Naturalism, classical theism, and first causes

Abstract

Enric F. Gel has recently argued that classical theism enjoys a significant advantage over Graham Oppy's naturalism. According to Gel, classical theism – unlike Oppy's naturalism – satisfactorily answers two questions: first, how many first causes are there, and second, why is it that number rather than another? In this article, I reply to Gel's argument for classical theism's advantage over Oppy's naturalism. I also draw out wider implications of my investigation for the gap problem and Christian doctrine along the way.

Introduction

Classical theism is unique among models of God in that it affirms four core theses: divine simplicity, timelessness, immutability, and impassibility. According to the Doctrine of Divine Simplicity (DDS), God has no physical, metaphysical, or logical parts. Traditionally, this is taken to entail that in God there is no distinction between essence and existence, subject and accidents, individual and essence, individual and properties, act and potency, and agent and the agent's actions.¹ Instead, God is God's essence. God is his existence. God lacks accidents in the sense of intrinsic contingent features. God lacks properties in the sense of exemplifiables distinct from himself. God is purely actual, lacking any potential for change or cross-world variance. And God is identical to each of his acts. Such are the commitments of the traditional DDS.

It will be useful to define ‘part’ before proceeding. At least in the context of DDS, a part of God would be any positive ontological item intrinsic to but distinct from God.² Thus, according to DDS, ‘[a]nything intrinsic to God is identical to God’ (Fakhri (2021), 10). Vallicella (2019) puts it equivalently: ‘God is ontologically simple … there is nothing intrinsic to God that is distinct from God.’ Other scholars follow suit.³

While much more can be said, that suffices for present purposes as a characterization of classical theism – or, at least, the version of classical theism with which I am here concerned. Let's now consider a view on the opposite end of the metaphysical spectrum: Oppy's naturalism.

Naturalism is variously characterized. Our focus here is Oppy's naturalism, which has been particularly influential in philosophy of religion.⁴ In its most recent and well-developed articulation (Pearce and Oppy (2022), ch. 2), Oppy's naturalism makes several core claims. Five important and characteristic such claims are:

1. (1) Natural reality exhausts causal reality: Every causal entity and causal property is natural, where natural causal entities and properties are those ‘recognized in ideal, completed, true science’ (ibid., 103).

2. (2) Mindedness is late and local: Minded beings are either relatively recently evolved organisms or products of such organisms.

3. (3) Nothing is divine: Nothing causal is divine, sacred, or worship-worthy.

4. (4) Shared history: Necessarily, any possible way causal reality could be shares an initial history (i.e. an initial world segment) with actual causal reality. In other words, for any possible causal reality, its history at some point coincides with actual history. Oppy thus affirms that ‘every possible world shares some initial history with the actual world’, which we can call natural causal reality's ‘initial state’ or ‘initial singularity’ (ibid., 222).

5. (5) Chance divergence: Necessarily, the only way that possible causal histories diverge or branch from actual causal history is through objectively chancy events.

Having characterized the two world-views being compared, let's get clear on my target. In a recent article, Enric F. Gel argues that ‘precisely because of its picture of God as a purely actual first cause of things’, classical theism enjoys ‘a significant advantage over’ Oppy's naturalism as a theory of the first cause (Gel (forthcoming), 2).⁵ Gel offers two argumentative paths in favour of this claim: ‘one where causal finitism is granted’ and another ‘where a foundational layer of reality is granted’ (ibid.). Gel's arguments advance debates about classical theism, the gap problem, and the comparison of rival theories in philosophy of religion. Ultimately, however, I do not find the arguments convincing. My task in this article is to explain why.

I begin by examining Gel's first argumentative path from causal finitism. Then I examine Gel's second argumentative path from a foundational layer of reality. I next examine two arguments that non-classical theisms enjoy the same advantage that – according to Gel – accrues to classical theism over Oppy's naturalism. Along the way, I also draw out wider implications of my investigation for both the gap problem and the Christian doctrine of Trinitarianism.

Gel's first argumentative path

Methodology

Causal finitism says that infinite causal histories are metaphysically impossible.⁶ If we rule out circles of causation, causal finitism – together with the relatively innocuous claim that something is a cause – straightforwardly entails that there is at least one uncaused first cause. At this juncture, Gel raises the following questions: first, how many first causes are there? And second, why are there exactly that many?

According to Gel, there are two possible answers to the second question: either (a) ‘it's a brute fact with no explanation’, or (b) ‘it's a necessary fact that could not have been otherwise’ (Gel (forthcoming), 3). It's not clear that this is right, however. Gel is importing both explanatory and modal notions into the answers to the second question. Given this, there are four permutations of possible answers: (i) it is a contingent, unexplained fact; (ii) it is a contingent, (non-causally) explained fact; (iii) it is a necessary, unexplained fact; and (iv) it is a necessary, (non-causally) explained fact.⁷ Ideally, we want our theory to answer (ii) or (iv), as these alone offer a gain in theoretical explanatoriness; moreover, we want the resultant gain in explanatoriness not to be outweighed by a cost in theoretical complexity.

Regarding the first question, Gel writes that ‘due to Ockham's Razor, out of two equally explanatorily efficacious accounts, we should prefer that which posits fewer entities and is, hence, simpler’ (ibid., 3). But Ockham's Razor doesn't merely advise us to posit fewer entities. It also advises us to posit fewer kinds of entities. (As Da Vee (2020) articulates it, Ockham's Razor enjoins us ‘to not multiple entities (or kinds of entities) without necessity’ (ibid., 3679, emphasis added).) The former relates to a theory's quantitative simplicity, whereas the latter relates to its qualitative or categorical simplicity. Moreover, given that Gel is critiquing Oppy, we must also take into account the three elements of simplicity Oppy employs to assess theories:

1. (1) Ontological simplicity: All else being equal, a world-view committed to fewer things and fewer kinds of things is superior to one committed to more things and more kinds of things (Pearce and Oppy (2022), 115).

2. (2) Ideological simplicity: All else being equal, a world-view employing fewer undefined primitive expressions is superior to one employing more undefined primitive expressions (ibid.).

3. (3) Theoretical simplicity: All else being equal, a world-view committed to fewer and less complicated fundamental principles is superior to one committed to more and more complicated fundamental principles (ibid.).⁸

With all these methodological points covered, we can now turn to Gel's main argument. I'll begin by examining the role of simplicity therein.

Simplicity

Gel nicely outlines his overall argument as follows:

[T]he advantage I want to ascribe to the pure-act theist's candidate for a First Cause is that, by adopting it, we are able to (1) get an answer to the question of how many First Causes there are; (2) explain why there is that number rather than another (namely, as we'll see below, because there can be no other) and (3) such a number could not possibly be lower. (Gel (forthcoming), 3)

The classical theist's answer to the question in (1) is obviously one. According to Gel, this also establishes (3), since one is the ‘simplest possible answer’ (ibid.). But this is not so clear given the abovementioned methodological points. For simplicity isn't merely a matter of the fewest number of entities posited (i.e. the quantitative simplicity element of Ockham's Razor and Oppy's ontological simplicity); it is also a matter of (i) the fewest kinds of entities posited (i.e. the qualitative simplicity component of Ockham's Razor and Oppy's ontological simplicity), (ii) fewer undefined primitive expressions (i.e. Oppy's ideological simplicity), and (iii) fewer and less complicated fundamental principles (i.e. Oppy's theoretical simplicity).

And here's the rub: it is not at all clear that classical theism wins out on simplicity considering (i)–(iii) (in addition to what Gel does consider – namely, only the quantitative simplicity element of Ockham's Razor). If anything, Oppy's naturalism seems to win out.

Let's proceed, then, through each aspect of simplicity. For each aspect, I will examine both (a) the simplicity of the respective theories’ postulates about the first cause (the number thereof, the nature thereof, etc.) and (b) the overall simplicity of the respective theories. I examine (a) because Gel himself seems to focus primarily thereon. (Gel, after all, is here comparing the simplicity of the theories’ answers to the question in (1), which is only about what the theories say concerning the first cause.) But it's crucial to consider (b). For even if one theory has a theoretically simpler view of the first cause, it may only secure this at the cost of a more complex overall theory. Indeed, I find that (b) is what really matters; for if T is simpler than T* in (b)'s sense but more complex in (a)'s sense, we should clearly prefer T over T*. When we compare theories, we don't compare their simplicity with respect to some small domain; we compare their simplicity full-stop, and we settle on the overall simpler theory (assuming explanatory parity). Thus, I think far more weight should be placed on (b). But because Gel seems to focus on (a), I will examine it as well.

The quantitative simplicity element of Ockham's Razor

First, in terms of how many first causes are posited (a), the answer for both classical theism and Oppy's naturalism is one. For Oppy, the first cause is the initial singularity, and there is only one such thing. As Oppy writes in his debate with Kenny Pearce, ‘Kenny says there is one initial thing: God. I say there is one initial thing, which I will call “the initial singularity”’ (Pearce and Oppy (2022), 271).⁹ Second, in terms of overall quantitative simplicity (b), Oppy's naturalism seems clearly superior to classical theism. For Oppy's entities are a proper subset of the classical theist's. The classical theist, no less than Oppy, thinks that the natural world exists. For Oppy, that's all. But for the classical theist, that's not all; there's also a purely actual God. Oppy's naturalism therefore seems to enjoy an edge in terms of the quantitative simplicity element of Ockham's Razor.

The qualitative simplicity element of Ockham's Razor

First, in terms of how many kinds of first cause are posited (a), it isn't clear how to assess classical theism and Oppy's naturalism. As explained above, both theories posit only one (causally and/or temporally) initial thing; we might conclude from this that they therefore each posit only one kind of initial thing. What makes this difficult is that one thing can be a member of multiple kinds. Oppy's initial singularity is a member of many kinds: natural things, necessary things, physical things, impersonal things, and so on. The classical theistic God, too, is a member of many kinds: supernatural things, necessary things, non-physical things, personal things, and so on.¹⁰ Thus, in terms of the theories’ answer to how many kinds of first cause there are, there is no clear winner. But there is, I think, a clear winner when it comes to overall qualitative simplicity (b). For classical theism introduces whole hosts of kinds into one's ontology in addition to those Oppy's naturalism admits. For Oppy, everything uniformly falls into the kinds natural, physical, limited, and so on. But classical theism not only recognizes such kinds but also postulates additional kinds: supernatural, non-physical, unlimited, and so on. Once again, because both Oppy and the classical theist are committed to the natural world (and all the kinds entailed therefrom) while the classical theist believes in God in addition, the kinds postulated by Oppy's naturalism are a proper subset of those of classical theism. Once more, then, Oppy's theory seems to enjoy an edge in terms of the quantitative simplicity element of Ockham's Razor.

Ideological simplicity

First, in terms of how many primitive undefined expressions are used to describe the first cause (a), there is once more no clear winner. To determine a winner, we would need a fleshed-out characterization of the initial state (from Oppy) and God (from the classical theist). We would then need to define as many of the terms as we can and, upon reaching analytical bedrock, count the number of primitive undefined expressions contained in the respective characterizations. Obviously enough, this is a task that lies far beyond the scope of the present article, and it's not clear which theory would edge out. What, then, about overall ideological simplicity (b)? Here again, Oppy's naturalism seems to win out. For the classical theist, like the naturalist, is committed to the natural world and, with it, all the primitive undefined vocabulary needed to describe it.¹¹ But the classical theist needs additional primitive undefined vocabulary to describe God. This is especially true for Thomistic classical theism – something to which Gel frequently appeals. For such a view makes considerable use of analogical expressions to characterize God. As Oppy points out in Pearce and Oppy (2022, 117–119 and 272–274), such expressions are new, primitive, undefined expressions. In short, the primitive undefined expressions used in Oppy's naturalism are a proper subset of those used in classical theism. Oppy's naturalism once more edges out.

Theoretical simplicity

For the same reason as the previous paragraph, there doesn't seem to be a clear winner when we restrict our focus to the theories’ principles concerning the first cause (a). But once more, Oppy's naturalism seems to edge out classical theism when it comes to overall theoretical simplicity. For Oppy, ‘we should treat all claims [principles] as fundamental except those that have been explicitly shown to be logical consequences of other things to which we are committed’ (ibid., 268). Now consider what an ideal world-view would say about the natural universe. It would provide a complete specification of the initial state of the universe, the laws of nature governing its evolution over time, and the outcomes of every chance event. By Oppy's lights, ‘logical consequence affords no possibilities to compress this information’, and so on Oppy's account, ‘we should suppose that it is all fundamental’ (ibid.). But nothing about God's nature logically entails facts about the initial state of the universe, the laws of nature, or the outcomes of chance events. Classical theism, then – no less than naturalism – must take such specifications as fundamental. But such specifications are all Oppy's naturalism needs, since Oppy's naturalism commits only to the natural universe. The classical theist, however, commits both to the natural universe and to God, and hence the classical theist will accrue theoretical complexity costs deriving from additional specifications about God.

I do not claim that the above considerations are decisive or insuperable. By my lights, they at least cast doubt on Gel's appeal to simplicity. Having covered simplicity, let's now consider explanatory power.

Explaining the number of first causes

According to Gel, classical theism enjoys an explanatory advantage over Oppy's naturalism when it comes to explaining exactly how many first causes there are. For the classical theist, there is exactly one purely actual first cause ‘because there can only be one purely actual thing’, that is, ‘because a purely actual thing is not susceptible of multiplication’ (Gel (forthcoming), 3). My criticism is that the arguments Gel offers for the impossibility of multiple purely actual beings do not succeed, and hence that Gel has not succeeded in showing classical theism's explanatory advantage over Oppy's naturalism.

Gel's first argument for the (necessary) uniqueness of a purely actual thing runs as follows. A purely actual thing's

essence would just be pure esse itself. But then, such a thing could not be multipliable, because it could not be subjected to any differentiating feature, as a genus (animal) is multiplied in its species (human) by the addition of a specific difference (rationality) or a species (human) in its individuals (Peter, Mary, and James) by the addition of matter. There is nothing outside pure being that could act, with respect to it, as a differentiating feature, as the specific difference rationality is outside the genus animal or as matter is outside form, because ‘outside’ pure being there is only non-being, and non-being is nothing. So pure being could not be differentiated, as pure being, into multiple instances of itself, such as pure being A, pure being B, pure being C, and so on. Hence, a purely actual reality that was pure being itself – and such is the classical theist's picture of God – would have to be unique, out of metaphysical necessity. (ibid.)

The argument, in short, runs roughly as follows:

1. (1) For there to be more than one thing that is pure esse, there would have to be some feature(s) that differentiates each from the other(s).

2. (2) But nothing that is pure esse could have such a differentiating feature.

3. (3) So, there cannot be more than one thing that is pure esse. (1, 2)

4. (4) But whatever is purely actual is pure esse.

5. (5) So, there cannot be more than one purely actual thing. (3, 4)

We should note, first, that Gel offers no justification for premise (1). Gel simply asserts (albeit implicitly) that multipliability requires the possession of some differentiating feature. (Gel's claim that ‘such a thing could not be multipliable, because it could not be subjected to any differentiating feature’ is true only if multipliability requires some differentiating feature.)

We should also note, second, that premise (1) amounts to the deeply controversial Identity of Indiscernibles (IoI).¹²

$$Identity\; of\; Indiscernibles\; ( {IoI} ) \colon\; \forall x\forall y( \forall F( Fx\leftrightarrow Fy) \to x = y) $$

IoI states that for any objects x and y, if for every feature F it is true that x is F iff y is F, then x is identical to y. In other words, if x and y share all and only the same features in common, then x is identical to y. Contraposing this, we get: if x is distinct from y, then there is some feature that one has that the other lacks. And this, in turn, is simply to say that multipliability requires some differentiating feature.

Given the controversial nature of IoI, it seems dialectically ill-suited simply to assume it. The principal motivation behind it seems to be explicability: if x and y share all and only the same features, in virtue of what are they distinct or individuated from one another? Given the nature of the case, we cannot pinpoint something had by one and not the other. And in that case, their individuation would seem to be primitive or brute.

It's not clear, though, whether we should be sanguine about this motivation. Why can't individuation or distinctness simply be primitive? In that case, there need not be some feature that grounds things’ distinction. At the very least, we need some positive argument in favour of the principle, since the onus of justification in this context is on Gel to demonstrate necessary uniqueness. Indeed, there seems to be a prima facie plausible argument that individuation or distinctness must ultimately be primitive. For we can equally ask: in virtue of what are those individuating features of x and y individuated? If they're not individuated by anything, then we have primitive individuation, which is precisely what IoI sought to avoid. If they have some further differentiating features, then we're off on a vicious regress. For we can further ask, of those features, in virtue of what are they individuated? And so on ad infinitum.¹³ It seems, then, that we must ultimately bottom out in primitive individuation.

What about premise (2)? Let's first consider Gel's proffered justification thereof: ‘There is nothing outside pure being that could act, with respect to it, as a differentiating feature . . . because “outside” pure being there is only non-being, and non-being is nothing’ (ibid.). I'm not sure what to make of this, mainly because I don't understand what ‘outside’ means in this context. It certainly can't mean ‘distinct from’, since there most definitely are things distinct from pure being. But if it doesn't mean distinction, I struggle to see what it could mean.¹⁴

Second, there seem to be several plausible candidates, at least in principle, for features that differentiate among beings of pure esse. Consider, first, that most Thomistic classical theists think that being pure esse is compatible with being Trinitarian (i.e. existing as three persons). But if that's so, surely being pure esse is also compatible with being (say) Unitarian (i.e. existing as one person). It is not as though Jews and Muslims are prevented from affirming the traditional DDS (and, with it, God's being identical to his existence) by dint of their Unitarianism. It would also seem intolerably ad hoc and inexplicable if Trinitarianism but not Unitarianism (or Binitarianism, or etc.) was compatible with God's being pure esse. If all this is correct, then we have on our hands a clear candidate for a differentiating feature among purely actual beings of pure esse: the number of persons in which they exist. In principle, one being of pure esse could be Unitarian; another could be Binitarian; still another could be Trinitarian; and so on. I do not claim these are genuine metaphysical possibilities; my point is simply that the argument that there cannot in principle be something that differentiates beings of pure esse fails.

Consider, second, the distinction between being identical to one's own act of existence and being identical to existence simpliciter or existence as such. Thomistic metaphysics already admits that there are (roughly speaking) different acts of existence. My act of existence, for instance, is not the same as God's act of existence (with which God is identical). If they were the same, then I would be a composite of essence and . . . God! God would be an internal principle that composes me as an essence-existence composite. This is clearly contrary to classical theism. As Kerr (2015) notes, the esse commune that creatures receive is ‘somewhat lesser’ (153) and ‘somewhat other’ (154) than the esse tantum of God.

God, then, is identical not to the existence of you or me or trees; he is identical to his own act of existence. But in that case, it's not clear why there cannot be two things which are identical to their acts of existence. They could presumably each be identical to their own respective acts of existence, which are different from one another. These acts of existence, moreover, could presumably be primitively distinct (which, as I argued in my discussion of IoI, seems possible).¹⁵

Finally, I would argue that Gel's argument is incompatible with Trinitarianism, for the exact same reasons for thinking there could only be one purely actual being of pure esse would equally motivate thinking that there could only be one purely actual divine person of pure esse.¹⁶ Just mount the following parody argument:

1. (1) For there to be more than one divine person that is pure esse, there would have to be some feature that differentiates each from the other(s).

2. (2) But nothing that is pure esse could have such a differentiating feature.

3. (3) So, there cannot be more than one divine person that is pure esse. (1, 2)

4. (4) Anything divine is pure esse. (Classical theism)

5. (5) Any divine person is divine.

6. (6) So, any divine person is pure esse. (4, 5)

7. (7) So, there cannot be more than one divine person. (3, 6)

In summary, Gel's argument, if successful, would equally vindicate the following argument: for there to be more than one divine person, there would have to be some feature that differentiates each from the others; but anything that is pure esse could not be subject to such a differentiating feature; hence, at least one such divine person would not be pure esse; but anything divine, per classical theism, is pure esse; hence, at least one such divine person would not be divine after all; and that's absurd. Hence, there cannot be more than one divine person.

Let's now consider Gel's second argument (or, at least, variant of the first argument) for uniqueness:

If a purely actual thing has to be absolutely simple or noncomposite (and plausibly it has to, given that any composite is in potency with respect to its parts, or is a mixture of act and potency), then for there to be two absolutely simple beings, A and B, A would have to exhibit feature X that B lacked, in order for them to be differentiated. But then A would not be an absolutely simple being, because it would have parts, contrary to the hypothesis – A would be what B is plus feature X. So, once more, there could not be any feature that distinguished two hypothetical absolutely simple things, without making it the case that one of them or both were not, after all, absolutely simple. (Gel (forthcoming), 4)

It's not clear why any composite is ‘in potency with respect to its parts’, and Gel simply asserts as much without justification.¹⁷ But set that aside. I think it's clear, now, that this second argument suffers from essentially the same problems as the first. It presupposes (without justification) IoI; IoI is deeply controversial, and moreover primitive individuation seems ultimately unavoidable; there are modes of differentiation entirely compatible with DDS; and the argument seems incompatible with Trinitarianism.

To draw out the last point, consider a parody argument:

On classical theism, whatever is divine is purely actual. And divine persons are divine. Hence, on classical theism, divine persons are purely actual. But whatever is purely actual is absolutely simple or noncomposite. Hence, on classical theism, divine persons are absolutely simple or noncomposite. Now, for there to be two absolutely simple divine persons, A and B, A would have to exhibit feature X that B lacked, in order for them to be differentiated. But then A would not be an absolutely simple divine person, because it would have parts, contrary to the hypothesis – A would be what B is plus feature X. So, once more, there could not be any feature that distinguished two hypothetical absolutely simple divine persons, without making it the case that one of them or both were not, after all, absolutely simple. Hence, there can only in principle be one divine person.

The same principles and inferential steps are operative in each argument. I don't think one can accept the original argument, then, without also accepting the parody.

There is yet another problem with Gel's argument. In particular, it does not follow that A would be complex merely from the fact that it has a feature that B lacks. Perhaps A has (and is identical to) feature X, whereas B has (and is identical to) feature Y, where X is distinct from Y. In this case, it is simply false, pace Gel, that A would be ‘what B is plus feature X’. In the case at hand, A is not a composite of everything B has plus feature X; A is simply feature X. And B is simply feature Y. And X is not Y. None of this requires A to have everything B has with X superadded. A is X, pure and simple. B is Y, pure and simple. And X is not Y; X and Y are primitively distinct pure esses. Thus, not only does Gel's second argument inherit the problems of the first, but it also faces a unique problem of its own.

According to Gel, what follows from his arguments is that ‘Pure-act theism . . . provides us with an explanation of why there has to be that number [of first causes] instead of any other – namely, because there could in principle be no more than one purely actual thing, so it is a necessary feature of a purely actual thing that it be unique’ (ibid.). I have argued, however, that Gel's arguments establish no such thing. For all Gel has said, classical theism does not explain why there is exactly one first cause.

Gel's second argumentative path: foundational layer

Let's now consider Gel's second argumentative path from a necessarily existent foundational layer of reality. Gel begins by supposing that there is a necessary foundational or fundamental layer of reality that explains (perhaps by sustaining causation, perhaps by grounding, perhaps by constitution, etc.) non-foundational, non-fundamental things.¹⁸ Gel then asks the same two questions as before: first, how many foundational entities are in this foundational layer, and second, why are there exactly that many?

The dialectic here then plays out much as before. Classical theism clearly answers the first question: one. And Gel relies on his arguments for uniqueness to establish that classical theism offers a satisfying explanation in response to the second question. I have already criticized Gel's arguments for uniqueness, however. As I argued, Gel has not shown classical theism's explanatory superiority on this front.

I do, however, want to address some distinctive things Gel says in his second argumentative path, beginning with the following: ‘Classical theism, then, appears both simpler and more fruitful than Oppy's naturalism, since it can answer more questions about the First Cause with fewer principles (namely, with just one purely actual thing)’ (Gel (forthcoming), 6). This, I think, is mistaken. Let's suppose classical theism answers more questions about the first cause with fewer principles. It clearly doesn't follow from this that classical theism is simpler than naturalism tout court.¹⁹ We don't judge theories by how simple they are only with respect to some small domain; we judge them by their overall simplicity. And once we recognize this, I think it's plausible that Oppy's naturalism is simpler than classical theism. Once again, the classical theist (no less than the naturalist) is committed to the natural world, including whatever fundamental physical things there are (whether they be superstrings, quarks, mereological simples, quantum fields, the universal wavefunction, matter-energy, or whatever). For Oppy's naturalism, that's all. But for classical theism, that's not all. There is, in addition, a transcendent, purely actual, absolutely simple God. This brings with it new vocabulary, new principles, and new ontological commitments that the naturalist does without.

Finally, I want to address Gel's consideration of a naturalist who accepts a purely actual foundation. According to Gel, ‘Both immutability and immateriality . . . seem unavoidable and straightforward once the pure actuality of the First Cause is embraced, since change is essentially the passage from potency to act and every material thing is both mutable and potential in many ways’ (ibid.). Unfortunately, Gel simply asserts that everything material is changeable and potential in many ways. Perhaps Gel takes it to be self-evident. To me, however, it is by no means self-evident. Consider atemporal wavefunction monism. According to this view, there exists a fundamental, physical, non-spatiotemporal entity: the universal wavefunction. This is a perfectly respectable view that has seen a blossoming of interest in philosophy of physics.²⁰ If we understand ‘material’ and ‘physical’ to be synonymous, then it simply follows that there are perfectly respectable views on which there is a fundamental or foundational, unchangeable, timeless, material thing. We can also suppose that (a) the fundamental layer of reality is necessary (as Gel himself supposes in his second argumentative path) and (b) the fundamental layer of reality is cross-world invariant. From all of this it simply follows that the fundamental atemporal wavefunction has no potencies for change, cross-world variance, or non-existence. We therefore seem to have a perfectly respectable naturalist view on which the foundation of reality is a material, unchangeable, purely actual thing. I don't claim that this view is true, of course; my point is simply that nothing Gel says rules it out. And yet Gel would need to rule it out in order to infer the immateriality of the purely actual foundation.

Non-classical routes to uniqueness

Gel concludes by considering two paths to uniqueness that don't require expressly classical theistic assumptions. Gel doesn't explicitly defend the first path, but I think it's worth addressing nonetheless. The first path comes from Hoffman and Rosenkrantz (2002), who ask us to consider two co-existent omnipotent beings named Dick and Jane. Hoffman and Rosenkrantz write:

If this were possible, then it could happen that at some time, t, Dick, while retaining his omnipotence, attempts to move a feather, and at t, Jane, while retaining her omnipotence, attempts to keep that feather motionless. Intuitively, in this case, neither Dick nor Jane would affect the feather as to its motion or rest. Thus, in this case, at t, Dick would be powerless to move the feather, and at t, Jane would be powerless to keep the feather motionless! But it is absurd to suppose that an omnipotent agent could lack the power to move a feather or the power to keep it motionless. Therefore, neither Dick nor Jane is omnipotent. As a consequence, it is impossible that there be two coexistent omnipotent agents. (Hoffman and Rosenkrantz (2002), 168)

The argument here is something like:

1. (1) If two (or more) omnipotent beings could co-exist, then the following conjunction is possible: (i) one such omnipotent being attempts to bring it about that a contingent state of affairs s obtains at time t, and (ii) another such omnipotent being attempts to bring about another contingent state of affairs s* at t, where s*'s obtaining at t is incompatible with s's obtaining at t.

2. (2) Necessarily, if (i) and (ii) are true, then one omnipotent being is powerless to bring s about at t while the other is powerless to bring s* about at t.

3. (3) But it cannot be the case that an omnipotent being is powerless to bring s about at t or powerless to bring s* about at t.

4. (4) So, two (or more) omnipotent beings cannot co-exist. (1–3)

The argument is valid, and both (2) and (3) seem plausible enough. Why, though, should we accept (1)? As far as I can see, it is simply a non-sequitur. Why would the possible joint satisfaction of (i) and (ii) follow from the mere possibility of two (or more) omnipotent beings? By my lights, (1) seems just as (im)plausible as (1*):

• (1*) If one omnipotent being could exist, then the following conjunction is possible: (i) this omnipotent being attempts to bring it about that contingent state of affairs s obtains at time t, and (ii) this omnipotent being attempts to bring it about that contingent state of affairs s* obtains at t, where s*'s obtaining at t is incompatible with s's obtaining at t.

Merely from the fact that there could be an omnipotent being, it doesn't follow that this omnipotent being could attempt to bring about two incompatible contingent states of affairs. Similarly, merely from the fact that there could be two (or more) omnipotent beings, it doesn't follow that they could both attempt to bring about jointly incompatible contingent states of affairs. Perhaps Dick and Jane's wills are perfectly (and essentially) in harmony. Perhaps both Dick and Jane are essentially perfectly rational and essentially omniscient and so necessarily coordinate their actions in a consistent manner. Or perhaps their attempting to actualize incompatible states of affairs is impossible not because two co-existing omnipotent beings is impossible but instead for the following reason: while each individually has the power to actualize the contingent state of affairs in question, neither has the power to actualize it while the other attempts to actualize an incompatible state of affairs. For the actualization of that power would entail a contradiction, and any proper analysis of omnipotence debars the power to actualize contradictions.²¹ Lots of entirely consistent hypotheses abound that would render (1)'s antecedent true and consequent false.

Gel's second path derives from ‘God's unlimited or perfect nature’ (Gel (forthcoming), 7). Gel writes:

[I]t seems plausible that there can only be one absolutely or maximally perfect being, because for there to be two of them, they would need to be distinguished either by a perfection that one of them had and the other lacked or by having the same perfections to differing degrees. But in each case, either one of them or both would not be an absolutely or maximally perfect being, contrary to hypothesis. (ibid., 7–8)

This argument, too, doesn't seem to work. It seems to rest, first, on IoI, and we've already seen why this is problematic. (Why can't there be two perfect beings, each with all the perfections to the maximal degree, that are simply primitively distinct?) It also seems to ignore ways that perfect beings might be differentiated. If – as Christians would have it – being Trinitarian is compatible with being perfect, surely being Unitarian is likewise compatible therewith. It's not as though Jews and Muslims, for instance, are prevented from affirming that God is perfect by dint of accepting Unitarianism. In that case, though, there could – at least in principle – be a feature differentiating perfect beings from one another – namely, the number of persons in which they exist. Finally, I think Gel's argument here is incompatible with Trinitarianism. Consider the following parody:

Whatever is divine is absolutely or maximally perfect. And divine persons are divine. Hence, divine persons are absolutely or maximally perfect. But, plausibly, there can only be one absolutely or maximally perfect divine person, because for there to be two of them, they would need to be distinguished either by a perfection that one of them had and the other lacked or by having the same perfections to differing degrees. But in each case, either one of them or both would not be an absolutely or maximally perfect divine person, contrary to hypothesis.

As before, the same principles and inferential steps are operative in each argument. I don't think one can accept the original argument, then, without also accepting the parody.

Conclusion

Gel has argued that classical theism enjoys a significant advantage over Oppy's naturalism in terms of its simple and explanatorily illuminating answers to the questions of how many first causes there are and why there is exactly that number of first causes. The goal of my article has been to critically engage with this claim.

I began by characterizing classical theism and Oppy's naturalism. I then examined Gel's first argumentative path from causal finitism in favour of the abovementioned claim. After covering some methodological preliminaries about simplicity, I argued pace Gel that Oppy's naturalism is importantly simpler than classical theism. I then argued against Gel's two arguments to the effect that classical theism explains exactly how many first causes there are. Along the way, I connected my findings to wider debates about the gap problem and the Christian doctrine of Trinitarianism.

I next examined Gel's second argumentative path from foundational necessity in favour of the above-mentioned claim. I argued that Gel's second path, like the first, doesn't succeed. After that, I considered two non-classical approaches to establishing God's uniqueness: one from omnipotence and another from perfection. I argued that both approaches fail.

While I have been critical of Gel's arguments, I think they are both innovative and thought-provoking. They advance debates about the gap problem, about the relative merits of classical and non-classical theisms, and about Oppy's naturalism. I hope my article similarly advances such debates into new territory.