What is panentheism? Three desiderata

Panentheism involves a claim about God’s immanence and his transcendence. The cosmos is in God but God is more than the cosmos. How should we go about developing a more philosophically precise interpretation of this view? One approach is to survey textual traditions that purport to be panentheistic, or are otherwise interpreted in a panentheistic fashion, and see what the texts in question say about the nature of the divine and its relationship to the cosmos. While panentheism is found in many religious and philosophical traditions, the Bhagavad-Gīta presents us with arguably one of the most well-articulated and well-known statements of panentheism, at least in the classical Indian tradition.Footnote ³ My methodology, therefore, involves deriving three desiderata from various verses in the Bhagavad-Gīta that I think a minimal account of panentheism must adhere to.

Since panentheism is a theological and philosophical view about the relationship between God and the cosmos, my three desiderata are conditions that any minimal account of this relation must satisfy. My first desideratum says the relation in question must be asymmetric.

An important passage in the Bhagavad-Gīta expressing Asymmetry is the following from chapter 9, where Krishna describes his relationship to the cosmos.

As Silvestre and Herbert (Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 19) note, a key term in this verse is stha, which means ‘to support’, ‘to abide in’, ‘to rest in’, or ‘to stand in’. They argue this term picks out a strong notion of ontological dependence; the very existence and/or essence of creatures depends upon the existence and/or essence of God. Insofar as creatures derive their existence from God, but God is not limited by them, this relation is arguably asymmetric.

Some understand panentheism as the interrelatedness of God and cosmos: the cosmos is in God and God is in the cosmos. And this implies a symmetric relationship rather than an asymmetric one. For example, John Culp (Reference Culp, Edward and Nodelman2023) writes, ‘Panentheism considers God and the world to be inter-related with the world being in God and God being in the world’. Similarly, Philip Clayton (Reference Clayton, Diller and Kasher2013, 372) says, ‘The “en” of panentheism is almost always a two-fold “in”: the transcendent is in the immanent, and the immanent is in the transcendent. Or, in the beautiful words of the Bhagavad Gita, “He who sees Me everywhere and sees all in Me; I am not lost to him nor is he lost to Me”’.

While the interrelatedness of God and cosmos may be a feature of some panentheistic traditions, I believe we should avoid this interpretation with respect to the Bhagavad-Gīta. Theologically speaking, strict interpretations of interrelatedness, according to which God is in the cosmos in the exact same sense in which the cosmos is in God, undermine God’s transcendence, where ‘transcendence’ is understood in terms of God’s ontological independence or fundamentality. If we interpret the ‘en’ of panentheism in terms of a dependence relation as Silvestre and Herbert (Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024) do, then strict interrelatedness amounts to saying that God is ontologically dependent upon the cosmos in the very same way in which the cosmos is ontologically dependent upon God. That would make God a dependent being. But the Bhagavad-Gīta very clearly affirms what the Western tradition calls ‘divine aseity’, from the Latin a se, meaning of or from itself. Aseity denotes God’s fundamentality or ontological independence. And the Bhagavad-Gīta is replete with statements of this doctrine, especially in later chapters. Consider the following verses.

These verses express clearly the notion of God’s aseity and the cosmos’s dependence upon God. Silvestre and Herbert (Silvestre and Herbert Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 21) agree, arguing that the Bhagavad-Gīta affirms that ‘God is the (only) absolute fundamental entity; he is the ontological foundation of all that exists’. To preserve God’s transcendence or fundamentality, therefore, we ought to say that the sense in which the cosmos is in God differs from the sense in which God is in the cosmos. For example, we might say that the cosmos is ‘in’ God in the rough sense that it depends upon God for its existence, while God may be ‘in’ the cosmos through his divine power or omnipresence.

My second desideratum says the panentheist’s God-world relation must be a relation of overlap. The relation in question must capture the important sense in which the cosmos is ‘in’ God, and so must be a quasi-mereological relation. I say ‘quasi-mereological’ because the relation can’t literally be the part-whole relation of standard mereology. Instead, I argue below that the grounding relation offers a non-technical sense in which two things can ontologically overlap by sharing their being or existence.

In my view, as we’ll see shortly, this desideratum expresses the essential core of panentheism that distinguishes it from rival accounts.Footnote ⁴ This requirement is, of course, implied by the very term ‘panentheism’ itself. And it’s clear from many passages in the Bhagavad-Gīta that the cosmos is in some sense supposed to be ‘in’ or ‘part of’ God, which implies there is no sharp distinction between God and cosmos.

A particularly important passage is when Krishna reveals himself to Arjuna in that terrible vision of chapter 11 of the Bhagavad-Gīta. Krishna says, ‘Behold the gods of the natural world, and many more wonders never revealed before. Behold the entire cosmos turning within my body, and the other things you desire to see’ (Easwaran Reference Easwaran2007, 94). Arjuna responds,

Here, we find the famous idea of the cosmos as God’s body that later influenced Rāmānuja’s theorizing.Footnote ⁵ Similarly, in a later verse Arjuna says, ‘Lord of the gods, you are the abode of the universe’ (Easwaran Reference Easwaran2007, 199). As we’ll see below, some thinkers offer interpretations of the ‘en’ of panentheism that fail to respect the mereological or quasi-mereological sense of the term used in the Bhagavad-Gīta. One of my key contentions, then, is that an adequate account of panentheism cannot forgo the idea of the cosmos overlapping with God.

My third desideratum says the panentheist’s God-world relation must be ontologically generative or productive. Simply by virtue of being related to God in the appropriate way, the cosmos derives its very being or existence from God. Thus, the relation in question must have the ability to generate or produce the existence of the relata in question.

We’ve already seen this idea of ontological generation or production in some of the quotes presented above. For example, Krishna expresses the idea of generation when he says, ‘I am the source from which all creatures evolve’ (Easwaran Reference Easwaran2007, 184), and ‘I am the seed that can be found in every creature, Arjuna; for without me nothing can exist, neither animate nor inanimate’ (Easwaran Reference Easwaran2007, 189). So, our third desideratum is a corollary to the second one since ontological overlap – the notion of two things sharing their being or existence – entails a notion of ontological generation or production.

Nonetheless, it is worth formulating our third desideratum explicitly because it precludes appealing to relations that connect entities whose existence is logically independent of their standing in that relation to begin with. For example, on a Humean conception of causation, cause and effect merely instantiate a regularity. When x causes y, it is not the case that x produces or generates the very existence of y. A Humean view of causation, therefore, wouldn’t be suitable for understanding the panentheist’s God-world relation. To have God as the source in virtue of which all things derive their very existence, the panentheist needs an ‘oomphy’ or inflationary relation that is ontologically productive.

I’ve presented three desiderata that an account of panentheism must adhere to. The Panentheist’s God-world relation must be (1) asymmetric, (2) quasi-mereological, and (3) generative or ontologically productive. Below, I argue that the relation of metaphysical grounding meets these conditions. Panentheism is the view that God’s existence grounds the existence of the cosmos. But first I consider several extant accounts of panentheism in the next section and show that they fail to meet the desiderata I’ve outlined here.

Extant accounts of panentheism

The difficulty with providing an adequate account of panentheism stems from, in part, the implausibility of straightforward mereological interpretations of the sense in which the cosmos is in God. To see why, consider Oliver Crisp’s (Reference Crisp2019, 32) toy account of what he calls ‘mereological panentheism’.

Given the spatial connotation of saying the world is in or overlaps with God, we might naturally understand panentheism in straightforward mereological terms. But everyone seems to agree it is implausible that the cosmos is a part of God in the same way that the legs are part of my desk. One issue has to do with dependence. As Silvestre and Herbert (Silvestre and Herbert Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 26) argue, if the cosmos is part of God in this sense, then God would be a dependent being because at least some mereological wholes are dependent upon their parts. Furthermore, if the cosmos is a part of God in this sense, then it follows God is a mereologically complex whole, which is incompatible with the doctrine of divine simplicity (Crisp Reference Crisp2019, 35). Crisp notes further issues as well, not least of which is the troubling result that God, who is supposed to be incorporeal, has a physical part.

However, rejecting a straightforward mereological account of panentheism provides little guidance for how to move forward. While there are many extant proposals, it isn’t clear to me that any of them meet the three desiderata I proposed above. For example, Benedikt Göcke (Reference Göcke2013) argues that classical theism and panentheism differ with respect to the modal status of the world. For the panentheist, the world is ‘an intrinsic property of God’, such that the world exists necessarily. Alternatively, for the classical theist, the world is an extrinsic property of God’s, and so exists contingently. As Göcke (Reference Göcke2013, 66) says, ‘According to panentheism, “God requires a world”’.

My concern with Göcke’s account is that it runs into problems with the asymmetry of dependence, which is my first desideratum above. If the world is an intrinsic property of God, then God has this property in virtue of God’s own nature rather than in virtue of anything extrinsic to God.Footnote ⁶ In other words, you can’t have the world without God. But this modal approach to God’s relationship with the world can’t capture an asymmetric relation of dependence between them. In fact, it seems to license a symmetric instance of dependence. Suppose we said the following.

The problem with this modal approach, noted by Kit Fine and others, is that it makes dependence symmetric.Footnote ⁷ On Göcke’s proposal, it is necessarily the case that if God exists, then the world exists. But it is also necessarily the case that if the world exists, then God exists too because the world is an intrinsic property of God. They are, in an important sense, inseparable. So, the problem is that this modal approach will license dependence running in both directions, making it symmetric. This is not to say that Göcke’s view on the modal status of the world isn’t an important aspect of certain accounts of panentheism. But it looks like we’ll need to appeal to something more than the modal status of the world to capture the asymmetric dependence between the cosmos and God that my first desideratum demands.

Ryan Mullins tentatively suggests an account that identifies God with time. And since the universe is in time (but not identical with time), it follows that the universe is in God (Mullins Reference Mullins, Göcke and Medhananda2023, 251). This approach promises to offer a literal interpretation of the cosmos being in God. While interesting in its own right, Mullins’s account also fails to meet my three desiderata. Consider the overlap desideratum, which I take to be central to panentheism. Any account of panentheism needs to say in what sense the cosmos overlaps with God. Mullins understands this condition in terms of the universe being in time. But rather than clarify what it means by the universe to be in time, Mullins instead focuses on clarifying what it means to say God is time. Mullins presents the following definition of time. ‘Time is (i) a natured entity that makes change possible, (ii) the ontological source of moments, and (iii) that which orders the moments’ (Mullins Reference Mullins, Göcke and Medhananda2023, 254).

Mullins suggests that God is identical with time insofar as he plays the roles of time specified in the above definition. For the first condition, God is a natured entity that necessarily exists and makes change possible. For the second condition, the existence, nature, and actions of God serve as the ontological source of moments in time. For the third condition, the exercise of God’s freedom can explain the ordering of those moments.

So, for Mullins, the sense in which the cosmos is in God is simply the sense in which the cosmos is in time. But to say that the cosmos is in time is not to say that the cosmos and God overlap in any sense. Rather, it is simply to say that the cosmos is temporal. Mullins’s account, therefore, seems to abandon a quasi-mereological interpretation of ‘en’ that I and many others have argued is essential to panentheism.

Georg Gasser (Reference Gasser2019) proposes to understand panentheism in terms of God’s omnipresence, or more accurately, God’s divine activity. On Gasser’s view, God is wherever he acts. Since God acts upon all of creation everywhere and anytime, God is therefore omnipresent. Thus, Gasser (Reference Gasser2019, 43) proposes to read the ‘en’ of panentheism in an agential sense: ‘God is in the cosmos by creating and sustaining it and the cosmos is in God by constantly being within the sphere of divine activity’. I think Gasser’s approach is initially plausible. It does have the resources to meet my first desideratum, Asymmetry, because God’s creative activity is asymmetric. The world is created and sustained by God, and not vice versa. Similarly, one might argue that the cosmos, as a derivative entity, overlaps with God insofar as it is constituted by God’s creative activity, which is more fundamental. And one might also think that God generates or produces the existence of the cosmos through his creative activity.

However, this approach is ambiguous when it comes to the notions of overlap and production. God’s omnipresence is undoubtedly an important component of panentheism, as I mentioned earlier. But it is unclear to me how the creative-activity relation, whatever it is, can capture the relevant notions of overlap and production that my desiderata target. If there is a unique non-grounding notion of overlap and production that the creative-activity relation delivers, then my suspicion is that this notion of overlap would be somewhat strained. Recall that the notion of overlap my desideratum targets is a quasi-mereological one. It is supposed to mirror mereological overlap without actually being a kind of mereological overlap. And it is unclear how the creative-activity relation can do this. My own view, presented below, is that we’ll need to appeal to a notion like metaphysical grounding to capture the relevant notions of overlap and production that I have in mind.

Silvestre and Herbert (Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024) provide an account of panentheism that I think is much closer to meeting my three desiderata outlined above. They propose to understand the claim that ‘the cosmos is in God’ in terms of the claim that ‘all that exists ontologically depends on God’, where x ontologically depends upon y just in case x depends upon y for its very being or existence and nature (Silvestre and Herbert Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 20). They develop this account in terms of God’s prakṛti, where prakṛti is ‘a metaphysical primitive denoting the intimate relation that exists between matter and conscious living beings on one hand, and God on the other’ (Silvestre and Herbert Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 18). On their view, all that exists is God’s prakṛti. They argue that prakṛti involves both intrinsic intimacy in the form of parthood and ontological dependence, which leads to the following thesis: ‘Law of prakṛti: if X is a prakṛti of God, then (1) X is part of God, and (2) X is ontologically dependent on God (like physical things depend on space)’ (Silvestre and Herbert Reference Silvestre, Herbert, Silvestre, Herbert and Paul Göcke2024, 27).

On Silvestre and Herbert’s view, the cosmos is part of God in the sense that the cosmos is God’s prakṛti. Prakṛti involves both an intrinsic notion of parthood and ontological dependence, as specified in their definition of ‘Law of prakṛti’ above.

I think Silvestre and Herbert’s decision to understand the panentheist’s God-world relation in terms of ontological dependence is the correct approach. Indeed, their account of panentheism is similar to the one I offer below because ontological dependence and metaphysical grounding are similar notions. Silvestre and Herbert affirm that ontological dependence is an asymmetric relation, which meets my first desideratum. But their account fails to explicitly address my second and third desiderata. Saying that the cosmos is God’s prakṛti doesn’t specify in what sense, exactly, the cosmos overlaps with, and is ontologically produced by, God. So, while I think Silvestre and Herbert’s account is on the right track, my account in the next section provides much-needed supplementation.

Grounding panentheism

Earlier, I proposed three desiderata that a minimal account of panentheism must adhere to. The panentheist’s God-world relation must be (1) asymmetric, (2) quasi-mereological, and (3) generative or ontologically productive. My view is that the relation of metaphysical grounding meets all three conditions. Panentheism, on my account, is the view that God’s existence grounds the existence of the cosmos.

Grounding is a non-causal form of metaphysical explanation and dependence.Footnote ⁸ While there are different approaches to grounding in the literature, my approach here understands grounding as a relation that obtains between facts. Grounding is non-causal in the sense that when some fact, f, grounds another fact, g, f doesn’t cause g to occur. Rather, the relationship between f and g concerns the constitutive generation of a dependent outcome (Schaffer Reference Schaffer2017, 305). In this sense, grounding is often distinguished from causation via a levels metaphor: grounding relates facts synchronically or non-temporally in a ‘vertical’ manner through levels of reality, while causation relates events diachronically in a ‘horizontal’ manner. As a result, grounding is also associated with a notion of absolute and relative fundamentality, whereas causation is not.Footnote ⁹ Those facts that are ungrounded are absolutely fundamental. And insofar as some fact, f, grounds another fact, g, f is more fundamental than, or ontologically prior to, g.

As I mentioned earlier, grounding is similar to ontological dependence. Both are non-causal relations, and both involve a kind of constitutive metaphysical dependence. But there are important differences. First, most assume that grounding is a primitive notion, and so opt to leave it undefined. Ontological dependence relations, by contrast, are typically defined notions. Second, many grounding enthusiasts believe grounding is a unified notion.Footnote ¹⁰ Ontological dependence, by contrast, is comprised of a family of such relations (Tahko and Lowe Reference Tahko, Lowe, Edward and Nodelman2025). Granted, ontological dependence does seem appropriate for characterizing panentheism in some respects. Various forms of ontological dependence are, for example, asymmetric. But it’s less clear whether ontological dependence relations are ‘quasi-mereological’ and generative in the relevant sense that grounding is. And given the vast array of different forms of ontological dependence, surveying each would simply take us too far afield. As I argue below, grounding uniquely meets all three desiderata outlined above.Footnote ¹¹ I address each of my proposed desiderata in order.

Asymmetry

First, there is a general, though not unanimous, consensus that grounding is an asymmetric relation. The standard or orthodox view is that grounding forms a strict partial order: it is asymmetric (anti-symmetric and irreflexive) and transitive (Raven Reference Raven2013). Several thinkers have challenged this view. For example, Gonzalo Rodriguez-Pereyra (Reference Rodriguez-Pereyra2015) argues grounding is not a strict partial order. Ricki Bliss (Reference Bliss, Bliss and Priest2018) maintains that grounding can be reflexive. And Naomi Thompson (Reference Thompson and Jago2016) argues that grounding can be non-symmetric.

While it is beyond the purview of this article to defend the asymmetry of grounding, I’ll simply note that denying the orthodox view of grounding as a strict partial order is incompatible with grounding’s theoretical roles. As David Kovacs (Reference Kovacs2018) argues, grounding’s theoretical utility is comprised of both an explanatory and structural role. To focus on grounding’s structural role, it is understood in terms of the notion of relative fundamentality. If f grounds g, then f is more fundamental than g. This imposes constraints on ground’s logic since more fundamental than is asymmetric (and transitive). For example, reflexive instances of grounding, where some fact f grounds g and f = g, would allow facts to be more fundamental than themselves. And anti-symmetric instances of grounding would allow facts to be both more and less fundamental than each other. Since an asymmetric notion of ground is introduced to capture the phenomenon of relative fundamentality in the first place, this rules out symmetric and reflexive cases.

While much more could be said here, getting any further into this issue would take us too far afield. So, I’ll simply assume the orthodox position moving forward. This conception of metaphysical grounding meets our first desideratum. God’s existence grounds the existence of the cosmos, but not vice versa.

Overlap

Second, there is a plausible sense in which grounding entails a non-mereological notion of overlap. In a technical sense, mereological overlap involves part-sharing. So, objects A and B overlap in a strict sense if they share a part with each other. One can also understand overlap in terms of spatiotemporal coincidence or colocation. For example, the statue and the clay out of which it is made are spatiotemporally coincident, and so overlap in that sense. But groundees arguably overlap their grounds in a non-spatiotemporal and non-mereological sense.Footnote ¹² Paul Audi characterizes this non-mereological notion of overlap as ontological overlap. On his suggestion, when A and B ontologically overlap, A and B ‘share their being’ (Audi Reference Audi2020, 579).

Audi (Reference Audi2020, 571) distinguishes between a making-conception of grounding and a building conception, and associates ontological overlap with the latter. On the making-conception, grounding is analogous to causation or what’s sometimes called ‘metaphysical causation’.Footnote ¹³ On the building-conception, grounding is more like constitution. Causation and constitution are ‘different breeds’, as Audi puts it. We tend to think of causation as a ‘horizontal’ relation and constitution as a ‘vertical’ relation. It’s for that reason that Audi suggests that the building conception of grounding, but not the making conception, requires the notion of ontological overlap. Grounding on the building conception is a relation of ontological priority that ‘builds’ or orders reality in a hierarchical manner. The lower-level more fundamental entities ground the higher-level less fundamental entities.

So, the core idea of ontological overlap is that when f grounds g, g overlaps with f in the sense that g derives its very being or existence from f. In some important sense, the being or existence of g is shared with f and so isn’t anything ‘new’. Contemporary metaphysicians working with the notion of metaphysical grounding often invoke various metaphors to characterize this idea. For example, many invoke a creation metaphor to illustrate the relationship between fundamental and derivative entities. Sara Bernstein (Reference Bernstein2016, 25–26) writes,

The idea here is that the non-fundamental or derivative facts ‘come for free’ because their being or existence is already accounted for by the fundamental facts, which they are derived from. And Ricki Bliss (Reference Bliss2019, 359–360) appeals to a seed metaphor, writing, ‘God, in His infinite cleverness, realised that He could save Himself a whole lot of work if contained within a small number of particularly important entities were the seeds for all the rest of it’. For Bliss, the derivative entities – ‘all the rest’ – are contained within the fundamental entities as ‘seeds’.

The notion of ontological overlap, therefore, nicely captures the sense in which the cosmos is in God that avoids the difficulties associated with a technical or strict mereological interpretation. To recall Crisp’s words from earlier, most panentheists would say something to the effect that ‘God does not create a world outside of Godself; he does not bring about something entirely distinct from Godself. Rather, he (somehow) “makes room” within himself for the created order’ (Reference Crisp2019, 32). Since panentheism, on my account, is just the view that the existence of the cosmos is grounded in God’s existence, it follows that the cosmos is in God in the sense that the existence or being of the cosmos ontologically overlaps with God’s. God doesn’t create anything entirely distinct from himself because the cosmos’s being simply is God’s being. To put it differently, when God creates, there are more beings, but there is not more being.Footnote ¹⁴ This implies that the distinction between God and creation is blurry for the panentheist. And I think this is the correct result given that many theists have been intent on affirming an absolute or sharp ontological difference between God and creation.

However, moving away from a technical notion of mereological overlap towards a less technical notion of ontological overlap does result in a loss of clarity. The idea of two things sharing their being is admittedly vague and metaphorical. And I’m not ultimately sanguine that we can get beyond these metaphors and offer a more literal definition of the notion. While some fault panentheism in this regard, I think this criticism is unwarranted. I argue in the conclusion that appeals to metaphor are unavoidable if we are to faithfully characterize the essential core of what panentheism is supposed to be.

Generation

Finally, many hold that metaphysical grounding is ontologically generative or productive. For Jonathan Schaffer (Reference Schaffer2010, 345), when x grounds y, y depends for its very nature and existence upon x. Kelly Trogdon (Reference Trogdon, Bliss and Priest2018, 189) writes, ‘a relation is generative just in case its instantiation brings things into existence. Grounding is generative given that grounded entities exist due to grounding’. And Sara Bernstein (Reference Bernstein2016, 23) says, ‘production does play an implicit role in concepts and elucidations of grounding which take it to be a kind of synchronic generation or “bringing into existence.” Production undergirds a “thick” concept of grounding according to which grounders transfer being to their groundees.’ Of course, this conception of grounding is controversial. Some thinkers, like Shamik Dasgupta (Reference Dasgupta2017) for example, advocate for a more deflationary conception of grounding. But again, wading into this debate here would take us too far afield. So, I’ll simply assume an inflationary conception of grounding in what follows.

This notion of ontological production or generation corresponds to Audi’s ‘building’ conception of grounding that I highlighted earlier. Indeed, Karen Bennett makes this notion of generation a condition for what it is for something to count as a building relation at all. On her view, all building relations are (1) directed (they are anti-symmetric and irreflexive), (2) necessitating, and, most relevant for our purposes, (3) generative, ‘in that the builders generate or produce what they build. Built entities exist or obtain because that which builds them does’ (Bennett Reference Bennett2017, 32). On Bennett’s view, grounding is a building relation and so is generative in the above sense.

If God’s existence grounds the existence of the cosmos, then our inflationary view of grounding entails that the very existence of the cosmos is produced or generated from the existence of God. As I mentioned earlier, this third desideratum is a corollary to the second one regarding overlap. If God and the cosmos share their being or existence, this naturally entails that the cosmos’s existence is produced by, or derived from, God’s. But to reiterate my earlier point, it’s important to state our third desideratum explicitly since specifying the panentheist’s God-world relation as a generative one precludes relations that relate distinct entities whose existence is logically independent of their standing in that relation to begin with. And like the notion of ontological overlap, ontological production or generation is admittedly vague and metaphorical. But I take it this notion of ontological production is just constitutive of what grounding is on the building conception. And since grounding is a primitive notion, the corresponding notions of ontological overlap and production are also primitive notions.

On my account, therefore, panentheism is the view that God’s existence grounds the existence of the cosmos. We’re now in a better position to understanding the panentheist’s slogan that the cosmos is in God but God is more than the cosmos. The cosmos is ‘in’ God in the sense that it ontologically overlaps, and so shares its being or existence, with God. But God is ‘more than’ or transcends the cosmos in virtue of being ungrounded or fundamental. Since grounding is an asymmetric relation of metaphysical dependence, the cosmos depends upon God for its existence but not vice versa. And since grounding is ontologically productive, God is the ultimate source of the very existence or being of the cosmos.

However, it isn’t initially clear how this understanding of panentheism results in a unique and rival view of God’s relationship to the cosmos. Indeed, at first glance, my account of panentheism looks indistinguishable from theism. I turn now to the demarcation problem.

The demarcation problem

An adequate account of panentheism needs to demarcate the view from rival forms of theism. Mullins (Reference Mullins2019, 204) has been a vocal critic of panentheism, arguing that ‘panentheism cannot distinguish itself from well-established and clearly articulated rival models of God such as classical theism, neo-classical theism, open theism, and pantheism’.Footnote ¹⁵ I don’t have the space here to offer a fully worked-out solution to the demarcation problem with respect to each rival model of God that Mullins identifies. Instead, I focus on pantheism and theism as rivals to panentheism. As we’ll see, demarcating pantheism from panentheism is straightforward. Not so for theism. In this section, I argue we can demarcate theism from panentheism vis-à-vis the doctrine of creation ex nihilo. While addressing all the relevant issues of this proposal in sufficient detail would take me too far afield, I believe laying out the broad contours will suffice to show that my proposal is a plausible avenue for panentheist’s to explore further.

First, my account of panentheism above offers a clear path to demarcating panentheism from pantheism. On my view, panentheism is the view that God’s existence grounds the existence of the cosmos. Pantheism, by contrast, maintains that God is numerically identical with the cosmos. While grounding is a strict partial order (asymmetric and transitive), identity is an equivalence relation (symmetric, reflexive, and transitive). The logical incompatibility of the formal properties of the grounding and identity relations precludes panentheism from collapsing into pantheism.

Second, I believe the doctrine of creation ex nihilo can help to illuminate the difference between panentheism and theism. Theists are committed to creation ex nihilo. As we’ll see below, I argue panentheists are not. Mullins isn’t sanguine about this approach. He offers, what is in my view, the following non-standard definition.

On this definition, creation ex nihilo means God and the created order are not co-eternal. Mullins maintains that panentheists reject creation ex nihilo, and so are committed to what he calls The Necessity of the Cosmos: God must exist with a universe or multiverse of some sort.Footnote ¹⁶ But he asserts that ‘nothing about affirming the necessity of the cosmos tells us anything terribly interesting or unique about the nature of God, which is precisely what a model of God needs to do’ (Mullins Reference Mullins, Göcke and Medhananda2023, 242). I think the necessity of the cosmos is irrelevant for the demarcation problem. Working with a standard conception of creation ex nihilo will arguably provide the resources for demarcating panentheism from theism.

What I call ‘the standard’ understanding of creation ex nihilo, at least in the Christian tradition, involves God creating the cosmos from no pre-existing ‘stuff ‘. As Wes Morriston (Reference Morriston2002) notes, the doctrine is often stated in Aristotelian terms. God is the efficient cause of the universe. But there is no material cause, no ‘stuff out of which God creates. Morriston (Reference Morriston2002, 24) defines creation ex nihilo as follows: ‘x is created ex nihilo by y if and only if i) y causes x to exist, and ii) y does not cause x to exist by transforming some other material stuff ‘. And Julius Lipner (Reference Lipner1978, 54) describes the standard interpretation of creation ex nihilo as the ‘thrusting into being, so to speak, of a reality not existing qua being hitherto … of being that had not pre-existed or remained hidden qua being before the creative act (except in the loose and related senses of being objectively possible to God and existing in him as seminal ideas)’. So, I think God’s freedom to create and God’s creating ex nihilo are independent issues. If we affirm that the cosmos has a material cause, I don’t see how it follows that God must exist with a cosmos, contrary to Mullins’s claim.

However, the notion of material causation that is operative in creation ex nihilo requires clarification. ‘Material causation’ is a placeholder for a more precise notion because the terms making and building, which are standardly invoked to characterize material causation, are ambiguous. As Bennett (Reference Bennett2017) argues, there are many different relations that capture the sense in which one thing is built out of or made from another. As I mentioned earlier, she calls them ‘building relations’ and identifies at least six candidates including set-formation, micro-based determination, constitution, realization, composition, and grounding. My contention is that ‘material causation’ most plausibly denotes the grounding relation due to both relations’ link with metaphysical or ontological priority.

Unlike efficient causation, material causation is presumably a kind of constitutive relation, at least in the sense that it is synchronic rather than diachronic. While we typically think of efficient causes as temporally prior to their effects, material causes are ontologically prior. For example, Felipe Leon (Reference Leon and Szatkowski2024, 293) says, ‘by material cause, I mean the temporally or ontologically prior things or stuff from which (though not necessarily of which) a thing is made’. So, unlike efficient causation, material causation is a kind of ‘vertical’ dependence relation.Footnote ¹⁷ And since grounding just is a constitutive or ‘vertical’ relation of ontological priority, it’s the best candidate out of Bennett’s building relations for clarifying what material causation is. On this view, saying that the pieces of wood are the material cause of the desk is equivalent to saying that the existence of the desk obtains in virtue of, or is grounded in, the existence and organization of its constituent materials.

We can then interpret ex nihilo in a non-causal or grounding sense. For example, when we say, ‘Nothing comes to be from nothing’, we can mean that nothing comes to exist without a ground. So, I propose the following definition of creation ex nihilo.

On this definition, God’s creation of the cosmos ex nihilo entails that the existence of the cosmos is fundamental.Footnote ¹⁸ God is the efficient cause of the cosmos at a finite time in the past. But God creates out of nothing in the sense that neither his existence nor the existence of anything else grounds the cosmos’s existence. In other words, there are no underlying facts or entities in virtue of which he creates. Theists are, of course, committed to creation ex nihilo. But since panentheism just is the view that God’s existence grounds the existence of the cosmos, panentheism therefore rejects creation ex nihilo as I’ve formulated it above. My grounding interpretation of creation ex nihilo, therefore, provides a way to demarcate theism from panentheism.

A pressing issue for my proposal is that it forces theists to deny what I’ll call ‘the Grounding Thesis’, namely, that God’s existence grounds the existence of the cosmos. On my definition, there can be nothing at all whose existence grounds the existence of the cosmos, neither God nor anything else. Creation ex nihilo, on this strong view, contradicts the principle, ex nihilo nihil fit: nothing comes from nothing.Footnote ¹⁹ But this is problematic given that many theists do affirm the Grounding Thesis. For example, Kenneth Pearce (Reference Pearce and Jonathan2017) argues that God is the foundational ground of the entire past causal history of the universe. Einar Bohn (Reference Bohn2018, 2) formulates a view called ‘Divine Foundationalism’ as ‘Anything distinct from God is existentially grounded by God’.Footnote ²⁰ One can also find affirmations of something like the Grounding Thesis among historical thinkers. For example, because Aquinas thinks created beings don’t exist by their very nature, they are therefore radically ontologically dependent upon God.

So, the theist might object that creation ex nihilo was never meant to contradict ex nihilo nihil fit. Instead, she may endorse a weak version of creation ex nihilo that is compatible with the Grounding Thesis, according to which x is created ex nihilo by y iff y is the efficient cause of x’s coming into existence and there is no z distinct from y such that z’s existence grounds x’s existence. On this definition, God creates the world ex nihilo so long as there is nothing distinct from himself that grounds the existence of the cosmos.

But this move seems to make creation ex nihilo synonymous with creation ex deo – creation out of or from God – which leads to an uncomfortably close relationship between God and cosmos for the theist. Augustine is sensitive to this issue. In his Confessions (Book 12, chapter 7) he writes,

Augustine’s worry is that if God creates out of himself or his substance, as the Grounding Thesis entails, then we blur the distinction between God and creation and thus undermine God’s transcendence. It seems Augustine is happy to affirm that God’s creation ex nihilo isn’t constrained by the principle, ex nihilo nihil fit since, as he says, there is nothing God cannot do.

For better or for worse, it looks like panentheism is the result of endorsing a weak version of creation ex nihilo and the Grounding Thesis. For example, Daniel Soars (Reference Soars2021) argues that, according to Aquinas (the paradigmatic traditional or classical theist), creation ex nihilo is really a form of creation ex deo and so doesn’t contradict the ex nihilo nihil fit principle.Footnote ²¹ Soars (Reference Soars2021, 957) then goes so far as to argue that it is only a short logical step from Aquinas’s views on causation, which in many ways resemble the notion of grounding, ‘to affirm that all created effects (viz. the world) must be pre-contained in their supreme cause (God) or, to put it in the slightly more daring terms not unknown to some medieval Christian mystics, that the world exists “in” God’. But that’s just panentheism. To avoid panentheism, the theist therefore needs to adopt my strong definition of creation ex nihilo and reject the Grounding Thesis.

This has revisionary implications for our categorization practices. So long as a philosopher or theologian affirms the weaker version of creation ex nihilo and thereby the Grounding Thesis, such that God creates out of or from himself, my account of panentheism says we should identify them as a panentheist despite any designation otherwise.