise, … we should expect [the universe we observe] to be only barely conducive to life, indeed almost entirely lethal to it (as in fact it is), since there are vastly more ways to get those universes by chance selection, than to get a universe perfectly suited to life throughout (indeed … by countlessly many trillions to one). Design predicts exactly the opposite (again, without a parade of convenient excuses).

Almost the entirety of our universe is a lethal radiation-filled vacuum, almost the entirety of its contents are lethal stars and black holes, and almost the entirety of what isn’t stars and black holes is a lifeless wasteland of rocks and dust on which nothing can naturally live. The universe is also billions (not mere thousands) of years old; and billions (not a mere handful) of light-years across; and life only slowly arose over billions of years of meandering, unguided natural selection from an initial, single, self-replicating molecule, which evolved into single cells, then into rudimentary colonies of cells, then into the advanced colonies of cells that we now call bodies; and exists now only as, indeed, a scaffolding of cooperating colonies of single-celled organisms (which we know as cells), an outcome only predicted by atheism—as that is the only way for intelligent life to arise without a God; whereas a God has no need of any such bizarre construction procedure, much less the billions of years of time it took.

So when we bring all the pertinent evidence back in, the evidence indicates support not for Theory B (intelligent design), but for Theory A (chance accident). Fine Tuning is therefore evidence against intelligent design. It could never be evidence for it, because gods don’t need fundamental constants at all, much less all the weird ones we have. No intelligent agent needs quarks or electrons or electromagnetism or even gravity—things can just behave as commanded or designed: where things need to fall, they just fall; where stars need to shine, they just shine; where things need to stick together, they just stick together. One might respond that, still, it is possible an intelligent engineer would choose all these weird and unnecessary ways to create and sustain life. But that is fully accounted for here. What matters is not whether it’s possible. What matters is how probable it is.

Because:

[I]f (a) we exist and (b) God did not design the universe, then (c) we should expect to observe several things, and lo and behold, those are exactly the things we observe; yet we do not expect to observe those things if God did design the universe. By definition that which is expected on x is probable on x; that which is unexpected on x is improbable on x. So if the evidence is probable if God does not exist and improbable if God exists, then that evidence argues against God, not for God.

Hence what matters is not what’s possible. What matters is its relative probability. In the case of Theory A, the probability of all these observations (the vast age, the vast size, the vast quantity of lifeless content, the vast lethality of the universe; and the bizarrely long, meandering, particular way life arose and developed into observers asking these questions) is essentially 100%. And you can’t get “more” than 100%. It’s as likely as likely can ever be. These observations are therefore maximally probable on Theory A. By contrast, none of these observations are at all expected on any plausible theory of intelligent design. Indeed, they are on Theory B predicted not to be observed.

Intelligent engineers aiming to create life don’t make the laboratory for it vastly larger and older and more deadly than is required for the project. Indeed, unless those engineers intend to convince that life that they don’t exist, they don’t set up its habitat to look exactly like a habitat no one set up. This is the least likely way they would make a universe. But set that point aside. The conclusion already sufficiently follows from the first point: there is no reason to expect God to have made the universe this way. It cannot be predicted that this is what a God would produce, or that it is what he would want to produce. Whereas it is exactly 100% predicted to be what we’d see if there was no God. So no matter what you try to propose, you can never get that probability to be 100% if there was a God. You can propose all sorts of excuses, all sorts of “maybes,” but you will never be able to prove those proposals to be 100% certain to be true. There will always be some significant probability that those “excuses” simply aren’t true, that God simply doesn’t have your imagined motives or limitations. And indeed, when there is no evidence for or against any one such motive or limitation, its probability simply is 50%. It’s as likely as not.

In other words, God is no more likely to have created a world that way, than in some other (indeed even more obvious) way. To change this assessment, you need to present evidence (not conjectures; but actual evidence) that God is more likely to have chosen these weird ways to do things instead. But even that won’t rescue your case. Because it wouldn’t even matter if you could get these probabilities to, say, 80% or even 90%. That’s still less than 100%. So unless you can prove it is logically certain that God would do all these weird things and not any more obvious things when creating a world for the purpose of supporting life, Fine Tuning is still evidence against the existence of God. Because the evidence of the universe’s construction (what its “Fine Tuning” actually did, such as merely allow the random appearance of life after billions of years, across billions of light-years, within a universe almost entirely hostile to life, and only along a slow and meandering cellular building process) is still “more” probable on chance accident than on intelligent design. And that’s what it means to say something is evidence for or against a theory: if the evidence is more likely on Theory A than on Theory B, then it is evidence for A and against B. And that’s that.

Without any evidence that God is any more likely than not to have made a universe to produce life in exactly the same way as a godless universe, the probability of each decision from God is at best 50/50. Whereas the probability is in every case 100% if there is no God, because the probability that we’d observe ourselves in a Fine Tuned universe of this kind is 100%. It may help to see it from the other way around. It is logically necessarily always the case that P(e|h) = 1 – P(~e|h); the probability of the evidence we observe is always the converse of the probability of having observed the opposite evidence (see Proving History, pp. 255-56). Yet the probability that we’d observe ourselves in a universe not Finely Tuned to produce life is exactly zero. Because, not producing life, it will never produce observers; so if God doesn’t exist, anyone observing themselves to exist will always observe themselves in a Finely Tuned universe. Thus, Fine Tuning is not a “peculiar” thing for us to observe. It is not distinctive of God-made universes; it is, rather, distinctive of godless universes. It is literally the only thing we could ever observe—unless God existed and made the universe. Because only then could the universe possibly have been made conducive to life without the Fine Tuning of our peculiar fundamental constants. Hence God-made worlds will tend to not be Fine Tuned.

This is a crucial realization. Fine Tuning of our observed fundamental constants is only necessary when a God is not doing the designing; it is only necessary when observers only evolve through billions of years of gradual cellular scaffolding, and life at all arises only by chance chemical mixing, and only after billions of years of the meandering random mixing of chemicals across a vast universe billions of light-years in size filled with random lifeless junk, which is almost everywhere lethal to life, and only hospitable to it in tiny specks of the chance arrangement of randomly mixed conditions. Only those conditions require Fine Tuning. Quite simply put: only Godless universes have to be Finely Tuned.

Which means when you observe a universe like ours (old, huge, deadly, and producing life only in the most awkward of ways and rarest of places), you can expect it to have been Finely Tuned by chance accident, not intelligent design. Intelligent design would more likely make a universe as large and old as needed to contain the life it was made for, and would create life directly (not employ billions of years of cellular scaffolding), and imbue the world with only those laws of physics needed to maintain it to its purpose (no weird fundamental constants, no weird fundamental particles). It would not produce a universe almost entirely hostile to life. There would be no lethal radiation-filled vacuum. No dead worlds or lifeless moons. Stars would not be uninhabitable monstrosities. Black holes would never exist.

You can easily do the math here:

So if there is no evidence for or against God choosing this one weird way to make a habitable universe, the odds favor the conclusion that the Fine Tuning we observe is a product of chance accident, not intelligent design, by at least 16 to 1. Hence it is sixteen times more likely we’d see the Fine Tuning we do if there is no God. Fine Tuning is therefore evidence against the existence of God. And it’s rather strong evidence at that. It means the probability that (at least a creator) God exists is less than 6%.

And this is being extremely generous to the God hypothesis. Because we are assuming it’s even half likely that an intelligent engineer would do any of these things, when we well know, from countless observations of how intelligent engineers do their jobs, not a single one of these things is anywhere near that likely. It’s at best thousands to one they’d make a universe this old, this large, and this lethal, and use a random, meandering, billions-years-long process of cellular scaffolding. They’d just make life. They’d give it a completely habitable environment. And they’d waste no time or space on it. Whereas if God does not exist, there is no possible way we could observe ourselves to exist other than in a universe this old, this large, and this lethal, following a random, meandering, billions-years-long process of cellular scaffolding. Hence the Fine Tuning we observe looks exactly like the kind we’d get if no God were responsible for it. The universe looks exactly like a world with no God. That’s weird if God were responsible for it. And weird is just another word for improbable.

There is no way to get around this. Even if you could somehow prove it’s, say, “90% likely” God would produce any one of these strange features in his life project, you still end up with odds against God having done it of around 1.5 to 1. Because it’s still 100% likely we’d see each and every one of these things if chance accident produced the Fine Tuning we observe. And that’s simply more than 90%. Compounded across four variables, it’s even less likely than that that God is the culprit. Between those two competing hypotheses explaining Fine Tuning, chance accident and intelligent design, Fine Tuning as we actually observe it is simply evidence for chance accident. It is therefore evidence against God.

This holds even for Bizarro Gods. You can try to gerrymander your way out of this fatal conclusion by just “insisting” the God that exists is full-Bizarro: he has a bizarre reason to make a vastly large universe, and vastly old universe, a vastly lethal universe, and use a vastly bizarre way of creating observers that extraordinarily looks exactly like the way they’d arise in a completely godless universe. If we “presuppose” each of these strange qualities in God, then we could say all this evidence is again 100% expected. But notice that still doesn’t get you evidence for God. Since 100% is not higher than the probability of seeing these things on atheism, Fine Tuning still is not evidence for God—even a Bizarro God. But it gets worse. Because the logical cost of all those presuppositions is simply moving the probabilities from the likelihoods to the priors. You end up with exactly the same odds against God.

Yes, you can argue that if God had a motive to, say, use billions of years of cellular scaffolding to produce worshipers, then observing that to be the case would be 100% expected. But what evidence do you have that God had that motive? In the absence of any evidence making that likely—or even any evidence making it unlikely (though we actually do: all the evidence of how creators intelligently accomplish their goals)—its expected probability is still at best 50%. Because there’s only a 50/50 chance in each case that your presupposition about God’s motives is true. So all you would’ve done is switch out an odds of 1/1 x 2/1 for 2/1 x 1/1. Or worse. But let’s be generous. It’s either a 50% chance God would on a whim use these weird methods, or a 50% chance he had a motive to (which amounts to saying the same thing). Either way, even at best, it still can’t be any more than 2 to 1 against that God existing on the actual evidence of Fine Tuning. And if you keep doing this down the list of all four oddities, it still ends up 16 to 1 against. You’re right back to less than a 6% chance any such God exists.

Conclusion

The usual response to this disturbing result is to say, well, okay, but “chance accident” still requires, you know, a chance accident. And that’s improbable, right? Maybe. But it doesn’t matter. Because the God hypothesis also requires positing a chance accident—indeed, one even more improbable. So that’s a wash. The only way to tell what is the more likely explanation of the Fine Tuning we observe—a chance accident of physics or a chance accident of theology—is to look at the evidence, to see what kind of Fine Tuning it looks like happened. And all that evidence confirms the kind we see is the kind produced by a chance accident of physics, not a chance accident of theology.

And that’s even if we just posit a brute, one-time, totally random selection of what would exist. In fact, there are dozens of extremely simple starting points that entail an inevitable limitless multiverse, which gives the “chance accident of physics” hypothesis an inherent probability of effectively 100%. When countless universes exist at random, it is simply inevitable that there will be many of them Finely Tuned for life in the way we observe. Per the Infinite Monkey Theorem: given enough tries, all probabilities approach 100%. So when we are comparing as hypotheses any of a dozen simple-initially-conditioned godless multiverse theories, and any God hypothesis, the priors always favor the former. Because they have much simpler starting points, requiring fewer lucky breaks. Indeed, arguably, the simplest possible theory that predicts a limitless multiverse is simply…nothing at all. And God is certainly a hell of a lot more complex a theory than that. So even the priors doom the God hypothesis; and even at best, can’t rescue it.

All the evidence of Fine Tuning therefore proves God does not exist—not the other way around. Of course by “proves” here I am not using the technical term in logic and mathematics to refer to a deductive proof; I am using the everyday English word that simply means showing sufficient evidence to warrant confidence in a conclusion; the concept of proof in science or a court of law. The actual evidence of Fine Tuning—the vast size and age and lethality of the universe, and the long meandering and peculiar process of evolution, all caused by the specific arrangement of physical constants observed—is simply not at all likely on the God hypothesis, but as near to 100% expected as makes all odds on any simple godless hypothesis—especially any standard Multiverse Theory today, any of which follows from far simpler proposed initial conditions than God.

Consequently, Fine Tuning is only what we’d expect to observe if there was no God. It is actually an odd thing to see if God exists, as such a God would have no need of Fine Tuning. Gods don’t need hyper-precise arrangements of bizarre physical constants to create and sustain worlds. That’s why none need be imagined to explain anything depicted in the first chapter of the book of Genesis. Real gods will just make stars glow and radiate heat, for example; they don’t need bizarrely complex nuclear reactions to do it. They don’t need electromagnetic forces. They don’t even need gravity. They can just make things fall or not fall wherever and however they want. Nor do they need universes that are almost entirely lethal to life and that span billions of light-years and spin their wheels for billions of years before producing any life at all. Whereas without God people could only ever observe themselves to be in a Fine Tuned universe; and will almost certainly always observe themselves to be in one Tuned to be vastly old, large, and lethal to life. Whereas even if on a whim a God used Fine Tuning to make a universe anyway (for whatever weird reason), they still would not likely have Tuned that universe to be so vast in size and age and lethality to life—as opposed to making their universe small and young and widely habitable, as our theists originally, and correctly, anticipated.

This doesn’t mean a God couldn’t have done that. Remember, we are here talking about what’s likely, not what’s merely possible. Regardless of what’s possible, it is still the case that the evidence of Fine Tuning reduces the probability that God exists. It does not increase it—because the observation of that evidence is 100% expected if God doesn’t exist, but significantly less than 100% expected on any typical God hypothesis. And that has the inescapable logical consequence of reducing the probability that God exists. This does mean that had we observed ourselves in a very young, small, widely habitable universe, all our species instantly formed, and the world obeying the physics of direct command without any underlying complex mechanisms like quarks or electromagnetism, then we could be conversely certain we were here because of intelligent design. But that just isn’t the way the evidence went. It went exactly the other way.

Fine Tuning is therefore evidence against God’s existence; and evidence indeed sufficient on its own to be confident no God exists. Because on the actual evidence of Fine Tuning, even the highest probability that any God could have isn’t even 6%; and for the reasons we just surveyed, it is almost certainly a great deal more improbable than that. And this is true even for Bizarro Gods, who, sure, could make the evidence 100% likely again, but only at the cost of lowering the prior probability that any such God exists in the first place by exactly the same amount. So you end up with the same improbability that any such God exists. We can therefore be confident none does.

Discover more from Richard Carrier Blogs

Subscribe now to keep reading and get access to the full archive.

Continue reading