God’s Multiverse?

Mere Creation

Beatrice: We spoke before about the problem of evil. I’d like to talk about a new defence:

First, I’d want to say that, even if there is no justification of the evils we see in the world, that they are nonetheless outweighed by the goods. The world is (pace the anti-natalists or negative utilitarians) net positive, 1 and the objection raised by the problem of evil is not that God did a bad thing in making the world, but rather fell far short of moral perfection.

Adam: And the second consideration?

Beatrice: The second is that it is good to bring positive things into existence. This can be weaker than a Total-view-esque it is as good to bring positive things into existence than it is to improve existing things by the same amount: just something like, “Given the option, it is better (ceritus paribus) to bring something good into existence”. 2

Adam: Okay. Where are you going?

Not the best of all possible worlds

Beatrice: God is capable of creating more than one world. On the face of it, perhaps he should: if it is a good thing actualize goods, then two good universes are better than one. So the object of God’s morally perfect act of creation would not be a single universe, but rather a multiverse. As our universe is on balance positive, then that is a pro tanto reason for God to actualize it, even if it has gratuitous evil: perhaps God has already actualized all the universes which do not have unjustified evil, but decides to add universes that do have some unjustified evil. Although they are not as good the previous universes, they are still valuable, and so actualizing them makes his creation even better.

If that is so, then it is plausible that among the multiverse God built, many, or even most, would have unjustified evil (but all would be on balance positive). So when we find ourselves in a universe that is on balance positive, but has apparently unjustified evils, it should not surprise the belief that God created it.

To infinity, and beyond!

Adam: This is really cool. Yet it seems to fall prey to a super-charged variant of the ‘no best world’ objection. In the ‘single universe’ case, there was a problem if a world could be made ‘even better’ without bound, as no matter how good a world God created, it could have been even better – so, perhaps, the notion of a morally perfect creative God is somehow inconsistent.

Yet at least there was the possibility (albeit one lampooned since the Candide) that there really was a best of all possible worlds. In the case of a multiverse, that seems hard. Even granting the identity of indescribables God can play with names and power sets of abstract objects without end. Even if he created an infinite set of good worlds, one could always task him to create the power set of whatever set that was, or perhaps some other way of creating even more universes that set theory is unable to describe.

Beatrice: Sure, but that might cut both ways. One could use this to defend against the problem of evil: Moral perfection probably shouldn’t oblige to do the impossible, and if it is impossible to create the best of all possible worlds, then there can be no complaint against God’s moral perfection for creating a universe (like our own) that could be better, as this would be true no matter the world God created. Perhaps so too here: if God creates a vast number of good universes, we should cut him some slack if he doesn’t create an even vaster number, for similar reasons. 3

Adam: There’s another worry. It is not clear whether adding more good universes adds value beyond a certain point. If you already have an infinitude (even a ‘merely’ countable infinitude) of good universes, the collection has an infinite value. Adding or subtracting finite increments (e.g., by adding further morally good worlds) does not change the total value. So if God already has produced infinite amount of value from a multiverse composed of universes without unjustified evil, why would he include a world like ours into the mix?

Beatrice: I don’t have a neat answer to these infinitarian worries. I would offer a promissory note that these problems could be resolved in a way that preserves God’s incentive to ‘add in’ good worlds to the already infinite sum. In fairness, these problems afflict many. Whatever you think about God, if you hold there’s some non-zero probability of the universe having infinite value, then whatever acts you do, for good or ill, will not affect the total (expected) value of the universe.

Adam: Fair enough, but the proposed solutions to ‘infinitarian ethics’ may still leave problems for this account. As most of them, whether they deploy hyper-reals, ordinals, ultra-filters, or anything else, want (at least) to say that an infinite set of 1 unit of value is less good than an infinite set of 2 units of value, and more generally that (infinite) collection A is better than (infinite) collection B if each and every member of A is better than every member of B and perhaps that each and every member of A is better than or equal to every member of B.

Yet if there is a ‘long way up’ in terms of value, then God could choose to ‘edit out’ worlds like ours and replace them with far better worlds. Instead of a set of worlds with a range of positive values, with our world as a non-maximal member of the set, God could Xerox the maximal member of the set and replace the other members with copies of it. 4 This Xeroxed set of worlds would dominate sets including worlds like ours via standard approaches to infinite ethics (if there can be said to be ‘standard approaches’ in such a recondite field). 5 So why didn’t he?

Beatrice: I think it depends. I’m not sure there’s yet a good approach to infinite ethics that deals with concatenating infinite sets. I agree that A>B when:

A: 2 2 2 ... 2
B: 1 1 1 ... 1

Yet it is not clear what should happen when there’s an option of adding an ‘extra’ set to B:

A: 2 2 2 ... 2
B: 1 1 1 ... 1; 2 2 2 ... 2

Of course, because of how transfinite sets work, the union of an ‘extra’ countable set into B does not change the cardinality of B, and A and B can still be put into one-to-one correspondence, and every member of A is greater than or equal to every member of B. Yet I still have the intuition that it is better to actualize B than A. I guess it gets a bit chicken-and-egg: it is good for God to add ‘extra’ good universes (albeit non-maximally positive) to a set of maximally good universes; it is also good to edit the non-maximally good worlds in the set to make them maximally good. And God can repeat both steps ad infinitum.

Another suggestion would be something along the lines of a ‘guardian angel’ account: in the same way we can imagine a person having a ‘guardian angel’ who would judge whether bringing that person into existence would be good for them or not, we could imagine each world having a guardian angel, who would beseech God to actualize the world only if it is good on balance. As our world is good on net, our ‘guardian angel’ would have appealed for it to be actualized. So I don’t think it should be that surprising to find our world among the vast set God would actualize. 6


Adam: I’m not sure that God would care about worlds; it seems God should care about individual lives. 7 Yet even if this world is net positive, I think there are lives which are not net-positive. Consider many wild or factory farmed animals, or the ‘horrors’ Marilyn McCord-Adams discusses. It is regrettably close to certain there is at least one life such as this that has existed on our world (and even more regrettable that it is similarly close to certain there will be more in the future).

If there were a guardian angel for everyone in our world, many of them would likely appeal for their particular person never to be brought into existence, even if they might agree between themselves that if they had to choose collectively, they would prefer all come into existence than none do. Yet it seems grotesque that God might ordain gratuitous horrors as effectively ‘collateral damage’.

Beatrice: Yes, I agree it seems grotesque. To be honest, it seems almost as incongruous with the imago dei that there are evils is merely ‘balanced off’ with outweighing goods, rather than being defeated or sublimated by them. I think that must happen, so I hope it will happen in the afterlife. I’m a universalist, and so everyone’s guardian angel would appeal them to be brought into existence: they would know, no matter how horrendous their charges lives would be, it would be made wholly good in the hereafter. 8 I’m not sure how people who aren’t universalists would deal with this problem. I guess they might be willing to accept that God would permit collateral damage, or that horrible lives without compensation in some way glorify God’s creation. I struggle to understand these views.

Adam: That may also lead to some strange conclusions. If this universalist take is correct, then worlds which are unimaginably bad are nonetheless be brought into existence by God. Suppose a world like ours but where everyone suffers the sort of horrors that would, without an afterlife, render their lives not worth living. Yet the guardian angels of the people in this world would vote unanimously in favour of the world being brought into existence.

Beatrice: I can see how that looks strange. But I don’t think it is that strange on reflection. If the sort of universalist take is correct, then all of these lives, horrendous as the initially are, are emphatically worth living given what they could enjoy in the hereafter (if you hold my convictions about the defeat of evil, then you’d say these lives will be just as emphatically valuable as those without earthly horrors). So although it might look grotesque that there is such a horrible world, and perhaps there are reasons why God would not actualize it, but it is not bad for these people to be brought into existence – heaven really will be that good. 9

Through a world-ensemble darkly

Beatrice: It is worth saying the the multiverse theodicy doesn’t necessarily entail a sort of restricted modal realism where all (on balance positive) worlds exist. Creating all of them may not be possible or the collection ill-defined. Even if not, a multiverse theodicy like this can only show a given world exists is not surprising given God’s moral perfection, not that God therefore has an all-things-considered reason to create it. God may have other considerations that weigh upon actualizing a universe or not. Yet if there is no reason for God, given his moral perfection, not to create a world like ours, the problem of evil loses its sting.

Adam: This sort of multiverse defence could apply a lot more widely than ‘just’ the problem of evil. Another branch of evidence that is suggest to ‘argue against’ God’s existence is God’s apparent hiddenness: lots of folks who appear to be epistemically virtuous (open-minded atheist philosophers of religion, for example) nonetheless do not think God exists. This seems odd if Theism is true: God wants to be known to the objects of his creation, and could easily make himself evident to them. Yet he is so non-evident that some of the people best placed to discover his existence nonetheless fail to do so.

There’s lots of discussion about whether this argument really does come out in the wash, yet the multiverse theodicy offers a straightforward reply: consider the guardian angel for this epistemically virtuous atheist. Although their life goes worse for never realizing that God exists, it looks plausible their life is nonetheless worth living. So god has at least a pro tanto reason to actualize this person, even if we grant their non-belief does not serve any divine purpose. So finding there are indeed these people is not so surprising after all.

Beatrice: I agree. God might have actualized all their believing counter-parts in other worlds. This sort of multiversal reply has a glib paraphrase that can be applied to most evidences against God’s existence: “Why didn’t God do this?” “Maybe he did, just somewhere else!”

Adam: Couldn’t this also undermine attempts to argue for God’s existence? Others have proposed that (among other things) the apparent first cause, the fine-tuning of the universe, or (for Christianity) the evidence surrounding the resurrection are harder to explain on Atheism than on Theism, and so point towards the latter.

Much of the discussion surrounding these reasons try and show that atheism has a good explanation, so the conditional probability on atheism is not as low as claimed. Yet this sort of multiversal account could undermine the denominator of the likelihood ratio. Grant the fine tuning data is really persuasive, but consider a nearby possible world which does not have any suggestive fine-tuning evidence. Although this might make the world slightly worse, my hunch would be the world definitely remains positive on net, and thus God would bring these worlds into existence. Yet then it is unclear whether if God exists we should expect to find ourselves in worlds with this sort of persuasive fine-tuning case. Mutatis Mutandis other proposals of natural theology.

Beatrice: I think it depends: I’m not a fan of ‘inscrutable’ likelihoods, so I think we should offer a guess as to how big the ‘fine tuned fraction’ of worlds created by God should be, even if there are error bars. So if the conditional probability of something given Theism is ‘not too low’ (but with error bars), but the conditional on Atheism is really really low, I think still provides some evidence, but I guess including multiverses in the mix might attenuate the confirmatory shove you would get.

Adam: I have saved the best until last. I wonder whether a multiverse might lead us to really sceptical Theism. For consider worlds where sceptical hypotheses are true: where I am a Boltzmann brain, a brain-in-a-vat, in the Matrix, and all the rest. In all of these cases I grant that my life goes worse compared to where I live in ‘the real world’, but it still seems (especially given the afterlife) the life is worth living. So God has a pro tanto reason to actualize worlds where sceptical hypotheses are in fact true.

So the question then becomes how common are worlds in which sceptical hypotheses are true out of those God would create. This seems inscrutable, but insofar as I have an intuition, it seems these worlds would be the majority: there are far more ways my beliefs can fail to track than succeed to track (replace as needed with your preferred epistemological account), and there seems no strong relationship between worlds that are on balance good and worlds where the sceptical hypothesis is false.

Yet accepting this looks like a heavy cost to bear: not just ‘for all we know we are a brain in a vat’ but perhaps ‘we are more likely to be a brain in a vat than we are to live in a universe as it appears to be.’

Population ethics for perfect beings

Beatrice: I’d still be inclined to appeal to some non-moral reasons. Even if these universes are good on balance, nonetheless the feature of there being a sceptical hypotheses makes them somehow unworthy of God’s multiverse.

Adam: Perhaps, but this line of reply could make evil problematic again. For even if the balance of good and evil is tolerable, one might think the particular quality or quantity of evil – even if ultimately defeated – puts a stain on the universe, and would make it unfit for God’s creation. I’m no divine being, but there seems a stronger all-things-considered case against actualizing a world with mass murder than a world with a Boltzmann brain.

Beatrice: If so, this would be step in Theism’s favour. Previously all one needed to do in the argument from evil was point to particularly bad things, and show it was a pretty hard sell to show how these things are nonetheless justified (or for all we know justified). Now it seems much more, “Okay, there is not a clear moral case for God including a universe like ours amongst the multitude, but because of various other reasons we should not expect God to create it”.

I accept there might be more work to do in getting a satisfactory account of what all-things-considered universes would be created, as it does look strange for God to (for example) almost entirely actualize Boltzmann brains or similar. Yet this seems a fertile project to attempt. Even if it isn’t, it seems a sceptical Theist case is in good stead here – it seems odd to expect us to have good access to God’s all things considered reasons for creation, especially when dealing with potentially infinite multiverses and stuff.

Adam: I agree. There remains a concern of ‘disconfirmed by’ versus ‘discomfirmed on’. The extension of the hypothesis of Theism to include creating multiverses may no longer be disconfirmed by the evidence of evil, yet the God who produces uncountable numbers of universes with evil might just by turns implausible and anthropomorphic (maybe perfect beings have no motive to create after all). I don’t know – it seems mysterious.

Beatrice: Yeah, I knew that already.



  1. Beatrice: I guess you might want to say that the world including its entire future will be net-positive, and so that entails optimism about how the future is going to go.
  2. Beatrice: I think this is generally plausible, but Theists should be particularly sympathetic. After all, if this were not true, why would God create anything?
  3. Adam: Okay, but this reply seems about as persuasive with or without a multiverse: if there is no best set of universes, then it does not reflect poorly on God to create a set with a single good universe rather than an infinite set of good universes, right?

    Beatrice: I suppose. Other considerations come into play. There might be something like a ‘blank cheque’ objection: if you were given a cheque to fill in as much money as you liked, it would be weird to write it for only £10, even if for any £N, you should instead put £N+1. What intuitions this rides on may not apply well to God’s creative agency, but perhaps it is a little odd if there’s a ‘long way up’ from our universe, why God didn’t pick a much better one. If he’s created (for example) all the universes from marginally net positive and increasing upwards without bound, this seems more defensible.

  4. Adam: If there is no maximal member because the set is infinite and the value of each world increases without bound, God could take a large finite sample and pick the maximum number of this set, or something along these lines
  5. Beatrice: One thing worth saying – given how recondite infinite ethics is, I’m fairly relaxed about if a multiverse has infinitarian worries. Few of us should take ourselves to have good access to the ‘right answers’ as to how to resolve the dilemmas in infinitarian ethics, so if the viability of a multiverse explanation for evil hinges on these dilemmas, then we might want to return to sceptical theist moves – even if the multiversal choices God is proposed to have made looks surprising to our moral faculties, these are unlikely to grant us reliable access to what a perfect being ‘should’ do with infinity at their disposal.
  6. Beatrice: Insofar as you think there might be ‘pattern goods’ of a multiverse of good worlds – better (ceteris paribus)variety worlds than copies of the same worlds, this might suggest other reasons for actualizing worlds with less total value, although a lot would have to be filled in.
  7. Adam: It isn’t even that implausible he might care about experience-moments, which makes the subsequent problems even more acute.
  8.  Adam: Couldn’t this reply be made without multiverses? If you think everyone goes to heaven, which defeats any worldly evil no matter how horrific, the prevalence of evils, horrors included is (pardon the flippancy) a minor irritation prior to the Eschaton.

    Beatrice: I don’t think so. The problem of evil is often accusatory. Hence the idea of ‘putting God on Trial’, “Why, O God, have you forsaken me?”, and others. In cases like this, the ‘problem of evil’ is perhaps little more than a fig-leaf over naked despair. For these people I would want to console them with my universalist optimism that God will make everything right.

    Yet I think some people err in thinking all examples of  discussions over evil are of this type. You might have a wonderful life, Adam (I hope you do), yet you could still soberly look at the balance of good and evil in the world and think it doesn’t plausibly comport with the sort of God I believe in. You wouldn’t need consolation, and if I told you how you shouldn’t worry about evil because of God defeating all evils in the hereafter, I imagine you’d say something like, “Well, that’s great if it were true, but I don’t think that answers my question: even if God will defeat all these evils, why would he permit them in the first place? It doesn’t seem to serve any purpose, and that’s surprising if God really does exist like you say he does. It looks much less surprising if God doesn’t exist and the universe is indifferent to our well-being, so I still think this is evidence against God’s existence.”

    The hope is that realization that God could (would?) create worlds that are net positive would make finding ourselves in a net positive world with unjustified evil unsurprising if God exists. As you say, individual lives may remain net negative, so the universalist answer removes this possibility – no matter what unjustified evils someone suffers, it is better they exist, and so God does not wrong them by bringing them into existence.

  9. Adam: This is more an intellectual curio, but one could genuinely have a ‘problem of good’: if we think that worlds with people in would generally have loads more evil than the actual world, but all of this evil can be defeated by a universalist hereafter, a universalist may expect the multiverse to be predominantly composed of worlds with lots of evil, and so the relative paucity of evil in our world is surprising. If this balance of evil is less surprising on atheism, then this could be an argue in favour of it (!).

Leave a Reply

Your email address will not be published. Required fields are marked *