Introduction

I start by arguing that it is difficult to get what Theists may in consistency believe about what the world would have been like were Theism to have been false to come out in such a way that it supports, rather than undermines, the natural theological project. Having argued that this difficulty must, however, be surmountable on pain of absurdity, I argue that it is best overcome by using some of the best-known natural theological arguments to argue for a more general Creative Supernaturalism (using something else (I suggest, the principle that simplicity is a sign of the true) to argue for Theism per se at a later stage). Or rather, I suggest that this is what Theists should do if they wish to believe in the soundness of their own arguments. I conclude by suggesting that, in disputations with Atheists, there is another, more dialectically hopeful, strategy: recast the relevant arguments so that they have some premises which are such that whilst a Theist must say they are false, an Atheist Naturalist must say they are true. The most rationally persuasive case for abandoning Atheistic Naturalism is one that the Theist must think is unsound.

***

Let us consider a few examples of natural theological arguments: the Cosmological Argument; the Design Argument; and the Argument from Religious Experience, as advanced by Richard Swinburne in a number of locations, most famously the second book of his trilogy in the Philosophy of Religion, The Existence of God.Footnote ³ In advancing these arguments, Swinburne enjoins on us the premise that if there had not been a God, then either there wouldn't (or probably wouldn't) have been a universe at all or, even if there were to have been a universe, it would have been very unlikely that it would have been conducive to life in the broad sense that our universe is conducive to life; or then, finally, even if there were to have been a universe and it were to have been conducive to life, and indeed even if it were to have had human life such as ours within it, it would not (or would probably not) have had in it religious experiences such as those we find in the actual world. And thus Swinburne reaches his conclusion: the existence of a universe; its being broadly conducive to life; the religious experience of mankind; and indeed various other things, are evidence that God exists. But why agree with this premise if one is a Theist? More generally, what should the Theist say about what the world would have been like had Theism been false? We obviously need to be clear what Theism amounts to before we may proceed.

In the first book of the aforementioned trilogy, The Coherence of Theism,Footnote ⁴ Swinburne sets out what he takes Theism to be. According to Swinburne, Theism is the thesis that there is a God and by a God is understood a ‘person without a body (i.e. a spirit) who is eternal, free, able to do anything, knows everything, is perfectly good, is the proper object of human worship and obedience, the creator and sustainer of the universe’.Footnote ⁵ Let us then take Theism to be the thesis that this sort of God exists. Swinburne is unusual amongst contemporary Theists in holding that the issue of such a God's existence is not one of metaphysical necessity.Footnote ⁶ But still, the majority of Swinburne's arguments do not rely on the differences between him and his fellow Theists; and Swinburne is rightly taken to be the pre-eminent expositor and defender of Theism in the present day.

So, we may ask, how should a Theist – someone who more or less follows Swinburne in understanding what Theism commits one to and in believing in Theism thus understood – best think about the nearest world in logical space in which Theism is false? The nature of that world should determine what the Theist says about what the world would have been like had Theism been false and thus what he or she may point to in the observable world as evidence that Theism is true. It seems that it has to be plausible on Theism that in the nearest world in logical space in which Theism is false, something is different from the way it observably is in the actual world for this natural theological project to get started. And it's getting it to come out consistent with Theism being true that, had it been false, there would have been (or probably would have been) a difference from what we observe that, I shall argue, is more of a problem than may appear.

It is instructive by way of a preliminary to note that Swinburne takes the view that although it is logically possible for substances to have haecceities – our own souls, according to him, are substances that do – God does not Himself have a haecceity.Footnote ⁷ If we assume, as is plausible, that haecceities are logically possible; that it's logically possible God have one;Footnote ⁸ and that a difference in haecceity is of logical necessity the smallest difference that can obtain between any two substances,Footnote ⁹ then, if we are Theists, we might be tempted by the following thoughts. Given that God does exist, in the nearest world in logical space in which He fails to exist, there is a being qualitatively identical to the actual God, differing from Him solely in having a haecceity. Or, if we suppose that Swinburne is wrong only on the detail of the actual God's not having a haecceity, then, in the nearest world in which the actual God doesn't exist, there exists another God, made different merely by differing in the haecceity that He has from that had by the actual God.Footnote ¹⁰ But these thoughts reveal that such a world is obviously not the world the Theist should have in mind when considering what the world would have been like had Theism been false for the purposes of advancing natural theological arguments that start from the observable world. Such a world does not differ in any observable way from the actual. If we grant God's having a haecceity as logically possible, we can see then that, when searching for the right comparator, the Theist shouldn't be considering a world which differs merely in that the actual God doesn't exist in it and anything consequent on that, but rather should be considering a world in which in addition nothing sufficiently similar to the actual God (to be God) exists; the comparator world needs to be a world where, as it were, there's nothing that does God's job for Him in His absence.

This in itself is not a problem for natural theology starting from the observable world and aiming at Theism, for it is an artefact of the way that we have defined Theism that the actual God's not existing in a world isn't logically sufficient for Theism to be false in that world (if it's logically possible that God have a haecceity). For Theism to be false in that world, it has to be true not just that the actual God doesn't exist there but ‘also’ (an ‘also’ presuming the logical possibility of God's having a haecceity) that nothing sufficiently similar to the actual God to count as the God of Theism exists there either. Though then it isn't in itself a problem for natural theology, it does suggest a problem: in travelling in his or her imagination out from the actual world into logical space, the Theist will leave behind worlds in which Theism is true long before he or she leaves behind worlds in which gods fail to exist altogether; and, as I am about to argue, the first worlds he or she will come to after leaving Theistic territory will have ‘functionally equivalent’ gods in them, functionally equivalent that is with regard to producing (or being as likely to produce) the observable characteristics of the world, the things some of which natural theological arguments need to take as evidence for Theism.

The problem is that, assuming Theism, the nearest world in logical space in which Theism is false should not be thought of as containing nothing godlike at all – that would be overkill, making the putative comparator world differ from the actual world in having a whole host of causal relations voided (all those the actual God is responsible for) and in having in it no supernatural substance such as there is, in the form of God, in the actual world. Rather then, the Theist should think of the nearest world in logical space in which Theism is false as having in it a being that, whilst it fails by the narrowest of margins to be worthy of the name ‘God’, does as much of God's job as is compatible with that failure.Footnote ¹¹ Let's give a name to such a god, ‘God Minus’.

A God Minus is by definition a being with properties that are as similar as possible to those that God actually has whilst nevertheless failing to be sufficiently similar for a world in which he exists to be thereby one in which Theism is true. It seems to me that the guiding thought behind Perfect Being Theology is that which holds the theistic tradition together. It thus seems to me that the closest being to God (who nevertheless fails to be God) is a being who fails by the narrowest of margins to be perfect. It seems to me sufficient to be imperfect that a being be ignorant of one truth, and that may be an entirely insignificant truth – the more insignificant, the narrower the margin of failure. It seems to me that there are an infinite number of maximally insignificant truths and thus that there are infinite ways of being a God Minus. So, for example, a God Minus might fail to know the trillionth number in the decimal expansion of pi, failing to come to any belief about it at all; in all other respects logically compatible with that, such a God Minus would be like God.Footnote ¹² But, be that particular putative example as it may, any plausible filling-out of a God Minus's properties clearly leaves him with the capacity to create and sustain a universe like ours. One can fall a long way short of omnipotence, omniscience, moral perfection, and so forth and yet still be able to do this. And the thing about insignificant differences is that they don't make any difference to the way that a being acts; a God Minus wouldn't have any reason to act differently from God.Footnote ¹³ As a God Minus could create a universe exactly like ours, so the Theist should say – to keep things as similar as possible – that in each of the nearest worlds in which Theism is false, a God Minus exists and creates a universe exactly like ours.

I have been assuming not only that God MinusesFootnote ¹⁴ could act in the manner that God acts with regard to creation, which is true by definition, but also that they would act in that fashion, which is not true by definition and may be disputed. If it's thought that the actual God has on balance more reason to act as He does in creation than not to do so, then of course so will every God Minus and thus every God Minus will act in creation as God acts. And most Theists pursing natural theological projects do have some thought along the relevant lines as a premise, but let's consider an alternative as an example of a possible ‘out’ at this stage. Let's suppose that one thinks that God was entirely indifferent over whether or not to create this universe; it was either this or nothing, but He was 50/50 between those options. On that view, as one moves out into logical space to God-Minus worlds, half of them have their God Minuses not creating anything and the other half have their God Minuses creating this world. This is just one example of how one's thoughts here might affect things. The point to take though is that however much one regards God as raising the probability of some happening, one should regard God Minuses as raising it to that same extent. Whatever the chance was that the actual God (who obviously could have done whatever it is He ended up doing with respect to creation as He did end up doing that thing) would do a thing, that's the chance one should think that that thing will be done by God Minuses in each of the worlds that makes up the ring of God-Minus worlds that surrounds the actual world, i.e. that it would have happened had Theism been false. So, although, for ease of exposition, I tend to speak in this article to the effect that given that God did do what He did, so God Minuses would do the same, that is a bit rough and ready.

Given then that worlds with God Minuses in them are, in observable features, indistinguishable from our world, so again this cannot be what the Theist who hopes to use observations of the actual world as evidence for Theism needs as his/her comparator class of worlds. But it is less clear on what grounds the Theist may object to this being the right comparator class of worlds (than it is in the haecceity case). As already observed, a God Minus is definitionally such that if he exists (but no God exists), Theism is false but the world is in all other respects as close as it can be to the actual world, given that a God Minus manages to do all the things that God actually does, including then creating a universe exactly like the actual universe. Thus how can the Theist avoid having to say that something like this is what would have been the case had Theism been false?

What would be needed in order to avoid this result would be for the Theist to be able to justify excluding from the close-in regions of logical space worlds containing those supernatural beings who fall short of being the God of Theism but who could and would nevertheless stand in for Him causally so as to produce (or as probably produce) a universe that is like ours. I take it as obvious that to narrow the necessary exclusion principle to such a proper subset of supernatural beings would be implausibly ad hoc, so I take it without argument that one would be best advised, if travelling this route, to try to justify excluding all supernatural beings from this zone. That having been done, one might indeed then say that the nearest world in which Theism is false has nothing supernatural in it at all; hence no ‘stand ins’; hence nothing replicating the phenomena which we may thus rightly take to be evidence in the actual world that God exists. But such a policy of exclusion is difficult to justify. Remember, according to Swinburne, the actual God's existing is a contingent fact. If so, surely the actual world must be surrounded in logical space by a ring of God-Minus worlds. If we deviate from Swinburne on the issue of the actual God's necessity, it might seem that we have more hope of consistently saying otherwise. But, I shall now argue, any such hope is forlorn.

Suppose we think that the issue of whether or not there's a God or a god is one of metaphysical necessity. We're Theists, so we think that there is a God, indeed we think that it's metaphysically necessary that there is. And suppose we also think that, if God weren't to have existed, then that would have to have been because there was something deeply impossible about there being anything godlike at all. Thus we might say that in the nearest worlds in which God doesn't exist, the necessity of His existence is ‘swapped out’ for the necessity of there being nothing godlike at all. So, in the nearest worlds in which Theism is false, there aren't any God Minuses, Platonic Demiurges, gods, angels, or so on. That is a consistent view and it would indeed exclude the problematic ‘stand ins’ from the close-in zone. But why make this ‘swapping out’ claim? If God actually exists of metaphysical necessity, then surely in the nearest world in which this metaphysical necessity is voided, it isn't ‘swapped out’ for another metaphysical necessity, but rather it just becomes contingent whether or not such a God exists (as Swinburne in fact would say it already is in the actual world).Footnote ¹⁵ We should as Theists say that near worlds are God-Minus worlds, in one of which perhaps the only respect in which its God Minus differs from the actual God is in its not having the metaphysical necessity that the actual God has. Anyway, we shouldn't say that such ‘swapping out’ occurs thereby evacuating the close-in zone of God substitutes.Footnote ¹⁶

One thing that the Theist might do at this stage is again to draw his or her inspiration from the work of Richard Swinburne.Footnote ¹⁷ Rather than using natural theological arguments in support of Theism exclusively, which then requires him or her to consider what would have been the case had Theism been false and generates the ‘A God Minus would have stood in’ answer, he or she might instead use them in favour of what we might call Creative Supernaturalism – the thesis that there is some supernatural creative force or agent (perhaps the God of Theism; perhaps a God Minus; perhaps a Platonic Demiurge or other god; perhaps some impersonal karmic creative energy). He or she might then use the principle of simplicity as a sign of the true at a later stage of his or her natural theological project to sift out the supernaturalist hypotheses that are thus supported, leaving the Theistic hypothesis on top only at this later stage. In seeking to use something observable about the actual world as evidence in favour of Creative Supernaturalism, the Theist obviously needs for his or her comparator to go out in logical space well beyond God-Minus worlds; the nearest world in logical space in which Creative Supernaturalism is false is – I have been arguing – much further out than the nearest world in which Theism is false. So, now there is no need to exclude the stand-ins from the close-in zone; one instead merely skips over that zone in the search for the comparator relevant to the more general thesis that one now wants to argue for. As well as having the advantage of avoiding the ‘A God Minus would have stood in’ problem, this manoeuvre is also attractive for at least the following two reasons. (a) It enables one to concede as true, but irrelevant, a slew of criticisms that are often levelled at natural theological arguments to the effect that even if they show some supernatural cause, designer, or what have you, they fall a long way short of exclusively confirming Theism. (b) It ‘meets Atheists where they are’ in the sense that most Atheists nowadays deny Theism not because they are happy to concede that something like God Minus exists, but not to concede that God exists; they deny Theism because they think that there is nothing supernatural at all.

Let's suppose then that we consider the Theist who seeks to use his or her natural theological arguments primarily to argue for Creative Supernaturalism. This Theist plans to use something else, such as the principle of simplicity being a sign of the true, later to get the arguments pointing to Theism more particularly. What should such a Theist say about the nearest world in which Creative Supernaturalism is false? Again, he or she must – if his or her natural theological project is to succeed – find himself or herself able to say of this world that it is in some way different from the world as we observe it.

We're now out beyond the realms of anything supernatural that's going to be causally relevant; and we're looking for a world that, given that, is as similar as it can be to the actual world, in which of course there's a universe like ours. Given this, it might seem that a good case can be made for saying that the Theist should think of the first world he or she comes to as containing a universe just like ours, just like ours except for its aetiological properties, of course. In this universe, the Big Bang; the sustaining of the laws of nature; and anything that's happened in the actual universe as a result of miraculous intervention, have happened as a result of chance. The case for this world's being closest would be premised on the qualitative similarity between its universe and the actual. The reason for calling it closest would be that, given that one has knocked out God and anything supernatural in travelling out from the actual world (in which Theism is true) until reaching the nearest world in which there is no supernatural causally-relevant being, one then does best to secure similarity with the actual world by having anything God is in fact causally responsible for – plausibly the Big Bang; the sustaining of the laws of nature; apparent miracles – happen anyway (otherwise there'd be needless difference), but happen without Him. We're beyond the help of any substitute god. Hence, the appeal to chance (i.e. uncausedness, i.e. nothing) to fill God's role. So, such an argument would conclude, the Theist should say that in the nearest world in which there's nothing creative and supernatural, there's a universe just like ours. But that again is not going to be a suitable comparator for the project of natural theology: the universe in this world is again exactly like ours in its observable qualities.

But at this stage one will rightly think that something must have gone wrong on pain of absurdity; if nothing had gone wrong, then no observable feature of the universe could, even in principle, be taken by the Theist to be evidence that Theism or even Creative Supernaturalism is true, but that would be absurd. We may be in a universe in which the evidence for Theism is not as good as Swinburne thinks; we may be in a universe in which the evidence for Theism is in fact very poor; but we cannot be in a universe in which the evidence for Theism is of necessity non-existent. So, what's gone wrong? And, following on from that, how do we put it right?

***

Let's go at this ‘the other way around’; let's look at what ordering in logical space worlds would need to be given for a natural theological project such as Swinburne's to succeed; and let's continue to focus on using these arguments merely to support Creative Supernaturalism in the first place. So, we're looking into logical space at the nearest worlds which lie beyond the existence not only of any God of the Theistic sort, but also of any God Minus; any Platonic Demiurge; any pantheon of gods; and so on. And we're looking to order these worlds in proximity to the actual world in whatever manner it is that would best serve the interests of such a natural theological project. Once we've established what that ordering is, we can then go on to see what the Theist might be able to do to justify its being the right ordering (on grounds other than its being the ordering that makes Theism more rationally supportable).

One point we haven't yet considered is the probable cumulative-case nature of any Swinburne-inspired natural theological project. Not all natural theological projects distribute their argumentative weight around in the same manner as Swinburne's of course, but most do say that the case for Theism is a cumulative one, built on more than one natural theological argument. So, let's start by making it a constraint on the ordering we're looking for that it allow for this. Not all orderings which would support individual natural theological arguments would support them as parts of a cumulative case.

So, for example, suppose that I say that if there hadn't been anything creative and supernatural, then there wouldn't have been anything natural. That being so, I say that the first world we come to in logical space, once we leave creative supernaturalist territory behind, has no universe at all in it. And I may thus take the fact that there is actually a universe as supportive of Theism (through being supportive of the more general thesis of Creative Supernaturalism). It will also of course be true that if there hadn't been anything supernatural, then there wouldn't have been a universe fine-tuned for life (in the sense that we may accept for the purposes of this discussion that our universe is fine-tuned for life). But, at this stage, that second thing will only be coming out as true for me because I'm thinking that if there hadn't been anything supernatural, there wouldn't have been a universe at all. So, if I want to think of the fine-tuning version of the Design Argument as adding something to the case for Theism above and beyond what is added to it by the Cosmological Argument, I'll need it to be that, even after we've crossed into the territory in which there is nothing supernatural, we can still, eventually, get to worlds in which there are universes. I'll need it to be that the first ‘batch’ of such worlds we get to contain universes with the same natural laws as our own, but most of them have different (and non-life-conducive) constants and/or boundary conditions. The thing I can't do is hold that the truth that there's never anything natural without something supernatural continues into logical space forever; I can't do that if I'm to think of the fine-tuning version of the Design Argument as adding to the case for Theism beyond the point to which the Cosmological Argument has already added to it. So as to add yet more arguments for Theism to one's cumulative case, one can put other universe-containing worlds into logical space in the right order later. As long as one hasn't blocked them by saying earlier things such as that universes could not (of logical necessity) exist without supernatural beings; fine-tuning could not (of logical necessity) have happened by chance; and so on, all will be well. So, further out still, one might say, are worlds with universes that are fine-tuned for life and even have humans in them, but, in the nearest such worlds, religious experiences do not exist (or are rarer, or do not have the character that they actually do), thus supporting the argument from religious experience and supporting it in a manner compatible with running it in addition to running the Cosmological Argument and the Design Argument in its fine-tuning mode. Only very far out indeed, one may say, does one get to the world in which there is simply an exact simulacrum of our universe produced by chance, the world in which Atheistic Naturalism is true. Thus Theism may be supported in a cumulative-case fashion by the arguments mentioned and potentially by others.Footnote ¹⁸

So then, there is a logically consistent ordering of worlds that would support the natural theological project. Indeed, I hope I've done enough to show that there are several, several even with the constraint that the project be a cumulative-case one. Suppose, for example, that one doesn't feel greatly taken the Cosmological Argument. One may then say that in the nearest world in which Creative Supernaturalism is false, there's still a universe with the same natural laws, but, in considering it, we're straight out to that ‘batch’ of universes, as I put it earlier, amongst which fine-tuned-for-life ones are a very small minority. And one may thus take up the story above at that stage.

What's stopped us so far in getting to one of the orderings that would seem to be needed to make a natural theological project work is that it's seemed that one can appeal to chance to make the nearest world in which Creative Supernaturalism is false exactly like ours (other than in its aetiological properties, of course). And of course one can state that chance could stand in once every creative supernatural being drops out; and thus one can state that it does in fact stand in in some relatively-close world, a world that's ipso facto relatively closer than an otherwise similar one in which it doesn't stand in. But the chances of chance producing a particular effect stay the same as one leaves behind creative supernaturalist worlds and that fact allows for some of what we observe nevertheless to have evidential value in favour of Creative Supernaturalism, in virtue of the general principle that in considering the evidential value of a given thing one needs to ‘bracket off’ that thing when specifying which worlds are the right comparators. So, for example, as one moves out past a certain point beyond Creative Supernaturalist worlds, one gets to that batch of universes with the same natural laws as our own and, if counting solely in virtue of qualitative similarity, one would have to count as closer-in amongst this batch that world which has in it a universe in which the constants and boundary conditions take by chance the values that they take in the actual world, but one must bracket off this as a reason for counting this particular member of that batch as closer when considering what evidential value our universe's having the actual constants and boundary conditions it does have has against the chance hypothesis being true. Thus, one may in consistency say that the fact that in the actual world the constants and boundary conditions take the values that they do is evidence that it wasn't chance that produced them – worlds in which chance produces them form only a tiny minority of worlds with the same natural laws. Thus Theism (plus of course God-Minus-isms and a variety of other supernaturalist hypotheses) are supported by the fine-tuning version of the Design Argument. (One can only say this on various other assumptions of course, but investigating those is beside my current purpose.) We're out in logical space beyond the help of any creative supernatural agency; we're considering the first batch of universes we come to with laws like ours and thus with just chance (i.e. nothing) selecting the values taken by various constants and boundary conditions; we're noticing that the proportion of them in which the relevant factors are selected so as to make their universes life-conducive is very small; were there nothing creative and supernatural, chance would have selected what values these took; but the chances of chance selecting life-conducive values is very small; thus their actually having life-conducive values is evidence that it's not chance but rather something creative and supernatural that's selected them, that the actual world is closer in than this batch of chance worlds; it's probably a world in the creative supernaturalist zone. (Or rather, on these other assumptions, it is.)