Is the existence of God logically impossible? I used to be suspicious of arguments that attempted to prove that, because they were usually so lame, and easily rebutted (although some stick, depending on which “God” you are talking about: see my discussion of this in Sense and Goodness without God IV.2.4, pp. 275-77; and for some serious, but not always successful, attempts at building these kinds of arguments, see the anthology The Impossibility of God; some other examples are cataloged at the Secular Web; but a very interesting example, quite pertinent to today’s post, is Evan Fales’ Divine Intervention: Metaphysical and Epistemological Puzzles). Yesterday I blogged an ontological argument for the necessary existence of our universe without God (Ex Nihilo Onus Merdae Fit, which I have since updated as The Problem with Nothing), and I had to stay on point there (it was long enough as is), but in developing that argument over the years I had already been thinking about one implication of it: if an infinite selection of all logically possible universes exists, then many of them will contain gods, if gods are logically possible. Today I cover that angle. I will also develop this argument further in The Argument from Specified Complexity against Supernaturalism, to which important background is the converse argument that Naturalism Is Not an Axiom of the Sciences but a Conclusion of Them.

Which God Was That Again?

To begin with, we can rule out the God of any monotheism, for the simple reason that if P1 is true (“in the beginning, there was absolutely nothing,” the key premise of the argument I developed in the previous post), then there are infinitely many more configurations with multiple gods than only one god. In fact, if we’re to ask about a true monotheistic God, such that no other gods exist at all (not even in other universes), then the probability that such a God will come into existence on P1 is infinitely close to zero. Because it’s infinitely improbable that of all the possible configurations, we’d get only one God out of all the universes whatever. Indeed, universes with many gods in them will vastly outnumber universes with only one god in them, even before we get to the possibility of no universes containing any gods except one of them. So you can’t rebuild monotheism on P1. At best you get polytheism. In universes we’re not in (because if we were in one, we’d have verified some gods in it by now).

And that’s even if God is logically possible to begin with. And lately I have suspected he is not. I have approached this question from two different directions in my random contemplation over the decades, and I see now they are approaching a common result, so here in one post I will discuss them both and how they reinforce the same conclusion. The first relates to yesterday’s post, and the general question of what sort of entities can logically exist (which connects to my published work on defining naturalism and the supernatural) and how likely they’d be if P1 there is true and we followed it here to the same conclusion reached there (that infinitely many universes exist, being a random selection of all possible universes). The second relates more specifically to our ability to conceive of disembodied minds (and disembodied mental powers even apart from minds) and whether that affords any evidence that such things must be logically possible (I conclude it does not, and that it may even prove the contrary).

From Boltzmann Brains to Boltzmann Gods

In my work on Defining the Supernatural I explored the difference between “natural entities” and “supernatural entities,” and demonstrated that the latter are ontologically basic mental entities, such that if no such things exist, then nothing supernatural exists. And if that’s the case, then all mental entities are not ontologically basic, but are instead reducible to interacting nonmental parts (like neurons, microcircuits, or what have you). And that being the case is what defines naturalism as a worldview. In my analysis I discussed the difference between supernatural gods and the kind of gods that could still exist if naturalism is true (skip to the section there called The Stoic and Epicurean Gods to see what those would be like). The latter would basically be animals, aliens, or computers, or mishmashes of all three, and most likely in a truly bizarre scifi way. There are infinitely many different kinds of gods like this that are logically possible even on presently known physics.

And that has to be acknowledged. Probability combined with the law of large numbers combined with the realities of cosmological scales of space and time entails some very weird things. Which are nevertheless certainly true. I’m not speaking of Nick Bostrom’s bizarre argument that we must be living in a simulated universe (Are you Living in a Simulation?), which doesn’t really work, because it requires accepting the extremely implausible premise that most civilizations will behave in the most horrifically immoral way imaginable, and for no practical reason whatever (in all good sense, by far almost all sims that anyone will ever generate will be games and paradises, not countless trillions of aimlessly tedious worlds with thousands of years of pointless wars, holocausts, plagues, and famines). Rather, I’m speaking of Boltzmann Brains. (For more on all of these points, see We Are Probably Not in a Simulation.)

If the universe were to slowly expand forever, even if it were to fade into a heat death of total equilibrium, even then, simply due to the laws of probability, the random bouncing around of matter and energy would inevitably assemble a working brain. Just by chance. It’s only a matter of time. Maybe once every trillion trillion years in any expanse of a trillion trillion light years. But inevitably. And in fact, it would happen again and again, forever. So when all is said and done, there will be infinitely many more Boltzmann brains created in this universe than evolved brains like ours. The downside, of course, is that by far nearly all these brains will immediately die in the icy vacuum of space (don’t worry, by far most of these won’t survive long enough to experience even one moment of consciousness). And they would almost never have any company.

Which is how we know we aren’t Boltzmann brains. Because we aren’t just floating around alone in random space dust. Yes, there will also inevitably be a completely random assembly of a whole working earth and civilization and so on, but that will be vastly (and I mean vaaaaasssstlyy) rarer, and again even then we would see we were on a weird isolated earth floating around in a frozen dead universe. And yes, there will inevitably be a completely random assembly of a whole working universe out to a visible horizon fourteen billion light years away that just by accident happens to look like it’s undergoing an accelerating expansion, and look like it began by a Big Bang but didn’t, and people in that world will be fooled. But of all the worlds that look like that, almost none of them are like that. Rather, most worlds that look like that got there the hard way. And when I say most, if I were to attempt to show you the ratio of real to accidental worlds that look like that, you would be unable to conceive of the number I came up with. So the odds are as good as a hundred percent that we’re in one of those real worlds, and not one of the weirdo accidental ones that look exactly the same. Although, if P1 is false, then our world probably is the product of a Boltzmann Big Bang (see my comment in The End of Christianity, n. 31, p. 411; see also A Hidden Fallacy in the Fine Tuning Argument). But it’s infinitely unlikely to be one of those accidentally deceptively assembled worlds.

Nevertheless, given infinite time, such worlds will exist. It’s a logically necessary truth. In fact, anything that has (and maintains) a stable nonzero probability of happening, will happen. Eventually. We can’t always be sure, though, what actually will maintain a stable nonzero probability of happening, and many things simply will not. Hitler will never be alive again. That probability is now zero. Because he’s dead. A copy of him might pop into existence some day, or some nearly identical sequence of events might produce someone nearly identical to him someday, but that would still be a different guy. And if universes don’t undergo eternal heat deaths, but collapse or explode on a regular basis (as ours is set to do), then there might never be a span of time enough to make a Boltzmann universe (although a Boltzmann brain, maybe). But even then, barring logical contradictions, even a sequence of short-lived universes might eventually make Boltzmann worlds. Ascertaining whether that will actually be impossible (and if possible, then it has a nonzero probability, and therefore will happen eventually) is a task perhaps beyond human ken.

But it doesn’t matter, because the point is, Boltzmann brains are an inevitability. In fact, because time will never end, there will be infinitely many of them. Boltzmann worlds, too. Which means Boltzmann gods are likewise inevitable…in fact, there will be (if there haven’t already been) infinitely many of them. What is a Boltzmann god? Think of a mind that is as near to perfection and power as could ever be physically made, a supermind, with a superbody, maybe even a body spanning and permeating a whole vast region of spacetime. The improbability of this is staggering. But remember, everything with a nonzero probability is going to happen, eventually. In fact, it’s going to happen infinitely many times. Only its relative frequency will be staggeringly low. Worlds without such lucky accidental gods will vastly (and I mean vaaaaasssstlyy) outnumber worlds with them. But the worlds lucky enough to get them will experience some pretty cool, or some pretty horrific, fates. In some, this god will be randomly evil and create civilizations just to torment them for fun (and let me reiterate: this may already have happened; in fact it may already be happening right now, in universes or regions of spacetime vastly beyond ours). In others, this god will be randomly awesome and create a paradise for his gentle children.

This will happen. It probably already has happened. It probably is happening as I type this. It’s a logically necessary truth. That’s weird. But there’s no escaping it. The only way this could ever be prevented is if time began, and were to end. Not the universe. Time itself. And not just time in our world, but in all worlds, all the spacetime continuums that exist beyond this one (if such there are). And there is no reason to believe that. Not only is there certainly no reason to expect time to end (there is no known physics on which it would–even the collapse of the universe will only create a ball of pressure so great that it will explode again into a new universe, with time still ticking; or seethe forever in a superdense state, time still ticking), but there is no reason even to expect that time ever began (we must assume it did if we grant P1, but if P1 is false…). Only Hawking’s nutshell model has time loop back in on itself at the Big Bang, and perhaps in an extremely unlikely scenario our universe may be one giant time loop somehow. But that’s just it: an extremely unlikely scenario.

Natural Gods or Supernatural Gods?

So Boltzmann gods are almost certainly an inevitability. Just immensely rarer than Boltzmann brains. Maybe even rarer than Boltzmann worlds, although many must surely be easier to randomly construct. They would be like the many different kinds of naturalistic gods I started talking about. Infinitely many configurations even under known physics would produce all kinds of gods of different sizes, powers, characters, of all degrees of intelligence and knowledge. These would in effect be alien gods, gods with bodies (however ephemeral or bizarre those bodies might be), gods with limitations. But they would be capable of anything gods of yore were, from immortality and superpowers, to intelligently creating universes and working scientific wonders (miracles, for all intents and purposes: see Clarke’s Third Law).

That’s what would distinguish them from just any Boltzmann brain, the ability to do those things; that’s what would classify them as accidental gods, and not just accidental people. And they need not even be accidental: we ourselves might one day create gods like this; we may even one day become gods like this (see How Not to Live in Zardoz). Barring an extremely unlikely disaster, a million years from now we will have the technology to accomplish either. And we are unlikely to be the only ones in this universe able to do this. In fact, in all probability, someone has done it already (statistically, we must have gotten started billions of years later than many civilizations in the cosmos). It’s just that, odds are, they are probably a billion galaxies away. And their gods, being physical beings with all the limitations that entails, won’t be able even to know we exist, much less communicate with us or lend us a hand.

Which is why many people don’t really allow these sorts of beings to be called “gods.” That is, this is not what people mean when they ask whether God exists, or insist that He does. Not even polytheists mean their gods to be distant aliens, accidental or manufactured. So the real question is not whether “gods” exist somewhere, in this or any of the universes that exist if P1 is true, but whether supernatural gods exist. Gods that don’t have the limitations and flaws of physical creatures. Gods that can be everywhere in the universe at once. Gods that aren’t slowed by the speed of light or weakened by the laws of thermodynamics. Invisible Gods that created our universe and hear our thoughts and meddle in our affairs, for good or ill. Gods that have constructed awesome heavenly places for us to go live in after we die (or horrific eternal prisons, as the case may be). Those kinds of Gods.

Certainly, if the question is, “If P1 is true, doesn’t that entail that there will be countless universes with all kinds of naturalistic gods in them, some accidental, some manufactured?” then the answer is “Yes.” Not only is that inevitably the case across any infinite array of purely naturalistic universes, but it’s inevitably the case in our own universe, where eventually there will indeed be Boltzmann brains, and far more rarely, but just as inevitably, Boltzmann gods; and sooner than either, gods of our own or alien manufacture. But what about supernatural gods? What about God?

The Probability of Supernatural Gods

As for monotheism, as I already pointed out, even if a supernatural One True God is logically possible (and as I’ll get to in a moment, I suspect it is not), then if P1 is true, the probability that this God exists is still infinitely close to zero: because infinitely many gods are possible, but God is here being defined as the one and only, and of all possible combinations of gods that could exist (in this universe alone, much less across all the infinite universes there would be), that only one would be selected to exist bears odds of many infinities to one against. It’s pretty much the most improbable thing humanly conceivable. In fact, it must necessarily have a probability of zero now, for the simple reason that once other gods exist, it becomes logically impossible for there to have been one and only one God. So if all the infinite multiverses born of the original nothingness did not at that singular moment produce one and only one God over all of them (and him the most marvelous and perfect of all the singular gods there could have been), then that ship has sailed.

Since we observe there to be no such God in our universe, we know that no such God came to be. Therefore, none ever can come to be. But if you let go of your dogmatic and emotional need to “believe” in that extremely improbable God, and instead just clinically examine what possible gods are left, there could one day perhaps be a really supremely awesome “supernatural” Boltzmann God (or may even already be countless many of them, scattered across other universes we’re not in). If supernatural gods are logically possible. So that’s the question. But before we answer that, let’s explore the logical consequences of assuming that such gods are logically possible.

As I’ve noted, naturalistic Boltzmann gods will exist, but will be so extremely complex and improbable we will almost certainly never meet one (the only naturalistic gods we are likely to encounter are those we build ourselves). And those that exist across the multiverse, created spontaneously by the instantaneous transformation of the original nothingness, will be extraordinarily rare (and that’s an understatement). But what about supernatural gods? Obviously, by definition, you can’t randomly assemble those out of nonmental parts. How would one come to exist by accident then? They could only come to exist by the random assembly of irreducibly mental properties, supernatural “parts” as it were, and since we have no known physics of that, we can’t really run calculations for it (in the way we can, for example, in statistical mechanics).

However, we can approach something like a conceptual analog. Note that even a supernatural God is vastly complex in its constituent parts. Any mind must necessarily be, much more so a mind with powers beyond those of mere thought (I explore this point further in The Argument from Specified Complexity against Supernaturalism). Theists will insist that God is somehow simple in the sense of having no parts, or all his parts logically entailing each other and therefore inseparable, and so on and so forth, but that’s all just handwaving. Sound proofs are always wanting. There is no conceptual basis for thinking that any mind “must necessarily” be omnipotent or omniscient in any sense, or for thinking that any supernatural spirit “must necessarily” be omnipotent or omniscient in any sense, or that any creative intelligence must be, either. Certainly lesser deities, lesser spirits, lesser supernatural minds are logically possible. Therefore a God could be a lesser being. If we are to randomly produce a God from among all logically possible Gods (as P1 would entail we do, if supernatural gods are logically possible), then we will certainly not get an omni God of any sort. We will get a God of some lesser knowledge, intelligence, virtue, and power than the best we could logically get (Ibid.).

Thus most Gods will be lesser deities. Few to none will be anywhere near omni. But how many will there be, in terms of per-universe frequency, say? First of all, most universes with those more common lesser gods in them, will have many, not just one. Because he number of logically possible universes with many lesser gods is infinitely greater than the number with single greater gods in them. Imagine if you rolled a die and got a god on a “6.” Now imagine you get to role ten dice. What is more likely? That you will end up with just one “6” or more than one? Most universes will end up so large and old that, no matter how rare Boltzmann gods are, eventually you will more likely get several than just one. And yet, at the same time, the number of universes with even just one Boltzmann god will be the rarest of rare. You will get Boltzmann people, worlds, and even Big Bangs more commonly than that (see my discussion of the First Premise of the Kalam Cosmological Argument).

So already, universes with just one god in them will not only in all probability have a lesser god, but such universes will be extraordinarily rare (and that’s a ridiculous understatement). Polytheistic worlds will be vastly more common. But secondly, a randomly produced supernatural god will not be much more probable than a natural Boltzmann god. The only thing in their favor is that, unlike natural Boltzmann gods, they won’t need a lot of superstructure to operate (e.g. no digestive system or equivalent, for example), and thus require a lot fewer parts. However, even natural Boltzmann brains do not require much superstructure relative to the complexity required of their brain to begin with. That is, almost all their improbability derives from the vast complexity of the brain itself (such as is required to generate a mind), in comparison with which the complexity of their bodies is trivial or incidental.

And that complexity will have to be shared by a supernatural mind, in two respects: (1) of all the possible combinations of mental contents and properties and their infinite degrees, godlike assemblies will be extremely few, relative to all the combinations that fall short of godlike; and (2) even mundane minds require vast complexity of organization, to produce a reliable system of beliefs and memories and of processing perceptions and contemplations. For example, to keep distinct the 500 or so faces our own brains are capable of memorizing, and correctly connect those faces to large systems of correct information about each face, and not get these connections all crossed up and confused, requires an extremely complex arrangement of neurons and synapses, any rearrangement of which would create confusion and error and literally eliminate information. A supernatural mind must also keep all this information inside it and also keep all these connections correctly linked up, which also requires a structure no less complex.

That structure might not be made of “stuff,” it might somehow be made of dreams or rainbows or bare supernatural brute facts or whatever, but the structure must still exist. Because of all the ways to connect up a supernatural mind, vastly more of them will be connected up all wrong, than will be connected up all right—much less connected up right for a superhuman scale of information and information processing. When picking random supernatural minds, most of them (by far) are going to be babbling lunatics or even completely nonfunctional spirits. After getting past those, of what remains, most (by far) will not be godlike. And after getting past those, of what remains, most (by far) will be truly minor gods. And after getting past those, of what remains, most (by far) will be merely mediocre gods. And so on. The number of gods who will be anything close to what Christians would want to worship, for example, is going to be infinitely fewer by comparison. In other words, the probability of any universe getting such a God in it is going to be well near infinity to one against. Even if the supernatural is logically possible.

So Is the Supernatural Logically Possible?

Still, if P1 is true, then it would still be the case that, in a broad sense, naturalism must be false, because supernatural things will inevitably exist, in some universes somewhere. If naturalism is true at all, it would only be true of our universe. But therein lies our first clue that the supernatural might in fact be logically impossible: the fact that we don’t observe anything supernatural operating in this universe. If it were in any way common for supernatural things to exist, certainly if they were as common as nonsupernatural things (and given P2, which is entailed by P1, why wouldn’t they be?), then our universe should be full of supernatural things, or at least have enough of them for many to have been scientifically confirmed by now. It would be extremely unlikely that we “just by chance” ended up in a completely supernatural-free universe (and no anthropic principle entails we would, either). Which in turn would entail that a supernatural god is impossible. Because if the supernatural is impossible, so are supernatural gods.

This is not a proof, however. There are extremely improbable ways that the supernatural could still exist and we would just happen never to have seen any. So at best this is evidence for the logical impossibility of the supernatural. Unless, of course, P1 is false. Then perhaps the supernatural is logically possible but just happens never to have become actual, owing to something (?) that prevents it. Although even then it would have to be possible to create supernatural things in this or some other universe. Because what it is to be a possible thing is to be a potential thing. That a triangle made of freshly severed dinosaur heads is “logically possible” means that any region of spacetime can (in principle) be configured to produce it. In fact, that very reconfiguration is what it means to be a triangle made of freshly severed dinosaur heads: it’s what the sentence “there is a triangle made of freshly severed dinosaur heads” means, such that if we didn’t know (at least in outline) what configuration of spacetime would make that statement true, then we literally wouldn’t know what that statement meant. (See Sense and Goodness without God II.2-III, pp. 27-208.)

So how exactly would we reconfigure spacetime to produce a supernatural property? That question is meaningless, because the supernatural by definition is not reducible to configurations of spacetime (it’s irreducible mental stuff, not spacetime stuff). So in what sense is the supernatural ever a “potential” property of anything in spacetime, much less of spacetime itself? I confess I cannot conceive of how it ever could be (whereas I can conceive of how everything natural is: see, for example, Superstring Theory as Metaphysical Atheism and The Argument to the Ontological Whatsit). But that again is not a proof, because many things we cannot conceive of are nevertheless true. Our inability to conceive of something only demonstrates our ignorance–which ignorance can be produced either by something not being possible or by our simply not knowing what makes it possible.

Maybe the supernatural is prohibited from existing by the laws of physics, which laws, if we could change them, would allow the supernatural again. I don’t know (I doubt it). But the point is, if P1 is false, then so is P2. So then the supernatural no longer has the same probability as the natural, and might even have a probability of zero, if something just happens to always have existed that prevents it from existing, something that does not necessarily exist, but just always has for no reason (the way God is supposed to always have existed for no reason). Although I confess that P1 is so surprisingly successful and simple an explanation for all that we observe (as I explained in The Problem with Nothing), I almost think it would be amazing if it wasn’t true. And anyway, I still have to ask whether the supernatural is logically possible. Since if P1 is true, then the truth of naturalism, in the grand scheme, requires the supernatural not merely to be nonexistent, but to be logically impossible. Because otherwise, the supernatural necessarily exists. Somewhere. Even if it’s not around here.

But again, that there isn’t any around here is clue number one that the supernatural is logically impossible after all.

On the Conceivability of Disembodied Minds

Although it’s obvious that an inability to conceive of something in no way proves it is impossible, I used to think that if something was conceivable, it must be logically possible (this is a working assumption evident, for example, in my 2004 critique of Reppert’s Argument from Reason). I now know there is one major flaw in that assumption, discussing which is how I shall end today’s meditation on the possibility of God.

First, it must be noted that many things (in fact, countlessly many things) are logically impossible that we do not know and at present cannot know are such (likewise things which are logically necessary). This is often overlooked, as it is assumed that if something is logically impossible, that fact should be obvious. But consider as an example Fermat’s Last Theorem, which simply states “no three positive integers a, b, and c can satisfy the equation a^n + b^n = c^n for any integer value of n greater than two.” Now, this statement was either necessarily false, or necessarily true. Thus, we either had a logically necessary truth, or a logical impossibility. Yet no one could prove which it was, for almost four hundred years. Only very recently did anyone prove that it was, after all, true (and therefore logically necessary). In fact, since it’s a statement declaring something to be logically impossible, by proving it’s true we had discovered something that is logically impossible. Many attempts had been made to prove this statement, which convinced people for a short time but were found upon further examination to be invalid or unsound (thus we can fool ourselves into believing something is logically possible or impossible, even when it isn’t). Finally, the proof that finally proved valid and sound, was over a hundred pages long. There was clearly nothing obvious about this statement of logical impossibility. And proving it required an extraordinarily arduous series of thousands of statements.

I provide this example to make a crucial point: if God is logically impossible, it could well be that the proof of this fact will require a hundred pages of propositions, and four hundred more years to discover. If anyone is even looking for it–unlike the quest to prove Fermat’s last theorem, to my knowledge no one is (they are only looking for simple proofs, of which there may be none). So it may never be discovered. Thus the fact that no such proof has been discovered is not a valid argument against the logical impossibility of God, any more than it would have been an argument against the logical impossibility of the equation in Fermat’s last theorem, or an argument against the logical impossibility of countless other logically impossible things we have yet to discover or may never do. Thus, suspecting the logical impossibility of God does not require a formal proof of its impossibility. We can have clues enough, as there were in the case of Fermat’s last theorem.

Second, it must be noted that concepts do not entail realities. Concepts, in the sense of potential entities (and not in the different sense of entertained or encoded thoughts), can exist necessarily, and exist always and everywhere, but concepts can’t think and act. Many ontological arguments for the existence of God are actually disguised arguments for the existence of the concept of God. But a concept of a God is not itself a God. That people can conceive of an entity is simply not the same thing as that entity existing. Fiction affords too many examples for me to have to belabor the point.

But here’s the rub. Even in terms of probabilities on cosmic scales, not every fiction will materialize. The movie Star Wars will never be acted out for real, not even in some vastly distant Boltzmann world, because it actually incorporates logical impossibilities (such as sound in a vacuum), systemic impossibilities (any civilization with such technology would not use humans to aim ship-to-ship weaponry, much less engage spaceships at such close proximities that even American naval commanders would consider absurd, which facts entail actual logical impossibilities between the human intelligence displayed in the film, and such stupid behaviors), and, of course, physical impossibilities (such as The Force, which as represented in the film no configuration of our universe, outside of computer simulated universes—take note—could ever produce, not at any probability).

This example is again crucial: we do not notice anything logically impossible about Star Wars; indeed, we are watching it, so how can it be logically impossible? But of course we are not actually watching Star Wars happen in reality. We are watching a dramatization that covertly persuades our brains to imagine that what is happening is happening, when in fact none of it actually is. Except, of course, in the “fourth wall” sense: filming the movie Star Wars is not logically impossible, but that’s all just trickery, the statements the characters make are literally false, the actors are not the characters they portray but are only pretending to be, the space battles were not filmed in outer space, and so on. When we actually try to translate all this into a real world system, only then will we notice the logical impossibilities that prevent any such drama from ever actually occurring, not at any probability, no matter how vanishingly small.

This is like Fermat’s Last Theorem: we can “imagine” that 2^45 + 3^45 = 4^45 is a true proposition. Especially if we don’t know how to run the math or lack a calculator to test it with. We can still understand every symbol, and the meaning of the statement as a whole. And there is no obvious contradiction among these parts. Indeed, we could even run the math with a calculator, make an error we didn’t catch, and thus conclude that that statement is indeed true after all! It happens to be false. In fact, not just false, but logically impossible. Yet that does not prevent our brains from imagining, even believing it is true. Thus, we can conceive of something as being true, that in fact is impossible. Therefore, our being able to conceive of something does not mean it is logically possible. And the more complex the thing we are asked to imagine, the easier it will be for us to overlook any logical impossibilities in its arrangement. We instead busy ourselves with imagining the parts and their juxtaposition. But that’s not the same thing.

The reason we are susceptible to this error is that when we imagine and conceive, we build models using pieces of things we know exist. We know light exists. We know swords exist. So why not a sword made out of light? We do not trouble ourselves with working out, first, how it could possibly be that a lightsaber has a practical finite length, and can be stopped (and with a loud report) by another beam of light just like the colliding of swords. Even lasers that could cut us in half cannot “ricochet” off of other laser beams like swords. They would pass right through each other (this is due to a property of light that is fundamental to the entirety of optical physics). But we understand how lasers work. We understand how swords work. So we build a model of laser swords in our head, borrowing the bits we want from each. But that in no way entails we could ever actually make a lightsaber. In this case such a thing at least is “logically” possible (the technology would be needlessly elaborate, whereas just cutting the guy in half with an actual laser would be easier; there’s a reason police and soldiers don’t fight much with swords anymore). But the point is, its physical impossibility does not for a moment deter us from imagining it, either. In fact, we can imagine it without even working out whether it is possible or not to build one. Because we are using models from other things and superficially combining them, not troubling ourselves with “checking the math” that would be necessary to connect them. Just like imagining, or even convincing ourselves, that 2^45 + 3^45 = 4^45.

In a computer simulated universe, of course, we could have working lightsabers, and sounds in space, and all manner of absurdly constructed worlds, but only because the elaborate physical machinery underlying the simulation connects all the logical and physical dots to make that happen. In other words, the logical possibility of all these things is dependent on being run as a simulation in a computer, whether a game console, a futuristic Matrix, or (as we usually settle for) the human brain. Take away that substructure, and many of these things are no longer physically or in some cases even logically possible. Thus being able to simulate something does not mean its existence is possible outside the simulation. Being able to simulate it only proves that it is logically possible for it to exist in a sim; in other words, it is possible as a sim. But outside a sim, it could well be logically impossible. Just as Star Wars certainly is.

And this is where we get to the problem of disembodied minds. When we simulate things in the computer of our minds, we can indeed simulate lightsabers without any of the elaborate technology that would actually be needed to make a real lightsaber, we can indeed simulate sounds in space, and impossible conjunctions of intelligence and stupidity (a common feature of fiction well mocked in Galaxy Quest: “What is this thing? I mean there’s no useful purpose for there to be a bunch of choppy, crushy things in the middle of a hallway!” “Relax, Gwen.” “No! I mean we shouldn’t have to do this! It makes no logical sense! Why is this here?!”). But we can only do that because we have a complex machine to do it with: our brains. So instead of the elaborate technology needed to make an actual lightsaber, we are using the elaborate technology of a brain to do all the same work. The logical and physical possibility of a lightsaber in our imagination is therefore dependent on a physical brain of staggering complexity.

For this reason, the mere fact that we can imagine something, that we can conceive of it, does not mean it is logically possible. Because it might only be logically possible inside a computer system simulating it (in this case, that being our brain). We can’t “conceive” of something apart from the machinery of our brain, so we can’t ever actually test the logical possibility of something outside our brain. We can only ever test the logical possibility of simulating something in a brain. And that’s a significant limitation we cannot overlook. We think we can imagine a god not dependent on a material brain, but in fact we can’t. We can only imagine a god dependent on a material brain: ours. The fact that “god sims” can only be run on a physical brain actually argues that god cannot exist outside of a physical framework to give rise to him. (Of course, I am assuming science has well nigh proven that minds do not exist but for complex physical brains: see Sense and Goodness without God III.6, pp. 135-60; The End of Christianity, pp. 298-302, 305-32; and my Argument from Mind-Brain Dysteleology.) So is God really conceivable outside of a complex simulation machine? Probably not.

Think about it: we know that a lightsaber in our brain is dependent on complex neural machinery to maintain its juxtaposition of properties, but that outside our brain that juxtaposition would require a vastly more complicated machinery (which in fact we know nothing about; and need not know anything about to simulate the device in our brains). So what reason do we have to believe God is any different? If you think about it, a lightsaber without any underlying machinery starts to look fairly inconceivable. And it may well be logically impossible. Because if it wasn’t, surely we would have seen countless examples by now of “property conjunctions” without the underlying machinery. Yet in fact, after trillions of dollars and billions of man hours of hard core science across four centuries, we have never discovered even one case. Instead, we have found, in billions upon billions of cases across every discipline and area of human experience, that nothing exists without that underlying machinery (see Naturalism Is Not an Axiom of the Sciences but a Conclusion of Them). Like Fermat’s impossible quadratic, this counts in the evidence column for God being just as impossible. All supernatural things, in fact. So clue number one is pretty strong.

Clue number two is the fact that mental things are actually structurally complex by definition. I made this point about minds earlier. What maintains the “structure” of God’s mind? Rainbows? Bare supernatural brute facts? What are those exactly? Because I actually can’t conceive of anything that would work. Again, that’s not a proof, but in the right context it can be evidence against. It seems obvious that structure requires actually existing things to hold that structure. If nothing exists holding it together, then by definition nothing is holding it together. It’s circular to argue that what’s holding it together is the disembodied mental concepts themselves. Concepts can’t think or act; so disembodied mental concepts can’t “do” things either. They can’t have physical relationships to each other. They can’t have structure. They can have the concept of structure, but again concepts can’t do things; and they can have structure when we simulate them in our brains, but that gets us right back to the point: that appears to be the only way they can ever actually exist.

This is popularly known as the Argument from Physical Minds. I’ve defended an empirical version of it before, but here I am making a logical argument from physical minds: my point is that a nonphysical mind appears to be logically impossible, and not merely non-existent. I cannot prove this. But as for Fermat’s Last Theorem, that proof may be too complex, and may never be discovered. That doesn’t mean I’m wrong. The consequence of this is that if the supernatural is logically impossible, then naturalism is necessarily true. Belief in the possibility of the supernatural is then merely a cognitive error (which psychological science has more or less confirmed, as Victor Stenger explains in The End of Christianity, p. 312, with refs. in n. 27, p. 416), akin to believing that Star Wars could really happen, or that 2^45 + 3^45 = 4^45.

Is that the case? I suspect it is. For more on why, see my Argument from Specified Complexity against Supernaturalism.