Why might God create?

Abstract

Some have held that this question cannot be given a satisfactory answer within the constraints imposed by a traditional theistic framework, a framework which sees the universe as contingent on God's free choice to create it. However, I shall argue that, while traditional theism certainly does place constraints on possible answers to the title question, on assumptions which are themselves plausible, these constraints leave open various possibilities for understanding the reasons God may have had as He decided whether or not to create something other than Himself.

Introduction

Some have held that the title question cannot be given a satisfactory answer within the constraints imposed by a traditional theistic framework, a framework which sees the universe as contingent on God's free choice to create it. Spinoza, for example, thought that God could not have any reason to create something other than Himself. Spinoza thus concluded that God must be identical to the universe, rejecting the traditional theistic framework in part on that very basis: God's creation of anything is impossible.¹ Leibniz, by contrast, reasoned that God would be required by the Principle of Sufficient Reason to actualize whichever logically possible world is the best and thus could not fail to do so. The Leibnizian (even if not Leibniz himself²) thus concludes that God's creation of exactly this universe, in all its particularity, is metaphysically necessary. I shall argue that the situation is more complicated than either Spinoza or Leibniz thought. The traditional theist can steer between the Scylla of Spinozism (God's creating anything and thus this universe is impossible) and the Charybdis of Leibnizianism (God's creating this universe is necessary). God's contingently creating this universe is possible.

Two constraints on possible answers imposed by traditional theism

First then, traditional theism says in answer to our title question that God didn't have to create. In some now-usual terminology, this is rendered as the assertion that the possible world in which God alone exists is an actualizable world. Contra Leibniz then, whatever we end up saying about God's reasons (if any) for creating something other than Himself, we cannot end up saying that He had reasons to create which were such that, had He failed to act on them by creating and instead remained the sole existent, He would have been in some respect deficient, imperfect in rationality or – if that isn't already entailed – imperfect in goodness and love.

Let's call the possible world in which God creates nothing ‘Just God’. Maintaining Just God as actualizable places one constraint on possible answers to our title question.

Obviously Just God is not the actual world; traditional theism, contra Pantheism, does not identify God with the universe. Rather, as well as God, there is this creation, including, at the least, us and the universe we see around us. And so we may further say that, second and contra Spinoza, God must have had it in His power to create at least one thing other than Himself, namely the actual creation. That the actual world is actualizable is of course even more obvious than that Just God is actualizable.

Let's call the actual world ‘God Plus This Creation’. The actualizability of God Plus This Creation means that one cannot, in consistency with traditional theism, say that God Plus This Creation is a world such that, in actualizing it, God acted in any way contrary to His essential nature, as a perfectly rational, good, and loving being.

Having God Plus This Creation as well as Just God as actualizable immediately generates various implications which are congenial to the traditional theistic world-view. That these two worlds are both actualizable entails that God may be pictured as free in the libertarian sense, where freedom (at least for Him) requires that He have had at least two alternatives metaphysically open to Him at the moment of choice. And, assuming this model is defensible, it is easy to show that God was not limited to choosing between only the two options we've so far been considering, Just God and God Plus This Creation, in that it is immediately plausible that there are many other worlds no less possible than these. The simplest way to see this is to consider universes which are like our own save that they differ in some very minor and value-neutral way from it. Clearly, if – as the model supposes – God were free to create our universe, He would have been equally free to create any of these. There are very plausibly an infinite number of such universes and thus it's very plausible that there is an infinite array of actualizable worlds. Thus, as Kraay puts the nature of God's choice as traditional theism depicts it, ‘In contemporary analytic philosophy of religion, this choice is construed as follows: God surveys the set of possible worlds, and then freely chooses exactly one for actualization.’³

What these constraints imply about God's reasons or lack of them in creation

Bringing together our first two points (contra Leibniz, that Just God is actualizable and, contra Spinoza, that God Plus This Creation is actualizable), we can see that the context in which God made His choice between Just God and God Plus This Creation must have been such as to have meant that God would have been no more reasonable in actualizing Just God than to have done what He did do, actualize God Plus This Creation, and no more reasonable in doing what He did do, actualize God Plus This Creation, than He would have been had He actualized Just God. And the same considerations of course generalize over any pair of worlds from among the infinite array which one must imagine God as equally well choosing between. Or rather and more carefully, we must say that He would not have been in any problematic way ‘more reasonable’ whichever way He had chosen; whichever way He had chosen, He would not have been, we might say, ‘irrational’ even if He might have been, we might say, ‘arational’ when an opportunity for Him instead to have been rational presented itself. We now need to take some time explaining why we need to be a bit more careful in this way and why we need to adopt this quasi-technical terminology. Doing so will help us see as possibilities things which Spinoza and Leibniz could not see as such. And it will thus help us avoid Spinoza's error (no reason could direct God to create and thus He could not do so) on the one hand and Leibniz's error (God's reason to create must have necessitated His choice in all its particularities and thus He could not but create and could not but create exactly what He has created) on the other.

One reason why we need to be a bit more careful is that there could be in play reasons of the sort Murphy calls ‘justifying reasons’.⁴ Although philosophical reflection on this type of reason goes back much further than him in time and although I disagree with Murphy at many points, Murphy's work has most significantly sharpened my own thinking on the topic of this article, so I shall use his terminology in what follows. To underscore one point then: that I use his terminology does not imply that he would agree with all the claims I make using it.⁵

Justifying reasons are reasons of a sort which, while genuinely reasons for action of a particular type, are such as to mean that, by not following them in performing the relevant action, one does not thereby become irrational even when not following them due to following stronger or as-strong countervailing reasons. Perhaps an example may help.

Suppose I am choosing a tie to wear with a white shirt and grey suit and that I have a selection of ties in front of me all of which, for the sake of argument, are qualitatively identical save in their colours. (I have specified the colour of the shirt and suit with the intention of generating as an implication that no particular tie-colour would ‘go’ with them more than would any other; if my reader doubts that I've succeeded in that intention, I leave it to him or her to fill in further details to generate it.) In such a circumstance, the fact that my favourite colour is navy blue and that a particular tie among those from which I'm choosing is navy blue jointly constitute a reason for me to select that tie over the others. But if I choose the colour of the tie that I'll wear without reference to that fact – entirely whimsically, as we may put it – while not acting on the reason that navy blue is my favourite colour and not indeed acting in respect of its colour on any reason at all (this is the weight ‘whimsically’ is intended to carry), then, while it could thus be said that I do act less reasonably than I might have done, this is not for me to act in any problematic way less reasonably. In a situation such as this, I choose a tie which I have by hypothesis less reason (as no reason) to choose over the navy-blue tie, but I do not, by being less reasonable than I might have been in this way, become thereby, as we might put it, irrational; I become simply, we might say, arational in this particular when an opportunity for rationality in this particular presented itself.

Justifying reasons are reasons of a sort which one may simply ‘put aside’, as it were, in one's deliberations and not be irrational for doing so. One may indeed put justifying reasons aside due to there being reasons in favour of a mutually exclusive course of action, but also – as the example just given illustrates – one may put them aside in favour of no other reason and simply act whimsically. I define acting whimsically then as acting for no reason, which acting then, I am suggesting is not always acting irrationally, even if it is always acting, as I am putting it, arationally.⁶ There are complications in that there can be levels of whimsicality. So, in the situation just imagined, it could be said that if I choose a tie without reference to the fact that navy blue is my favourite colour, I act whimsically with respect to the colour of the tie that I pick, but, even so, if it is the case that sartorial conventions give one a justifying reason⁷ to wear a tie with a suit, then my choice to pick a tie as such may have been made by reference to those reasons and thus not have been whimsical. If so, I am not acting whimsically ‘all the way down’, as it were. A change in the thought experiment to a situation in which I am wearing a more causal outfit, composed in part of a jacket of a sort that may or may not be worn with a tie and for an occasion which is sufficiently balanced with respect of formality/informality so as not to provide me with a reason one way or another on the tie issue, would provide a situation where it might be whimsical that I choose a tie at all and also whimsical that I choose the colour of any tie that I do choose. So, my own view is that God too may be whimsical and indeed whimsical all the way down. But while it is relatively uncontroversial that humans can be whimsical in the sense defined, it is much more controversial that God can be whimsical, at least all the way down. And it's not just ‘extremists’ such as Leibniz who reject it as a possibility; Murphy also appears to reject it.⁸ However, as I have argued elsewhere for God's being able to act whimsically and do so all the way down,⁹ I shall not do so again in this article. Rather, I shall, as my argument proceeds, mark where possibilities open up only on the assumption that divine whimsicality at least at some levels is possible, so that readers sceptical about divine whimsicality may the more readily discount these as possibilities. As one would expect, if one allows for divine whimsicality, the constraints on possible answers to the title question are looser than if one does not; a fortiori, if one allows for it all the way down.¹⁰ But even if one does not allow any divine whimsicality, the constraints are not, I hope to show, uncomfortable ones for the traditional theist. I should also acknowledge at this point that not everyone believes in the existence of justifying reasons, though positing them as potentially in play in the divine mind is less contentious than positing that divine whimsicality may be in play there. Still, similarly, it is beyond the scope of this article to argue for them (I refer readers to the works mentioned in an earlier note) and, as they are less controversial, so I shall proceed more boldly in utilizing them, leaving more work to the reader who doubts them to subtract from my conclusions the options which require them.

Murphy well sums up the difference between justifying reasons and reasons of a sort with which we are perhaps more familiar, which he calls requiring reasons:

Requiring reasons are rational constraints: a requiring reason for A to Φ rationally necessitates – that is, makes rationally necessary – A's Φ-ing in the absence of contrary considerations. Justifying reasons provide rational opportunities: a justifying reason for A to Φ rationally possibilitates – that is, makes rationally possible – A's Φ-ing in the absence of contrary considerations.¹¹

Another reason we need to be a bit more careful then is that there could be, in Murphy's terms, requiring reasons in play. We'll look at that next.

In order to keep Just God and God Plus This Creation both actualizable, we shall need to say that if at God's moment of choice there were requiring reasons in favour of His actualizing one of these worlds, then there were also requiring reasons which were no less and no more weighty in favour of His instead actualizing the other; and indeed in favour of His actualizing each other world in the infinite array of actualizable worlds. This is because requiring reasons are those which one may not simply put aside in one's deliberations without thereby becoming irrational. On the contrary, one needs ‘balancing’ reasons against doing what one has requiring reasons to do if one is not to be rationally compelled to do it (or in fact rationally compelled to refrain from doing it, if the reasons against it ‘over-balance’ the reasons for it). A point worth flagging then, in that it may be surprising given the name, is that requiring reasons for an agent's performing a particular action do not as such rationally require the agent to perform that action. They do not do so as there might be sufficiently weighty requiring reasons against his or her performing the same action, either sufficiently weighty so as to balance out the requiring reasons in favour of performing it (leaving either performing it or refraining from doing so equally reasonable) or even more weighty, so as to outweigh the reasons in favour of performing it, making performing the action in favour of which there are genuinely requiring reasons something which it would be all-things-considered irrational for the agent to do.

If there are, across the whole infinite array of metaphysically possible worlds, such balancing requiring reasons, then again God would be free to choose from among all members of that array without risk of irrationality. Indeed, the array of metaphysically possible worlds must be conceived by the traditional theist as being constituted as it is due to the balancing of any requiring reasons that there are in play across all of its members: within the array are all and only those worlds for the actualization of which there are no unbalanced requiring reasons; outwith (logically possible but not metaphysically so, given God's existence and nature) are all those worlds against the actualization of which there are unbalanced requiring reasons. While in the end I myself shall evince scepticism that there could be any requiring reasons in play across the array,¹² it is open to the theist to speculate that perhaps God was in a requiring-reason-balanced situation vis à vis His choice whether to actualize Just God, God Plus This Creation, or some other member of the array. Another example may help to illustrate the nature of requiring reasons.

Most will think that one has moral reasons (of a requiring sort¹³) to keep one's promises (at least all other things being equal) and that one has moral reasons (of a requiring sort) to help one's friends (at least to some extent and presuming that one can do so at negligible cost to oneself). If that is granted, then one should be able to construct in one's imagination a case where a given person, let's call him ‘Stew’, is faced with a choice between hurrying on his way so as to keep his promise to be on time for an appointment with a given friend, let's call him ‘Nick’, but thereby failing to assist another friend, let's call him ‘Tim’, who has just had a minor accident in his company or delaying so as to assist Tim, but thereby being later than promised for Nick. As one can vary the solemnness with which Stew made the promise; the severity of the accident; and other relevant factors through their own continua, so one should be able to generate in one's imagination a case where one judges that the requiring reasons for Stew keeping his promise to Nick are neither stronger nor weaker than those for him assisting Tim.¹⁴

In such a requiring-reason-balanced scenario, while Stew may not simply put aside the moral reasons for doing whichever of these he decides not to do, he may – without any irrationality – not act on the reasons on the one side but instead act on the reasons on the other. Whichever way Stew ends up deciding, if he is not to have failed of rationality, he will need to do what he ends up doing for the relevant requiring reasons, not (or at least not solely) for other reasons. It will be permissible for Stew to fail to assist Tim in order to keep his promise to Nick or for him to fail to keep his promise to Nick in order to assist Tim. But it will not be permissible for him to fail to assist Tim merely in order to enjoy a brisk walk or to fail to keep his promise to Nick merely in order to enjoy a more leisurely walk. This is what is meant by claiming that requiring reasons cannot simply be put aside as can justifying reasons.

Similarly, it could be maintained that God could have had requiring reasons in favour of actualizing Just God, ones which were neither stronger nor weaker than the requiring reasons He had in favour of actualizing God Plus This Creation. And indeed, as mentioned, the same would then need to be true of each of the other actualizable worlds; we need a balancing of any requiring reasons across the array. Given what choice we know He did make – actualizing God Plus This Creation – and that requiring reasons are ones which cannot simply be put aside in deliberation, we can thus infer that if God did have requiring reasons for actualizing God Plus This Creation, these will be the reasons for which He acted (or at least among the reasons for which He acted). We will not be able to infer the same if God had only justifying reasons, for He may simply have put these to one side when choosing, just as I sometimes choose a navy-blue tie to wear entirely whimsically from among my set of ties, without reference to the fact that navy blue is my favourite colour. Or at least that is so if divine whimsicality is allowed as a possibility.

So let's sum up where we've got to so far. And, in order to keep things simple (well, as simple as they can be kept), let's continue to focus on God's choice of which world to actualize as if it were a choice between merely Just God and God Plus This Creation, though the results generalize across the infinite array of actualizable worlds. On the traditional picture, we may say that God's reasons for creation, if any, could be justifying reasons, not requiring ones. He may have had justifying reasons to actualize God Plus This Creation and justifying reasons to actualize Just God. Alternatively, He may have had justifying reasons only on one side, though if one holds that this is possible, then one needs to hold that God could engage in purely whimsical action so as to keep both worlds actualizable. Or again, God may have had justifying reasons on neither side, presuming again that one allows Him whimsical action. If God did have justifying reasons in favour of actualizing God Plus This Creation, it may have been that He acted in creating on those reasons, but, presuming one allows divine whimsicality, He may have actualized God Plus This Creation whimsically, that is, created it but not for the reasons which were in principle available to Him and for which He could have created it, just as I sometimes happen to choose a navy-blue tie to wear without reference to the fact that navy-blue is my favourite colour or for any other reason. Alternatively, God's reasons, if any, for actualizing God Plus This Creation may have been requiring reasons, but, in that case, He would have had to have requiring reasons of a sort which were not weaker (or stronger) to actualize Just God and indeed the same would need to be true of each member of the array of actualizable worlds, not simply the two on which we have been narrowing our focus. And then, if God did have requiring reasons in favour of actualizing God Plus This Creation, He must have acted (at least in part) on the basis of those requiring reasons in creating, for requiring reasons are precisely those which cannot simply be put to one side in one's deliberations. And then finally again, if one allows divine whimsicality, God may have had no requiring reasons as well as no justifying reasons, necessitating His deciding between His options entirely whimsically.

While this article is mainly concerned with articulating the statuses of God's possible reasons – as justifying or requiring – and the connections between those statuses and the axiological statuses of the possible worlds between which traditional theism depicts God as choosing; and while progress in that task may be made without committing to any substantive view about the content of those reasons, nevertheless, before moving on, I shall venture a few speculations as to what reasons God may have for and against creating something other than Himself in general and what reasons He may have had to create what He has created in particular. Having a few examples of answers to the title question which we can then see operating in the structure of justifying and requiring reasons and show relating to the axiological statuses of the relevant worlds will ground the abstract points and, in that way, make them more tractable to assessment. So – again stressing that if the suggested answers to our title question are not plausible to the reader, he or she should be able to substitute others without loss to the argument – what might God's reasons for and against creation have been? Let's consider against first.

First, Just God has a certain value-purity about it; God exists there in the most splendid of isolations, reality's value being unalloyed by anything that is to any extent less than completely perfect. There's nobody there other than Himself to benefit from this of course, but there's a certain unblemished pristineness and ultimate simplicity about reality in Just God. It seems plausible (to me anyway) that maintaining that value-purity would have been a justifying reason for actualizing Just God. If it would have been a justifying reason, then God would of course have been free to put it aside and actualize God Plus This Creation regardless of it. Only if one were to posit that maintaining value-purity would have constituted a requiring reason should one worry about a potential threat to God's freedom not to act on it, a threat which one would then need to counter by positing balancing requiring reasons in favour of actualizing the other worlds in the array.¹⁵ Of course, it's also open to the theist to say that the nature of Just God does not provide God with any reason – even a justifying reason – to actualize it, merely no requiring reasons not to actualize it (though one might then need to allow for divine whimsicality, if one said that the nature of God Plus This Creation does give God justifying reasons for its creation¹⁶).

Now let's look at possible reasons to create.

Second, if God were to create something which instantiated values not present in His own self, He would add to the value-variety of reality and that gain in value-variety could itself be argued to be a good thing. So, for example, it's plausible (to me anyway) that the creation part of God Plus This Creation contains various values which are not present in the God part (or thus in Just God). By way of example, one might consider those values which cluster around sentient and morally conscious free agents such as ourselves making significant choices. We sometimes act courageously, self-sacrificially, and with strength of will. God never has need – or indeed ability – to do any of that. If so, then God would gain in value-variety were He to actualize God Plus This Creation or indeed actualize a host of worlds other than Just God, every one the creation element of which has values not present in God's own self. And that increase in value-variety could itself be argued to provide a justifying reason to fail to actualize Just God and instead to actualize God Plus This Creation or one of the other members of the array with such creations in them. The values peculiar to our creation (and creations relevantly similar) could be argued to give God justifying reason to create our universe (or one relevantly similar) in particular. Again, one could hold that, instead of being merely a justifying reason, adding to value-variety would have been a requiring reason for God to act as He has done, but that would again entail that one needed to hold that there were balancing requiring reasons across the array so as to preserve the actualizability of Just God and a host of other worlds. And casting this as a requiring reason would threaten to exclude from actualizability worlds whose creation elements contain creatures who instantiate only values already present in God's self or worlds which contain only value-neutral elements. And, to me anyway, that doesn't seem credible, though I must acknowledge that it is not so absurd a consequence that someone who eschews both justifying reasons and whimsicality need have much fear from needing to draw it.

And then finally, again it's also open to the theist to maintain that the nature of God Plus This Creation gives God no reason – even a justifying reason – for actualizing it, merely failing to give Him any requiring reasons against doing so, though one would then need to allow divine whimsicality to explain why God had done that which He ex hypothesi had no reason at all to do, actualize God Plus This Creation.

If one does allow divine whimsicality, one may say that even if God had justifying reasons on one or both sides, it may be that He did not act because of the reasons that would have spoken in favour of the decision that He in the end took; sometimes, I pick out a navy-blue tie to wear from a selection without my doing so being explained by navy blue being my favourite colour or by any other justifying reason; it may even be that I have no reason to wear a tie rather than not that day. And God, one may say, could act likewise, creating arationally, but not irrationally.¹⁷

What implications these being the possibilities for God's reasons has for the axiological properties of actualizable worlds

We are now in a position to map a couple of potentially navigable routes through the next bit of territory which traditional theism needs to traverse on its way to explaining how it may assert that there are answers to the title question which are compatible with it.

Given the theistic God's necessary perfection (in particular as it pertains to His beneficence), the claim that Just God is actualizable and the claim that God Plus This Creation is actualizable together entail that either (a) Just God and God Plus This Creation do not differ in their overall axiological status such that one is better than the other or (b) they do differ such that one is better than the other, but God is not compelled by His perfection to actualize the better of them. Option (a) quickly entails that God could have no reason of a requiring sort to actualize the one over the other and thus resolves our Spinozian/Leibnizian dilemma. Option (b) in turn entails that either (i) the relative difference in overall axiological status between Just God and God Plus This Creation does not provide a requiring reason for God to actualize one rather than the other (at most it provides a justifying reason) or (ii), if it does provide a requiring reason, then it is a requiring reason that is balanced by a requiring reason on the other side (and the balancing stretches over the whole array of actualizable worlds, i.e. obtains between any two members of the array). And this too would resolve our dilemma. Let's look at the two disjuncts of this conclusion to which traditional theism commits one – (a) and (b) – in order, (a) first.

How then, we may ask, could it be true that (a), Just God and God Plus This Creation fail to differ in their overall axiological status such that one is better than the other? There are two ways this could happen.

Either (a1) the values instantiated in them are commensurable such that these two worlds may be ranked relative to one another, but when so ranked they come out as of exactly the same overall value or (a2) there is value incommensurability between them such that one cannot be ranked relative to the other and thus neither can be said to be better than the other. There are two ways of arguing for (a1), one of which, (a1.1), makes claims about the net zero value of creation and the other of which, (a1.2), makes claims about the infinite value of God. We'll look at those ways of arguing for (a1) in order, (a1.1) first. And we'll see that, on the assumptions of value commensurability and a certain understanding of how to compare infinities which we may take as the starting points for the (a1) line of thinking, (a1.2) does indeed provide good reason to endorse (a1) and thus (a) even though (a1.1) does not. However, this is, it will be concluded, something of a Pyrrhic victory as these starting assumptions themselves are probably wrong. After looking at arguments for (a1), we'll turn to consider arguments for (a2) and we'll see that this provides a good reason to endorse (a) which is not so dubiously founded. Unrankability, while not following of necessity from incommensurability, is, I shall assert, plausible on it, and we can grant it at this stage in the argument of this article as we'll turn later to consider how things would play out if it were not granted. And thus we'll be in a position to conclude that, whether or not one is attracted to value commensurability, (a) looks like a route for the traditional theist to travel that is not without hope of navigation. However, it is, we will have seen, only navigable on one branch of the argument if, in addition to commensurability, a certain understanding of how to compare infinities is granted and, on the other branch of the argument, if, in addition to incommensurability, unrankability is granted.

Having considered the issues in connection with defending (a), we'll then turn to consider (b), the assertion that Just God and God Plus This Creation do differ such that one is better than the other (they can be ranked and when ranked one comes out as higher ranked than the other) but that God is not compelled to actualize the better, the highest ranked of them. By this stage, we'll already be in a position to see that on the assumption of value commensurability and a certain understanding of how to compare infinities, (b) is not plausible – given the (a1.2) manner of argument which by then we'll have endorsed as valid – and thus we'll be in a position to see that (b) is only going to be plausible if one grants value incommensurability or eschews a certain understanding of how to compare infinities. As already indicated, incommensurability does not entail unrankability and while by this stage it will have been asserted that unrankability is plausible on incommensurability, asserting the rankability of Just God and God Plus This Creation is by no means absurd. Similarly, even if one endorses commensurability, if one eschews the way of understanding how one may compare infinities that was one of the starting presumptions of the (a1.2) manner of argument, rankability again becomes a possibility. Thus, the Leibnizian view that agents are always required to actualize the highest-ranked state of affairs causally available to them still threatens traditional theism's capacity to maintain the possibility that there is an answer to our title question which is compatible with it. However, we'll see that there are two reasons for realistic hope on the part of the traditional theist in this context.

First, we'll see that it is rather implausible to maintain that a state of affairs being the highest-ranked state causally available to an agent provides that agent with a requiring reason to actualize it when failing to actualize it would not harm anyone. The idea that requiring reasons are always creature-affecting reasons is itself plausible and grounds this conclusion. It's rather more plausible to maintain that a state of affairs being the highest-ranked state in such circumstances provides a justifying reason for actualizing it. But of course justifying reasons are precisely those which one may put aside in one's deliberations without irrationality. Second, we'll see that at least over that part of the array with creations in its worlds, it's also rather plausible that there is no highest-ranked state of affairs available to God, Him thus being in what is usually called a ‘no-best-world’ scenario. In such a scenario, on the assumption that one always has to be able to do something which does not render one irrational, A's being a higher ranked state than B cannot be said to be a requiring reason for actualizing it (for otherwise, per impossibile, B's being higher than C would be a requiring reason to actualize it; and C's being higher than D. . .; ad infinitum) unless one is also prepared to say that it is an always-balanced reason, balanced (for example) due to there being a requiring reason against actualizing it because in favour of actualizing the world one better than it. And we may thus conclude this argument by saying that either being better than an alternative is a justifying reason and as such one that God is free to put aside or it is a balanced requiring reason and as such one which God is again free not to act on. He thus has carte blanche over which world to create from among this portion of the array.¹⁸ So finally, we'll be in a position see that, whichever way one calls it on the issues of commensurability/incommensurability and rankability/unrankability, there is plenty of reason to think that Spinoza and Leibniz were wrong: there is plenty of reason to think that traditional theism can allow for there being an answer to the title question which respects the constraints it imposes.

How Just God and God Plus This Creation could fail to differ in overall axiological status due to their having the same status

The first way that Just God and God Plus This Creation could fail to differ in their overall axiological status such that one is better than the other, which we are labelling (a1.1), is if Just God and God Plus This Creation exactly tie with one another for overall value due to the fact that the creation part of God Plus This Creation is of zero value. The simplest way for the value of creation to be zero would be if creation had no valuable or disvaluable elements in it at all, if all parts of creation were entirely, as we might put it, value-neutral. Sadly (from the point of view of simplicity), this is manifestly false. Had the creation part of God Plus This Creation been merely a single momentary point particle, then it might have been plausible that there was nothing good or bad about the creation part of God Plus This Creation; it might have been plausible that creation was value-neutral through and through.¹⁹ But that is not what the creation part of God Plus This Creation is like. The creation part of God Plus This Creation has a huge and complex universe in it and this universe is shot through with values and disvalues. This universe is one with elements of great beauty, but also elements of great ugliness. It is one with at least billions of morally sentient, significantly free creatures, persons such as ourselves (there are that many on this planet and there are billions of planets in the universe). And it is one in which these persons are daily helping one another and hindering one another; being virtuous and being vicious; acting with good intentions and acting with bad. Thus, there clearly are a manifold of both good things and bad things in the creation part of God Plus This Creation. And so the view that the value of creation is zero due to creation having no valuable or disvaluable elements in it at all is not tenable. (It is also, of course, in itself incompatible with traditional theism.)²⁰ The next most simple story to the effect that creation is of zero value would have it that while creation has good parts and bad parts, these precisely balance one another out, leaving creation of zero value net. This story, while not manifestly false, is highly implausible. All of this being so, I suggest we may pass by this way (a1.1) of understanding the first way (a1) that (a) could be true and turn to the other way of understanding it, the way we have labelled (a1.2).

While (a1.1) adopted a tactic of seeking to render credible the thought that the contribution to the value of the world made by its creation part is zero and use that (alleged) fact to explain the (alleged) equality of value between Just God and God Plus This Creation, (a1.2) seeks to explain this (alleged) equality of value by rendering it plausible that, whatever the value of creation, it is beyond negligibility given the relative value of God.

The idea behind the tactic we are calling (a1.2) is expressed succinctly by Nozick at one stage: ‘God isn't going to create a world in order to increase his own value or goodness, or whatever – that's already infinite. Neither is the total amount of value there is going to be increased; adding a finite amount to an already infinite amount doesn't make that any bigger.’²¹

This simple ‘swamping’ by God's value of the contribution made to the axiological status of a world by its creation part follows if we assume complete value commensurability; that cardinal arithmetic provides the best way of thinking about infinities in this context; and that God has infinite positive value. On those presumptions, the conclusion follows unless the creation part of a world is itself of infinite negative value. But, at least in the case of God Plus This Creation, that seems unlikely enough to be ignored as a possibility. And it's arguable that if there is a God, then of metaphysical necessity there couldn't be anything of infinite negative value anyway (there can't be anything as bad as God is good), so worlds containing such elements are ruled out as actualizable on those grounds. The target conclusion (a) is secured: Just God and God Plus This Creation have the same value and thus neither is better than the other. In short, I think that, given its starting presumptions, this swamping argument works. However, these starting presumptions should probably be rejected.

First, the commensurability of all value is something which, to me anyway, is deeply implausible, but a quicker case to make against the swamping argument focuses on the alleged adequacy of cardinal arithmetic to the task of thinking about infinities in this context. It is not an entailment of commensurability that cardinal arithmetic is the best way of thinking about value and it seems as if those attracted to commensurability would in fact be well advised to eschew it. So, consider a range of logically possible worlds, each containing a creation composed of a single angel enjoying the beatific vision. The first is a world in which the angel everlastingly enjoys the beatific vision at a given level of hedonic intensity, which we call level one; call this world ‘God Plus Angel in Beatific Bliss Level 1’ and say its ‘n-number’ is 1. The second is a world in which the angel enjoys the beatific vision at double that level, ‘God Plus Angel in Beatific Bliss Level 2’, n-number 2. The third is a world in which the angel enjoys it at three times the level enjoyed in God Plus Angel in Beatific Bliss Level 1. And so on. So, in this range of worlds, if we suppose that that which admittedly some would say is metaphysically impossible, namely that each such creation has already existed for an infinite amount of time, then each creation has infinite beatific bliss in it (and ex hypothesi no disvalues). Nevertheless, clearly there is a sense in which higher n-numbered worlds are better than lower – there can be a ranking. In Kraay's terms: angel-worlds form a cluster, each of which has infinite bliss in it, yet each of which can be ranked nonetheless relative to the others.²² So, the swamping (a1.2) argument is one which I believe should in fact be rejected as sound even by those who favour value commensurability, although it is valid.

How Just God and God Plus This Creation could fail to differ in overall axiological status due to incommensurability and unrankability

Let us turn then to consider the second way (a) might be true, the way which we labelled (a2): there is value incommensurability between Just God and God Plus This Creation such that one cannot be ranked relative to the other and thus neither can be said to be better than the other. In short, I shall argue that value incommensurability and (though it does not follow from incommensurability) unrankability are plausible and so (a2) is a plausible alternative route to (a); indeed, given the doubts one may reasonably have about the presumptions of the swamping argument, I venture it as a preferable route

One should, I suggest, think that there is incommensurability between at least some of the values in Just God and some in God Plus This Creation. So, as sketched previously, a finite agent's self-sacrificially doing what he or she believes (but does not know for sure) is the right thing after struggling with temptation to do what he or she believes (but does not know for sure) is the wrong thing seems to me to instantiate a cluster of values which are different from and incommensurable with those instantiated in God's own self. In such a situation, the finite agent shows a certain moral courage; a strength of will; a willingness to put others above himself or herself; a certain type of heroism; and so on. God never does anything at all like that²³ and the goodness of this cluster of values does not seem able to be cashed-out in the same terms as all other goods, in particular the goods instantiated in God's own self. That being so, I think that one should in fact conclude that some values in the creation part of God Plus This Creation are qualitatively distinct from and incommensurable with some values in the God part, which latter are the only values that there are in Just God.²⁴ However, as already indicated, incommensurability, while compatible with unrankability (and indeed a necessary condition of it) does not entail it (it is not a sufficient condition for it). Nevertheless, for myself, I incline to think that Just God and God Plus This Creation cannot be ranked.²⁵ As I think about the values in each, I conclude that neither is better than the other; they are just different. Of course, this is a first-order value judgment; I have offered no argument for it; and one might contest it. We'll look at how things should be configured if it is wrong in a moment. But if it is right, we may say that God could have remained the sole existent and, had He done so, then He could have taken as His justifying reason for doing so the fact that by doing so no disvalue of any sort would have been allowed into reality – complete value-purity would have been maintained. Given what He did in fact do, actualize God Plus This Creation, we may speculate that He may have taken as His justifying reasons the fact that by actualizing this world values of the sort peculiar to us get to be instantiated, but He may simply have acted so as to instantiate value-variety and then chosen our world on no other basis than that it fulfilled that requirement. Or He could – if whimsicality is allowed – have created with no justifying reasons at all. He may have put all potentially justifying reasons aside. It is possible to cast these (or other reasons) as requiring reasons too – as long as one posits that they balance across the array – though I have not suggested this is advisable and will not explore it further here.

What then if my first-order value judgment is wrong and these worlds can be ranked? On such a supposition, in order to be consistent with traditional theism, one needs then to maintain the (b) line of thinking: God is not compelled to actualize the better. How could this be so?

Suppose one held that Just God is in fact the highest-ranked world, due to one's eschewing the swamping argument and holding that the absolute and unalloyed value-purity of Just God makes it qualitatively and incommensurably better than all other worlds; Just God has, one might say, lexical value-priority over all other possible worlds. Even so, what we called the (b) lines of thinking seem navigable: it seems implausible that its having this status would provide God with a requiring reason to actualize it, for the quite general reason that it is implausible to maintain that a state of affairs being the highest-ranked state causally available to an agent provides that agent with a requiring reason to actualize it when failing to actualize it would not harm anyone.²⁶ Who would God be ‘letting down’ (in the manner that Stew would be letting down one of his friends in the example given previously) were He to fail to actualize Just God and instead actualize another world, with a creation element? And so in virtue of whom could He have the alleged requiring reason not to create? It is hard to think these questions are going to have credible answers that would provide reasons to worry the theist.²⁷ Requiring reasons are grounded in what one might follow contemporary scholarship and call their ‘person-affecting’ nature, though, given that we plausibly have requiring reasons with regard to sentient creatures who are not persons (dogs, for example), it would perhaps be better to call it their ‘creature-affecting’ nature. If any agent is to have a requiring reason to perform some action, that must be because some actual morally significant being will be affected for good or ill by that agent's performing that action, but – prior to creation – there were no such creatures in whom requiring reasons could be grounded. And it could not sensibly be suggested that God Himself would be harmed in some way by the lack of value-purity were He to create, well, not on traditional theism as I understand it (those with weaker versions of aseity and impassability might have more trouble here). While the well-being of those who do not yet exist and indeed whose existence itself depends on whether or not one performs a given action can give one requiring reasons to act in a certain way, one cannot have requiring reasons generated by their possible well-being to bring creatures into existence. The negative well-being of potential creatures can give God a requiring reason not to create them. However, the positive well-being of potential creatures cannot give Him a requiring reason to create them.²⁸ This asymmetry has been long noted in the context of human population axiology. It receives what is to my mind irrefragable support from intuitions such as the following. A healthy couple who could irradiate their gametes so as to conceive a child genetically prone to great suffering are required not to do so, while if they could irradiate their gametes so that the child conceived would have some genetic advantage (perhaps be a little bit less prone to develop Type II Diabetes than the average), they are not required to do that.²⁹ However, all that having been said, the fact that a given state of affairs would be the best from an impersonal (non-creature-affecting) standpoint does, it seems to me, give one some reason to actualize that state of affairs; so if it is granted that Just God is the highest-ranked world, its being so plausibly gives God a justifying reason for actualizing it. Though, it being only a justifying reason, God could of course have put it aside without any irrationality. The fact that the theist is committed, if he or she configures things in this way, to positing that this is what God has done is therefore no weakness in his or her position. He or she may maintain that Just God would have been best, but it being the case that nobody would be harmed by God's actualizing less-than-the-best, God was not required to actualize the best.

If one brackets off consideration of Just God for a moment and considers from among the infinite array all and only those worlds with things other than God in them (i.e. worlds with creations), it is plausible that – from among this proper part of the array – either (given commensurability; cardinal arithmetic; and God's infinite value) all worlds are equally highly ranked (due to the swamping argument) and thus there is no unique highest-ranked world or (given incommensurability and unrankability within this proper part of the infinite array) there is no unique highest-ranked world. (We may be granting at this stage for the sake of argument that Just God has ‘lexical priority’ over all members of this proper part of the array, in the sense that all members of this proper part are rankably lower than Just God even if not themselves rankable such that one emerges as highest-ranked among them.) If so, then God is in a ‘no-best-world’ scenario at least as He contemplates this proper part of the array. And thus again it is hard to see how He could have had any unbalanced requiring reasons (the (b 2) line of thinking) – or indeed, for my money, any requiring reasons at all, rather than merely justifying reasons (the (b1) line of thinking) – directing Him with regard to His choice of which world from among this proper part of the array to actualize. There would be justifying reasons in favour of actualizing many of them. Many of these creations (and thus the worlds of which they are a part) instantiate values peculiar to them and God might have acted on the reasons which that fact occasions – in order to add to value-variety (relative to Just God) and/or in order to instantiate those peculiar values.

Of course, it would be compatible with God being in a no-best-world scenario with regard to the infinite array as a whole for there to be clusters of worlds the creation elements of which do have commensurable values and which are thus rankable relative to one another. All that is required for the unrankability of the array is that some of the clusters are not rankable by reference to others. Some have been tempted to say that were there to be clusters of this sort, then God would have requiring reasons to create the highest-ranked world within any cluster if He created any world from among those within a cluster. We do not need to take a position on this topic, for, even if it is granted, if we grant inter-cluster unrankability, God does not have a requiring reason to select from within any one cluster rather than from within another or rather from within no cluster at all (should any world not be clustered with at least one other).

So, if we are considering solely the proper part of the infinite array which is all worlds other than Just God, one may say either of two things.

First, one may take the (b2) approach and say that if A's being higher ranked than B is in fact a requiring reason in favour of actualizing A over B, then A's being a higher ranked world than B cannot be an unbalanced requiring reason for actualizing it (for otherwise, per impossibile, B's being higher than C would be an unbalanced requiring reason to actualize it; and C's being higher than D . . .; ad infinitum). So, each such reason, if it is requiring, must be balanced out by the ‘next’ requiring reason along. Second, one may say that A's being a higher-ranked world than B is merely a justifying reason for actualizing it over B (and B's being better than C is a justifying reason for actualizing it over C; and so on). Thus, the reason-giving force of this better-than relation may be put aside by God without irrationality.³⁰ Either way then, God may actualize any world from this portion of the array, assuming He may choose from this portion and is not compelled to actualize Just God, Just God being a world which we have said we may grant for the sake of argument is lexically better than any other world in the array. And we may assume this now, as we've just seen that even if Just God is best from an impersonal perspective, it is not credible that being best from an impersonal perspective gives God a requiring reason to actualize it. Even if Just God trumps all other worlds in this fashion, that it does so provides at most a justifying reason to actualize it, not a requiring reason.³¹ So, whichever world God actualized from among the whole array (i.e. that including Just God as well as the part of the array with worlds containing creations), He would have had at most unbalanced justifying reason against actualizing it. Thus, whatever He had done, He would not have acted irrationally. Of course, such a view fails to conceive of God as of necessity unsurpassable in the fashion that some pushing the Problem of Evil in the manner of Rowe seek to present it as desirable for the theist to think of God as unsurpassable. However, for myself, I find that merely an additional reason to take the route plotted here. That this move sidesteps this formulation of the Problem of Evil is an additional reason for the theist to make it, not a reason to avoid making it.

Conclusion

In conclusion, there is plenty of reason to think that Spinoza and Leibniz were wrong: traditional theism allows there to be an answer the title question which explains why, contra Spinoza, God may create, but also, contra the Leibnizian, He need not create. That a given action would maximize value in a situation where no-one would be harmed by one's failing to perform it gives one at most a justifying reason for performing that action. And as no mere-possible creature can be harmed by not actually being brought into existence, so, while God may have had a justifying reason to create (e.g. adding to the value-variety of reality), its being merely justifying entails that He was not required to act on it – He could have created nothing, actualizing thereby Just God instead (and perhaps He would have had a justifying reason for doing that – e.g. maintaining value purity). While it is true that if He decided to actualize a world other than Just God (as we know He in fact did, given that there is a creation), the love constraint³² would have demarcated an outer limit to the range of possible creations open to Him, nevertheless there are many – plausibly an infinite number – of possible creations that could have satisfied this constraint. And, presuming as we are in this context that the Problem of Evil has a solution, this – the creation we find ourselves in – was obviously among them. Contra Spinoza, God could create; contra the Leibnizian, He didn't have to do so.