CT and contingency

Creator theology, even though the most natural starting point for those for whom biblical literacy was formative, does not get the respect in philosophical theology that is accorded perfect being theology. The source of philosophical discontent is not hard to trace, and I want to begin by clearing the deck of the misconceptions involved in such discontent.

First, a preliminary word is in order about the starting point for each position. PBT begins from the Anselmian idea that God is the most perfect possible being, that than which a greater cannot be thought. These two claims are not identical, and part of the project of PBT is to refine this starting point to make it precise enough for theoretical purposes. Analogously, CT starts from a position that needs refinement as well: it starts from the idea that God is the creator of all. Refinements are in order for CT initially because the notion of a creator will seem to many to imply that the universe cannot be eternal, and it would be best not to require such an assumption of CT. So, one refinement involves replacing the idea of a creator with the more generic notion of sourcehood: God is to be understood as the source of all. A second refinement is required as well, since it would be perplexing to require that God be the source of God as well as everything else. So one should understand CT as the position that God is the asymmetrical source of all else. This claim, too, needs further explication, but we can leave such details for later, as we see the need for further comment below.

It is nearly universally acknowledged that an adequate theology cannot be derived from CT. The grounds that motivate this conception of inadequacy is that CT is assumed to derive its conception of God from cosmological and teleological considerations, calling for a first cause and a designer. It is well known, however, that such considerations give grounds for a limited deity only – one with awesome but not unlimited power, wisdom, knowledge, and concern for something other than self. John Stuart Mill provides an excellent representative for such a position. Mill (Reference Mill1874 [1957]) grants the existence of evidence for a demiourgos or artisan/craftsman for the universe, one of considerable power and knowledge as well as goodness. He noticed as well, however, that nature's expediencies are far from perfect, aimed at the preservation of individuals and species rather than at well-being, and thus provide evidence against any artisan or craftsman of unlimited knowledge, power, and goodness.

This argument from inefficiency is not without criticism. As Tom Morris notes,

Morris notes here that Mill's focus on the imperfections of nature's contrivances presupposes an understanding of the artisan's goals, plans, and intentions in order for what is seen to count as an imperfection. If so, what one infers about God from cosmological and teleological considerations will depend on some prior conception of God's plans, and this prior conception will thus have to appeal to some more fundamental metatheology than any metatheology of the sort Mill is using. So, if CT is to be developed on this Millian model and if the Morris criticism is well taken, CT has no prospect of being the fundamental metatheology. It perhaps generates too limited a conception of deity, but more importantly, it can't generate any conception of deity without relying on some other approach before beginning to draw inferences.

I note this standard way of understanding a theology focusing on God as creator, not to praise it, but to bury it. For this entire discussion is epistemologically driven, and is thus out of place as a contrast to PBT, which is clearly metaphysically driven. CT, from such a metaphysical perspective, is not an account of the nature of God derived from whatever attributes of God can be supported by cosmological and teleological considerations. Instead, it is the view that God is (as I argue below) the asymmetrical source of all else, a starting point on which the fundamental nature of God involves aseity and independence from all else. The idea is to start with the characterization of God that is central to CT and see what can be learned about God from that fundamental starting point. The mistake in the literature is to conflate the metaphysical project with an epistemological one, where one derives a conception of God from epistemological sources rather than exploring what can be derived from a beginning point concerning the nature of God on which God is the asymmetrical source of all else.Footnote ¹

To appreciate better the difference between an epistemological approach and the metaphysical project, note that perfect being theology doesn't begin, either historically or theoretically, in dependence on the ontological argument. Instead, perfect being theology begins from a fundamental conception of God as a perfect being, and develops a theory about God from that starting point. If theory construction goes as Anselm and other defenders of the ontological argument imagine, the theory culminates with a demonstration of God's necessary existence. But that argument comes late in the theoretical game. It comes only after making a list of great-making properties with intrinsic maxima, as in Plantinga (Reference Plantinga1974), and then finding a way to include something involving existence on that list. Perhaps it is possible existence, or existence-at-a-world. My preferred approach is to talk about possible beings that have maximally fragile existence, where any change to a world in which they exist results in a world where they don't exist; then define durability as the capacity to continue to exist despite changes, with maximal durability implying necessary existence. If all goes well here, we discover that God is not only an omni-being, but also necessarily existent if possible. Notice, however, the metaphysical nature of the project: we start with the perfect being characterization, and derive the ontological argument in the process of characterizing what maximal perfection involves. We don't start with an ontological argument and then embrace a limited theology constrained to include only the attributes of being a possible being, being necessary if possible, and thus being both a necessary and an actual being. PBT's promise is first to generate an exalted conception of God, and then consider whether it can be demonstrated that such a being exists.

Just so with CT. One starts with a basic conception and builds a theory, a theology, from it. Such a project is independent of the epistemological issue of whether and how one might come to know or have reason to believe that the being so described exists. So let us consider what such a metaphysical project looks like, in contrast to the more standard epistemological approaches to the nature of God as creator.

Creator theology involves the central claim that God is responsible for what there is, and I use the term ‘creator’ here because of familiarity, even though it can mislead as well. Talk of a creator communicates the idea of a specific activity at some point in history when things came into being, but we should not saddle CT with such a claim. Instead, we should think of the view in terms of sourcehood and ontological independence of an asymmetrical sort.

I thus favour a more generic characterization of CT. God is the source of all else, and is ontologically independent of all else, features that get us at least very close to what philosophical theologians have in mind when they use the language of aseity to describe God. In some narratives, as in Genesis, such sourcehood takes the form of creation at a certain point in history (‘in the beginning God created’), though it is easy to imagine alternative possibilities of sourcehood that do not involve creation, such as Aristotle's eternal universe finding its source in an Unmoved Mover.

Theological development then proceeds by addressing how sourcehood and independence are to be understood. One can note, however, that even without any elaboration, the view has a pristine and restrained ontology of the divine: there can only be one asymmetrical source of all else, by definition. Moreover, this definitional result is not achieved simply by the definitive description beginning with the word ‘the’. Even if we had said that the nature of deity involves being an asymmetrical source of all else, the same result follows: there can be only one such asymmetrical source, in virtue of being the source of all else.

Note this difference from what we get with Anselmian PBT, which originates from the phrase ‘that than which a greater cannot be thought’. There is nothing in this beginning point alone that yields monotheism. One could get such a result by replacing this language with that of the greatest possible being, but that gets the monotheistic result on the cheap. What is theoretically desirable is to show that there can't be ties for being the greatest, and so the beginning point of PBT does not yield monotheism as directly as does CT.

Notice as well that in both cases, we could avoid any commitment to monotheism by resorting to logical plurals. Doing so would have us say that the gods, however many there may be, are either the asymmetrical source of all else or the greatest possible beings.Footnote ² Such polytheistic-friendly versions face difficult issues almost immediately. First, it is no virtue of a theory to be ontologically profligate, so an explanation would be needed why more than one deity is important to affirm. Second, complications arise in characterizing the community of deities itself. Are there conflicts? Are there differing roles played by each member? Are some in higher positions of authority than others? Such questions, depending on the answers, can quickly lead to the suspicion that what is being characterized is, at best, only a collection of divine beings rather than one of deities. And things can get worse, for if conflict can arise in the community, who emerges triumphant? If there is a winner, one would think that it is the winner who would be God, not the loser.

It would be easy to be sidetracked by such questions, but without any reason for positing a multiplicity, we needn't take the detour. Furthermore, reasons for positing a multiplicity will not be found in the general features of inquiry that arise for any metatheology, but would presumably arise in particular religious communities and involve claimed support from divine communications of one form or another. To begin CT or PBT from an ontologically parsimonious point does not imply that the starting point can't be revised in light of further information of this sort. But it does imply the need for further information before doing so. So, for present purposes, we begin from the grammatical singular, noting that such a version of CT directly and immediately entails monotheism.

My intention here is to take a cautious approach as to how much of a benefit this result confers on CT. For the present, it is enough to note that multiplicity in a theology generates theoretical difficulties, some of which are noted above. An approach that can derive monotheism immediately is thus advantaged, but it is hard to see why this advantage would be conclusive. If PBT can find some other, less immediate, way of deriving the same result, the immediacy of the derivation is of little value if any. Though there are attempts to show that there cannot be more than one omni-being,Footnote ³ the arguments generally fall short by failing to note that it isn't impossible for multiple beings to be in full modal accord in heart, mind, and will. Moreover, multiplicity with this kind of modal accord may be all that is needed. Multiplicity comes in two forms, one with the kind of embarrassing interactions and conflicts common in the mythology of the Greek pantheon, and another form where the multiplicity in question must be in harmony. Once we have such essential harmony as a requirement on multiplicity, we also get a kind of unity within the divine realm that is at least close to a social Trinitarian view common in recent Christian thought.Footnote ⁴ If such a social Trinitarianism is not a black mark against a theology, then perhaps the failure by PBT to easily derive monotheism is not a problem either, at least so long as they can sustain a derivation of the claim that the divine realm must be harmonious. At the very least, however, the failure is not debilitating on its own.

Once theoretical development of CT begins, a primary question concerns the nature of sourcehood itself, and the first stage of inquiry examines whether such sourcehood should be understood as agential or non-agential. I have argued elsewhere, in Kvanvig (Reference Kvanvig and Kittle2019), that mysteries are best avoided by thinking about sourcehood in agential terms. If so, sourcehood involves dependence on the will and intellect of the creator, where mind and will are the minimal conditions for being an agent. The best way to develop this view is to hold that contingent truths depend on the will of God and necessary truths find their source in the operations of the Divine Mind.Footnote ⁵ There are philosophical concerns at each stage of this account, but let's pass on these issues for now to see their implications.

The central implication of concern here is that the will of God, on this account, is utterly unconstrained, except insofar as its operation answers to the mind of God. Metaphysical necessities constrain what can be done, of course, but no constraints beyond these can be found when we are talking about the asymmetrical source of all else. Moreover, if metaphysical necessities are themselves explicable in terms of operations of the Divine Mind, these constraints are internal rather than external constraints. The will of God is thus completely free of influence or direction from any outside forces, and thus creation itself is a free act of God. We thereby generate a pleasing account of contingency: anything that finds its source in the will of God comes indelibly stamped with the mark of contingency. That is, not only are the things so-sourced contingent, they could not fail to be so: there is no such possibility as merely contingent contingency, and it is part of the essence of any contingent nature that it is contingent.

The same is not to be said about things that find their source in the operations of the divine mind, for here we find the landscape of modality itself.Footnote ⁶ It, too, is marked indelibly: with respect to that which is explained solely in terms of the operations of the Divine mind, there is no such thing as contingent necessity. Here, there is no potentiality in terms of such operations that would result in the operations being one way but capable of another. The argument for this conclusion is simply the incoherence of Cartesian voluntarism regarding necessary truths.Footnote ⁷

CT thus generates a two-part account of natures and their essential properties, so that both contingency and necessity are built into natures in virtue of their source in either the mind or will of God, and the resulting divine ontology is strictly monotheistic. Let's now turn to see how PBT can be and needs to be developed, returning to the question of contingency once we have a picture in mind of what the metatheology of PBT looks like.

The metatheology of PBT

PBT begins by identifying God as a maximally perfect being. This metatheology can be developed in two ways, top-down or bottom-up. The latter is the more familiar form, with Plantinga (Reference Plantinga1974) as the most sophisticated version of the view. On Plantinga's account we first identify great-making properties that have intrinsic maxima and then derive the claim that God is an omni-being, perfect in power, knowledge, and goodness. Once this characterization is in place, the bottom-upper aims to get existence itself, or something involving it, included in the collection of great making properties, so that necessary existence will count as the maximum value possible. (My preferred way of expressing this idea is to talk about possible beings that have maximally fragile existence, where any change to a world in which they exist results in a world where they don't exist; then define durability as the capacity to continue to exist despite changes, with maximal durability implying necessary existence.) So God is not only an omni-being, but also necessary if possible.

If we then add the claim that it is possible for there to be a maximally great being, the ontological argument achieves lift-off. My goal here, though, is not to address whether that argument succeeds (or what the relevant notion of sucess is supposed to be), but rather to understand the structure of PBT. The crucial element of this structure is where the property of being necessary if possible enters the story. The ontological argument depends on the claim that there is a possibly necessary being, and this property of being possibly necessary is derived from two other properties: the property of being possible and the property of being necessary if possible. These latter two properties are added to a collection of great-making properties that have intrinsic maxima, so the structure of PBT involves two collections plus a possibility claim. The first collection is the set of great-making properties that purportedly yields the omni-being conception of God, what Plantinga terms the collection of ‘maximal excellences’, where a maximal excellence (ME) is the maximal form for a great-making property with an intrinsic maximum. So, where P is a property and ME _P says that P is a maximal excellence, we get a collection C(ME) defined as follows:

C(ME): {P:ME _P}

The second stage involves adding to C(ME) the property of being necessary if possible, on grounds that durability of possible existence is itself a great-making property with an intrinsic maximum. That maximum is, of course, necessary possible existence, or the property of being necessary if possible. This collection Plantinga terms the collection that defines the notion of maximal greatness: maximal greatness is C(ME) plus the property of being necessary if possible. So, where p is a proposition:

C(MG): {P:ME _P} ∪ {λp(p→□p)}Footnote ⁸

The final stage in the development of PBT is to claim that the collection C(MG) is metaphysically coherent:

◇PB: ◇(all the members of C(MG) are co-exemplified).

That is the basic structure of bottom-up PBT: C(ME), C(MG), and ◇PB. If successful, it generates an understanding of God in terms of an omni-being and demonstrates the existence of such a being through the ontological argument.

The challenges to this bottom-up approach concern the coherence of both C(ME) and C(MG) and the possibility claim, ◇PB, itself. Perhaps the properties in C(MG) are not coherent, either internally, or with each other, or with certain known facts about the world. Moreover, even if C(MG) is logically consistent, ◇PB does not follow from that fact, since such consistency does not imply metaphysical possibility.

In response to the first concern, one might resort to a top-down strategy, where instead of building up a collection of maximal excellences by identifying the maximal value for each great-making property included in the collection, one instead starts with the most impressive and possible total collection of great-making properties.Footnote ⁹ Such a holistic strategy offers, first, a replacement for C(ME), one that guarantees metaphysical coherence among the members of the set. If C(ME) is itself metaphysically coherent, the top-down approach yields the same result as the bottom-up approach. But if C(ME) isn't coherent, all is not lost. Instead, the top-down approach simply generates a set as close to C(ME) as is coherent.Footnote ¹⁰ Call this set ‘C(ME)-lite’. The advantages of this top-down holism are clear. If omniscience is incoherent, no problem: the holism then only requires that God be as knowledgeable as is compatible with all other individual great-making properties. If omnipotence conflicts with omnibenevolence, no problem either: just withdraw from the lofty peaks a bit, so that God is very powerful and very good. C(ME)-lite is very forgiving when it comes to the potential problems for bottom-up PBT.

As articulated in Nagasawa (Reference Nagasawa2017), one then constructs ‘C(MG)-lite’ from C(ME)-lite by joining together the latter with the property of being necessary if possible. One then replaces ◇PB with ‘◇PB-lite’, the claim that the properties in C(MG)-lite are possibly co-exemplified, running the ontological argument on this premise instead. For that argument to succeed, once the possibility premise is defended, we need only the distinctive S5 claim that anything that is possibly necessary is necessary.Footnote ¹¹

The advantage of this top-down approach is that the notion of coherence invoked to construct C(ME)-lite is best thought of in terms of metaphysical coherence, and thus saying that C(ME)-lite is coherent is the same thing as saying that the co-exemplification of this set of properties is metaphysically possible. That makes it appear at first glance, at least, that we can easily derive ◇PB-lite from this coherence claim, but that appearance is misleading. For ◇PB-lite isn't a claim about C(ME)-lite, but rather about C(MG)-lite. Furthermore, even if the members of C(ME)-lite can be co-exemplified, that provides no guarantee that the members of C(MG)-lite can be co-exemplified, for the same reasons that one might worry that the members of C(ME) might not be able to be co-exemplified: combining coherent properties doesn't always preserve coherence.

So, here's the argument for being a top-downer. C(ME) has no guarantee of coherence, whereas C(ME)-lite does, so there is one way for incoherence to obtain for C(MG) that can't obtain for C(MG)-lite. So, be a top-downer!

Advantages can dissipate quickly on closer inspection, and such is the case here. Notice the recipe to generate the above advantage for top-down PBT. The first step is to find a collection of properties that are guaranteed to be metaphysically coherent, and the second step is to hope the ship can still hold water when we add to this collection the further property of being necessary if possible.

We had better not be committed to the coherence of such an addition for any coherent collection of properties. Allowing such a manoeuvre bloats our ontology and threatens reasonable assumptions about contingency itself.

Begin by distinguishing metaphysically coherent sets of properties from natures. A nature is complete, whereas a set of properties need not be. The notion of completeness here is vexed, but if we adopt both bivalence and excluded middle, it is easy enough to characterize: a complete nature is something that can be exemplified, and includes, for every property P except those properties that entail exemplification, either P or its complement $\overline P $ , that property exemplified by everything that lacks P. If either bivalence or excluded middle are abandoned, a different account of completeness will be required, but we need not pursue the details here, for all we need to note is than any actual being in any possible world has a completeness that is to be contrasted with the incompleteness of fictional entities. There is no fictional fact of the matter about whether Sherlock has a mole on his left calf; there is an actual fact of the matter for each of us whether we have such a mole.

I note that it is essential to the methodology of either version of PBT to move from properties that are (purportedly) individually coherent, to sets of compossible properties that are clearly less than complete natures, to the possible exemplification of such a set together with the additional coherent property of being necessary if possible. It is here, in the gap between collections of compossible properties and complete natures, that problems arise for PBT's understanding of contingency.

With this distinction in hand, turn to the issue of what happens when we assume that any compossible set of properties will remain compossible when the property of being necessary if possible is added to the original set. Take any nature and divide it into those properties that metaphysically guarantee contingency and those that don't. A plausible division here will have the intrinsic properties of any object in the group that doesn't guarantee contingency, so that the set of intrinsic qualities of my coffee cup do not guarantee contingency. Any object that has all of the intrinsic qualities of my coffee cup will be indistinguishable from it. So now take this compossible set of intrinsic qualities and add to it the property of being necessary if possible. If this supplemented collection is also metaphysically coherent, we now get a stunning duplication in ontology: we have both my actual contingent coffee cup and its metaphysically necessary doppelgänger. For the intrinsic character of my coffee cup is clearly exemplifiable (because it is actually exemplified!), and possibility entails necessary for any compossible set of properties that includes being necessary if possible.

The same recipe can be followed for any contingent thing whatsoever, so loose scruples on when the addition of the modal property preserves metaphysical coherence lead to the absurd conclusion that every contingent thing has its own necessary doppelgänger. This result combines a problem of profligacy with a problem of contingency: we get too many things, and the additional things we get threaten our understanding of contingency. My coffee cup, as I understand its nature, is something that can be destroyed so that it ceases to exist. If being so requires that it have a necessary doppelgänger that can't be destroyed, I think I don't understand this notion of contingency, I think we must have become confused in the process of finding a need for such doppelgängers, and I certainly don't understand how this necessary thing can be the doppelgänger of my coffee cup, once the cup is destroyed and the necessary being now shares no properties with this cup.

Can defenders of PBT take refuge in the claim that the intrinsic qualities of my coffee cup actually do guarantee contingency? Well, perhaps if they had an argument for the claim. There is an attractive one available (though I will stay short of endorsing it here), but it undermines the methodology PBT needs Here's the argument. If an object is contingent, it is necessarily so. So, when we strip natures of some of their properties, and end up with a coherent subset of properties, we can only get to the conclusion that this set is possibly exemplified by retaining in the set all the essential properties of possible exemplifiers. In the case of my coffee cup, we thus have no reason to think that stripping its nature of contingency leaves us with anything other than the empty set.

This argument is intriguing, but deadly to the methodology of PBT. For if the argument is endorsed, any property or set of properties includes contingency or necessity in the properties themselves, and in a way that can't be detected simply by looking at the identifying description of that property. The intrinsic properties of my coffee cup – being oval in shape, white in colour, having a mass of 273 grammes, etc. – carry no conceptual or analytic connection to being contingent. So a defender of PBT simply can't take collections of great-making properties – either the omniproperties themselves or coherent limitations of some or all of them – and be guaranteed to have a further coherent collection of properties when adding the property of being necessary if possible. The only way to have such a guarantee would be for these properties to be free of the contingency implication that is supposed to plague the intrinsic character of my coffee cup. But such a result upends the order of inquiry central to PBT: it puts the needed conclusion in place before the derivation can be completed.

Could a PBT-er take refuge in the idea that there simply aren't any contingent things at all?Footnote ¹² No. On such an ontology, the distinction between necessary and contingent has to be replaced by something else, such as the distinction between being concrete and being merely abstract. So reality is still bifurcated in an important sense, and the property appended to coherent collections for purposes of the ontological argument will not the the property of being necessary if possible, but rather the property of being necessarily concrete if possible. All the same problems re-emerge, just under a different label.Footnote ¹³ So, let's abandon the idea of re-labelling and press on with the problem under the more usual labelling.

Generating these problems of profligacy and contingency is straightforward. Both top-downers and bottom-uppers hold that the modal property of being necessary if possible can be coherently added to some metaphysically coherent collection of great-making properties. But if we grant that claim, we need an account of when and where such appending preserves metaphysical coherence.

A defender of PBT can resist some of the above, at least. For the argument for the existence of necessary doppelgängers depends on generalizing the central move of PBT, and a defender of PBT can rightly point out that such a generalization was never part of PBT itself. That point is correct. But a related problem will still be present, and it results in the disappearance of the advantage noted above for top-down PBT over bottom-up PBT. Bottom-uppers restrict how the modal property of being necessary if possible gets into the story of the nature of God: it gets in only because there is a great-making property regarding which this modal property is maximal. Top-downers don't get to limit great-making properties in this way, since they don't want C(ME)-lite to be hostage to the metaphysical coherence of the maximal properties involved in being an omni-being. So, top-downers have to countenance less than maximal great-making properties in their account of the nature of God, and so have to embrace a recipe for constructing C(ME)-lite and C(MG)-lite that doesn't restrict membership to maximal great-making properties.

This feature of top-down PBT engenders profligacy concerns not generated by the recipe endorsed by bottom-uppers. Notice that C(ME)-lite is trivially coherent, in virtue of the recipe for its construction. Moreover, if p is logically stronger than q (i.e. if p entails q but not vice versa), then if p&r is metaphysically possible, then so is q&r. Hence, if C(MG)-lite is coherent, any further weakening of C(ME)-lite will also cohere metaphysically with the property of being necessary if possible. There are indefinitely many such weakenings, and thus the coherence of C(MG)-lite provides guarantees of the coherence of all the lower-calorie lites to come. Bottom-uppers can now rejoice, because their recipe only allows the conjoining of properties that are both maximal and great-making; top-downers have to allow conjoining of less than maximal but still great-making properties. They thereby generate a divine embarrassment of riches: not one divine being at most, but as many as one wants to count, one for every combination of great-making properties that is weaker than C(ME)-lite. It is an inconvenient profligacy in ontology akin to the Greek pantheon. Add the further property of behaving in ways befitting a soap opera and there you are.

Bottom-uppers can insist that we not tread this path, that we shouldn't append the modal property to any and every metaphysically coherent collection, but must generate inclusion of this modal property into a set by a well-constructed recipe that forces its inclusion. Such a restriction, if defensible, will keep the problems of profligacy and contingency at bay, and in this way the earlier advantage touted by top-downers ends up dissipating once we notice this advantage of the bottom-up approach. There is thus at least an interesting dispute about which version of PBT to embrace, but one not to pursue at present, for things get worse for both versions.

CT and the problems of contingency and profligacy

Things get worse precisely because no version of PBT can offer an explanation why the property of being necessary if possible must be combined only with some metaphysically coherent collections of properties and not others. Restrictions are needed to avoid the problems noted above, but it is one thing to insert restrictions to avoid a problem, and quite another to have an independent explanation of exactly why such restrictions are appropriate. It is the lesson of a reductio: when disaster follows, you have to give up something. To retain PBT, you have to restrict. But there is a further option: give up PBT itself or modify it in some way. To retain PBT in its present form, one thus needs a reason for requiring the restriction needed to avoid the difficulties noted, one that explains why such a restriction is needed in a way that doesn't simply point to the difficulties that arise apart from such a restriction. The problem is that of determining to which coherent combinations of properties we can coherently append the property of being necessary if possible. The solution to that problem will imply or depend on an independent account of which combinations of properties must be contingent and which must be necessary. Only such an account can give us a way of determining when adding the property of being necessary if possible will result in a metaphysically coherent combination or a metaphysically impossible one. PBT has no theoretical resources to generate such an answer, so its only recourse is to restrict merely on grounds of needing to avoid the disaster.

Here CT, at least in its classical formulation, has a story to tell. Recall that CT is the claim that the starting point for developing an adequate theology is the claim that God is the asymmetrical source of all else. The first task for the classical view, after noting that there can be only one asymmetrical source of all else, is to show why such sourcehood requires that God is a person. I have argued for this claim in Kvanvig (Reference Kvanvig and Kittle2019), so won't repeat the argument here; but it is crucial for the classical formulation of CT to affirm the personhood of God, since that version of CT finds the source of contingent truth in the will of God and the source of necessary truth in the operations of the divine mind. Moreover, defenders of CT can insist that being the source of a thing isn't a matter of finding a compossible set of properties and co-exemplifying them, since that account will encounter the problem of incomplete objects noted above. Instead, God's responsibility for the existence of things involves instantiating natures, where a nature is complete except for the singularity involved in its exemplification. So, smaller collections of properties are capable of co-exemplification only when they are part of some complete nature or other. Finally, it is an essential property of every nature that it finds its source either in the will of God or in the operations of the divine mind. Contingent things are thus necessarily contingent, just as possibly necessary things are actually necessary. So, since all natures depending on God's will are, of necessity, contingent,Footnote ¹⁴ we have a recipe for determining to which coherent collections of properties we can coherently append the property of being necessary if possible: only to collections that are part of a nature that finds its source in the operations of the Divine mind.

I hasten to point out that I am not claiming that this position is defensible or even coherent. In particular, there is a serious need to articulate what is involved in explaining necessity in terms of the operations of the divine mind.Footnote ¹⁵ This version of CT is, in this respect, in precisely the same situation as, for example, bottom-up PBT: there is the position as articulated, and there is a need for a defence of the position's coherence and plausibility. All I am claiming here is that there is a standard formulation of CT that has little difficulty addressing the two problems that I claim plague PBT.

Defenders of CT can thus be rightfully suspicious of the major premise of the ontological argument, when presented without some accompanying account of contingency to make clear whether the appending of the property of being necessary if possible is mere chicanery. (It is probably worth noting in passing that the CT account provides no help in defending the premise in question, since the natures to which the modal property can be appropriately appended are precisely those that are already assumed to be necessary beings.) A CT-er can have substantive reasons for limiting the range of natures to which one can coherently append the property of being necessary if possible, and substantive reasons for rejecting such an appendage to any nature not needed to account for some aspect of necessity itself. Such reasons may themselves be overridden or defeated by further considerations beyond the scope of this article, but it is one thing to have no story to tell, and it is another thing to have a defeasible one.

In contrast, not even bottom-up PBT is in such a position. It might insist on conjoining the property of being necessary if possible only to other great-making properties with intrinsic maxima. If pressed as to why such conjoining is not OK in other contexts, all that can be done is to cite the reductio resulting from the problems of contingency and profligacy. Doing so leaves one in an uneasy state, wondering what it is about the idea of maximal excellence that makes it a legitimate exemption here. Without such an explanation, the vestigial remains of the problems of contingency and profligacy plague even bottom-up versions of PBT.