So you think you understand Existential Inertia?

What is existential inertia? Is it a thesis? A phenomenon? Why would anyone believe it? Why would anyone disbelieve it? Does it explain persistence, or merely describe it? Does it render persistence inexplicable or brute?

This post will answer all these questions and more. It’s a comprehensive guide to existential inertia. I’ll discuss (i) common mistakes people make when discussing existential inertia, (ii) what existential inertia is, (iii) motivations for existential inertia, (iv) arguments against existential inertia, and (v) resources on existential inertia. Buckle up.

The Beauty of Pi - The Flerlage Twins: Analytics, Data Visualization, and  Tableau

Outline

1 Common mistakes
2 The basics of existential inertia
3 Clarifying the Existential Inertia Thesis (EIT)
—–3.1 Scope
—–3.2 Persistence and Relativity Theory*
—–3.3 Modal Register
—–3.4 Dependence and Destruction
—–3.5 Metaphysical Accounts
4 A rigorous articulation of EIT*
—–4.1 Temporal Ontology*
—–4.2 Taxonomic questions: Summary
5 The metaphysics of EIT
—–5.1 Tendency-disposition Accounts
—–5.2 Transtemporal Accounts
—–5.3 Law-based Accounts*
—–5.4 Necessity Accounts
—–5.5 No-change Accounts
6 Motivating EIT
—–6.1 Theoretical Virtues
—–6.2 Aristotelian proof argument
—–6.3 Divine Temporality
—–6.4 Bayesian argument
—–6.5 Moorean argument
7 Arguments against EIT
—–7.1 Alleged Counterexamples
—–7.2 Red Chairs
—–7.3 Hsiao and Sanders
———-7.3.1 Counterexamples to EIT
———-7.3.2 First Metaphysical Argument
———-7.3.4 Third Metaphysical Argument
—–7.4 Proportionate Causality
—–7.5 Form-matter Interdependence
—–7.6 Contingent Natures
—–7.7 Vicious Circularity
—–7.8 Aristotelian proof causal principle
—–7.9 Neo-Platonic proof causal principle
—–7.10 Thomistic proof causal principle
—–7.11 Rationalist proof causal principle
—–7.12 Nemes’ argument
—–7.13 De Ente argument
—–7.14 Nemes and Kerr on the De Ente argument
8 Resources
—–8.1 Articles
—–8.2 YouTube videos
—–8.3 Blog posts

Notes:

First, an asterisk (*) indicates that the section is highly technical. If the reader does not understand the section, that is alright—he or she can move to the subsequent non-asterisked section without loss.

Second, as a tip for navigating this post, use your computer’s command F function and search for the relevant section(s) of interest from the above outline.

Third, I’ll be using footnotes in this post. The footnotes of a given section are at the bottom of the section as a whole. For instance, any footnotes within Section 3 (including 3.1 through 3.5) will be at the very bottom of Section 3 immediately before Section 4.

Fourth, please cite this post if you will be using its content in any way. Fair use rules apply to the whole post. 🙂

1 Common mistakes

Mistake #1: Ambiguity

One common mistake is to entirely overlook that ‘existential inertia’ is ambiguous. In particular, it is ambiguous between (i) the thesis of existential inertia, which is a proposition that purports to describe the way reality is and (ii) the phenomenon of existential inertia, which is the manner in which things persist according to the Existential Inertia Thesis (EIT). The thesis is just the claim that some temporal concrete objects persist in the absence of both external sustenance/conservation and sufficiently destructive factors. The phenomenon is the actual persistence of one or more temporal concrete objects without external sustenance and destruction. To refer to the thesis, I’ll typically use ‘EIT’. To refer to the phenomenon, I’ll typically use ‘inertial persistence’.

Mistake #2: Explanation

Another common mistake is thinking that EIT provides (or purports to provide) an explanation of persistence. This is not true. EIT merely purports to describe the manner in which some temporal concreta persist. EIT leaves open (i) whether there is an explanation for persistence and (ii) what that explanation is (if there is one). EIT is a descriptive thesis, not an explanatory thesis. But this doesn’t mean that EIT renders persistence (or inertial persistence) a brute fact. For EIT can be (and should be) supplemented with one or more metaphysical accounts of EIT. These accounts aim to provide an inertialist-friendly explanation of persistence, i.e., explanations of persistence on which EIT is (or could be) true.

Mistake #3: Bruteness

A third mistake is thinking that EIT renders persistence brute (i.e., unexplained or, perhaps, inexplicable). This is false. EIT merely purports to describe the way temporal concrete objects (or some subset thereof) persist. It entails nothing about whether there is an explanation of persistence and, if there is, what that explanation would be (apart from being compatible with EIT). EIT thus says nothing about whether persistence is brute. And as we will see later, there are whole swathes of inertialist-friendly explanations of persistence. (An ‘inertialist’ or ‘existential inertialist’ is one who accepts or leans toward the truth of EIT. An explanation is ‘inertialist-friendly’ if it explains persistence in a manner consistent with EIT.)

Mistake #4: It’s a property

A fourth mistake is thinking that existential inertia is a property. Ed Feser, for instance, has said this, and I have argued at length that he is mistaken here on many fronts. The first mistake in this thought is its failure to disambiguate ‘existential inertia’ (cf. Mistake #1). It could refer either to the thesis (EIT) or the phenomenon (inertial persistence). The former obviously isn’t a property. But the phenomenon of inertial persistence by itself doesn’t entail that there is some property exemplified by concrete objects corresponding to inertial persistence. What would such a property amount to? The property of being such as to persist in the absence of both external sustenance and sufficiently destructive factors? That’s a ludicrously gerrymandered property. There is no such property as this. Moreover, nothing in EIT as such entails that inertial persistence is a property or attribute exemplified by temporal concreta. And as we’ll see when discussing metaphysical accounts of EIT, it is simply false that EIT requires there to be some attribute or property corresponding to inertial persistence.

Mistake #5: Universal scope

A fifth mistake is thinking that EIT, if true, applies to everything—every event, state of affairs, property, object, and so on. But this is not true. EIT, as I and other authors articulate it, simply says that temporal concrete objects or some subset thereof persist in the absence of external sustenance and sufficiently destructive factors. This only quantifies over (i) a sub-set of (ii) temporal (iii) concrete (iv) objects.

Mistake #6: Atheistic

A sixth mistake is thinking that EIT is inherently atheistic. This is false. Suppose—along with neo-classical theists—that God is temporal. And suppose God creates and sustains everything apart from himself. In this case, EIT is true: God is a temporal concrete object that persists without being externally sustained and without being destroyed. Hence, some subset of temporal concrete objects persist in the manner EIT describes. Hence, EIT is true. Hence, EIT is not inherently atheistic. EIT only debars a scenario in which every temporal concrete object is sustained or conserved in being from without. This is perfectly compatible with theisms that render God temporal, and it’s also compatible with theisms that afford God no role in explaining the continued existence of (some) temporal concrete objects.

Mistake #7: Burden of proof

A seventh mistake is thinking that defenders of EIT always have a burden to positively justify the truth of EIT. This is false. Whether the defender of EIT has a burden to positively justify the truth of EIT depends on the dialectical context at hand. If the context is one wherein someone is proffering a positive argument for classical theism one of whose premises requires the falsity of EIT, all the defender of EIT needs to do is argue that nothing in the argument—that is, nothing in the relevant premise or what is said on its behalf—gives those who accept (or are neutral on) EIT sufficient reason to abandon their position. The onus, in this context, is not on the defender of EIT to positively justify EIT and prove that the premise in question is false (by proving EIT true). They need only show that the argument (and what’s said on its behalf) fails to prove EIT false.


2 The basics of existential inertia

Here’s the simplest way to articulate EIT: temporal concrete objects or some sub-set thereof persist in the absence of both (i) sustenance or conservation from without and (ii) sufficiently destructive factors operative. An object that inertially persists, then, is one that persists in the absence of both external sustenance and sufficiently destructive factors.

But reality isn’t simple. Complications arise.  Different authors characterize the thesis differently. We’ll explore those below. For the next few paragraphs, I just want to give you a sense of what EIT says and what it doesn’t say.

In simplest terms, EIT is the claim that at least some temporal concrete objects persist in the absence of both (i) sustenance or conservation from without and (ii) sufficiently destructive factors that would destroy the object(s). EIT does not aim to answer that in virtue of which objects persist; instead, it purports merely to describe the way they persist. EIT can (and should) be supplemented with an answer to the aforementioned question. Such answers represent inertialist-friendly metaphysical accounts of persistence.

Here is a slightly more formalized articulation of EIT. This more formalized articulation is cast I terms of endurantism. But nothing hangs on this—a similar thesis can be developed mutatis mutandis for perdurantism. [I develop such theses in Section 4 below.] Here, then, is the thesis:

Existential Inertia Thesis (EIT): For each member O of some (proper or improper) sub-set of temporal concrete objects and for each time t such that O exists at some time t* earlier than t, (i) at t, O does not ontologically depend on the existence or activity of some concrete object O*, where O* is not a (proper or improper) part of O, and (ii) if O is not positively destroyed within the temporal interval [t*, t], then O exists at t.

First, some definitions. To positively destroy O at time t is to actively bring about O’s cessation at t, such that O endures through [t*, t) but not [t*, t], where t* < t. Ontological dependence is a hierarchical or concurrent dependence of a less fundamental object on the existence or activity of another, more fundamental object. (Examples include sustaining or conserving efficient causation, grounding, and functional realization.)

Second, some notes. First, my EIT quantifies only over a sub-set of temporal concreta. This allows the inertialist to affirm that at least some temporal concreta (e.g., perhaps non-fundamental, composite physical objects) are sustained or conserved in being from without. Second, my EIT only specifies that every inertially persistent object O does not ontologically depend on concrete objects that are not parts of O. This allows for views on which composite objects ontologically depend on their parts.

Third, a simpler statement of EIT. The thesis is simply that one or more temporal concrete objects persist in existence in the absence of both (i) concurrent sustenance or conservation from without and (ii) factors sufficient to positively destroy the object(s).

Now we can survey other authors’ articulations of EIT. Warning: they vary wildly. In Section 3, I’ll bring significant clarity to these variations. For now, though, I want to give you a survey.

As Beaudoin characterizes existential inertia, “[a]n object enjoying existential inertia will continue to exist, without being sustained in existence by any external agent, until something else comes along and destroys it” (2007, p. 86). Along similar lines, Adler views the thesis as affirming that “[c]ontingent individuals continue in existence, once given existence, until counteracting causes intervene to deprive them of their existence” (1980, p. 125). Feser follows suit: “the world of contingent things, once it exists, will tend to continue in existence on its own at least until something positively acts to destroy it” (2011, p. 239).

Audi’s characterization differs from the above: “Whatever exists is poised to continue to exist if undisturbed and inactive” (2019, p. 2), where x’s being poised to A implies that “there are circumstances C such that if x is in C, x will A” (2019, p. 4). Audi continues: “By ‘undisturbed’, I mean not being causally influenced by anything else. By ‘inactive’, I mean not undergoing internal causal processes” (2019, p. 4). Moreover, for Audi, the domain over which the thesis quantifies is unrestricted: natural or supernatural, structurally simple or complex, fundamental or non-fundamental, necessary or contingent.

Schmid’s (2021) characterization [yep, that’s a me, Mario] differs further still in terms of the domain of quantification and modal register:

Existential Inertia Thesis (EIT): Necessarily, concrete objects (i) persist in existence (once in existence) without requiring a continuously concurrent sustaining cause of their existence and (ii) cease to exist only if caused to do so. (Schmid 2021, p. 203)

Given the wide variety of (incompatible) characterizations of the inertial thesis in the literature, much conceptual clarity is needed in the debate. In particular, both defenders and detractors of EIT need clarity on each of the following: (i) the thesis’s domain of quantification, (ii) the thesis’s modal register, (iii) the kind of ontological dependence the thesis denies of inertially persistent entities, (iv) the factors that do or can destroy inertially persistent entities, (v) that in virtue of which inertial persistence obtains (if inertial persistence obtains at all), and (vi) the thesis’s relation to temporal ontology. Surprisingly little attention in debates concerning existential inertia has been paid to disentangling and precisifying each of these distinct elements of the inertial thesis. That’s what I’ll redress in the next section.

3 Clarifying the Existential Inertia Thesis (EIT)

My purpose in this section is to uncover a series of questions to clarify and taxonomize inertial theses. While I’ll take sides on some of these questions, my main focus is the taxonomic questions themselves, since the taxonomic questions serve to bring clarity and precision to the existential inertia debate. The taxonomic questions also lay the foundation for our characterization of existential inertia in Section 4.

3.1 Scope

What kinds of ontological items inertially persist? Audi’s inertial thesis, as we’ve seen, is completely unrestricted in scope: whatever exists is poised to continue to exist if undisturbed and inactive. This seems implausible, however. Suppose abstracta (non-spatiotemporal, acausal entities like universals, propositions, mathematical objects, etc.) exist. Because abstracta don’t enter into causal relations, they are undisturbed and inactive. Audi’s unrestricted thesis therefore entails that such abstract objects are poised to continue or persist in existence. But surely this is a category error—abstract objects, qua non-temporal, cannot ‘persist’ or ‘continue’ in existence. Persistence is similarly a category error if applied to non-temporal concreta (if they exist).

We therefore have the first two scope questions that any inertial thesis must answer: Does the thesis quantify over temporal or non-temporal things (or both)? Does the thesis quantify over concreta or abstracta (or both)? For reasons just articulated, I think the thesis should quantify only over temporal (or spatiotemporal) concreta.

Assuming we quantify over temporal concreta, a third scope question arises: what kind of temporal concreta—objects, events, concrete states of affairs, etc.—inertially persist? Beaudoin quantifies both over concrete objects and ‘the world’, while both Adler and Schmid quantify solely over concrete objects (i.e., individuals like trees, particles, etc.). If events, concrete states of affairs, etc. exist in addition to concrete objects, then Audi quantifies over them as well.

By my lights, the standard differences between concrete objects and events renders this latter commitment of Audi’s thesis implausible. At least in debates concerning inertial persistence, objects are standardly pictured as enduring continuants that persist by being wholly present at each time at which they exist. But events are not like this.[1] Standard accounts of events picture them as perduring occurents that take up (rather than persist as wholes throughout) time and have various parts or stages at different times (Casati and Varzi 2020; Simons 2000; Mellor 1980). Thus, events as such (standardly conceived) are not candidates for continuance or persistence over time.[2] Similar points apply to concrete states of affairs. For this reason, I restrict my domain of quantification solely to concrete objects.

A fourth scope question concerns whether all temporal concrete objects inertially persist or only some proper subset thereof. In particular, should the thesis only quantify over the most fundamental (i.e., basic, foundational, ultimate, not explained in terms of some further underlying reality) temporal concrete objects? Both Beaudoin (2007, p. 86) and Schmid (2021, fn. 6) express the possibility of the inertial thesis quantifying solely over the fundamental components (whatever they may be) of temporal concrete reality.

In principle, then, the inertial thesis leaves open whether it quantifies over all temporal concrete objects or instead only some subset thereof. This leaves open potential explanatory relations (grounding, efficient sustaining causation, realization, etc.) that might obtain between the fundamental or foundational temporal concrete objects and the non-fundamental, non-foundational ones. Consequently, the inertial thesis, by itself, is compatible with some concurrent ontological dependence among temporal concrete objects. As we shall see, my articulation of the inertial thesis will leave the fourth scope question open.

Note, moreover, that nothing in the inertial thesis demands quantification solely over material objects. Perhaps neutral monism is true, or perhaps there exist non-physical temporal concrete objects (like a panentheistic or neo-classical theistic God, say). I take no stance on such matters here. But so long as such temporal objects persist in existence without some kind of conserving activity that continuously keeps them in being, they inertially persist. Existential inertia is fundamentally about continuance in existence in the absence of both (i) conserving or sustaining activity from without and (ii) sufficiently destructive factors. The inertial thesis thus applies to temporal concrete objects satisfying (i) and (ii) regardless of whether those objects are material or immaterial, physical or non-physical, necessary or contingent.[3]

Disentangling these scope questions (and the questions in the subsequent subsections) has largely been overlooked in the literature thus far. This is unfortunate, since ruling out or justifying one inertial thesis doesn’t automatically rule out or justify other inertial theses.

3.2 Persistence and Relativity Theory*

  • Thanks to philosopher of physics DJ Linford for co-authoring this section as well as parts of Section 4. Fair use rules apply to the whole post (i.e., cite the post if you’re going to use its content). For this section, I’ll use ‘we’ and ‘our’, since this is co-authored with DJ.

So much for taxonomic questions concerning scope. We turn now to another series of taxonomic questions pertaining to the nature of persistence. In particular, we will ask: what does persistence amount to? Will the existential inertia thesis understand it in endurantist terms, perdurantist terms, or something else altogether? How does persistence (and, thus, existential inertia) relate to the findings of contemporary physics?

To explore how one might update the existential inertia thesis for relativistic space-times, we begin by introducing a three-fold distinction between two non-relativistic space-times, i.e., Newtonian and Galilean space-times, and one simple relativistic space-time, i.e., Minkowski space-time. In Newtonian space-time, at each time, the universe is a three-dimensional Euclidean space consisting of a collection of objectively simultaneous spatial points. The spatial points persist over time; moreover, objects can be at absolute rest by occupying the same spatial point(s) at successive times. In Galilean space-time, the universe, at each time, is again a three-dimensional Euclidean space composed of a collection of objectively simultaneous spatial points. However, absolute rest does not exist in Galilean space-time; for that reason, spatial points do not persist through time, even though material objects may. On either Newtonian or Galilean space-time, material objects can either be wholly located at each of the times that they exist and so endure or have temporal parts located at each of the times at which they exist and so perdure. 

The simplest relativistic space-time is Minkowski space-time, that is, a relativistic space-time without gravitational effects. Like Galilean space-time, the points of Minkowski space-time are not typically thought to persist because relativity does not allow for absolute rest. However, in contrast to either Newtonian or Galilean space-times, the standard conception of Minkowski space-time holds that there are no objective simultaneity relations between any two numerically distinct space-time points. And since there are no objective simultaneity relations between any two numerically distinct space-time points, we should not think of Minkowski space-time as being objectively divisible into three-dimensional spaces located at successive times.

In Newtonian and Galilean space-time, relations of objective simultaneity allow one to define absolute time. That is, two objects that are objectively simultaneous are said to be at the same time. But without objective simultaneity, no two numerically distinct objects can be objectively said to occupy the same time. Leaving aside the possibility of extended simples, any non-point-like material object O has some spatiotemporal extension and so has some collection of numerically distinct proper parts. If no two numerically distinct objects can objectively be said to occupy the same time, no two numerically distinct proper parts of O can objectively be said to occupy the same time.

Standard conceptions of endurance and perdurance are not obviously compatible with relativity. In order for O to objectively and wholly occupy a time—as is required on a standard conception of endurance—all of O’s proper parts need to occupy the same time, that is, all of O’s proper parts need to be objectively simultaneous with each other. On the other hand, on a standard conception of perdurance, while O is a space-time worm, each of O’s temporal parts occupies a distinct time. In order for a temporal part p of O to occupy a time, the proper parts of p need to be located at the same time, that is, to be objectively simultaneous with each other.

In order to introduce relativistic conceptions of endurance and perdurance, we need to first place some conceptual machinery on the table. Following Gilmore (2008), we can define a binary locational predicate ‘is weakly located at’. For O to be weakly located at a space-time region R is for O to spatiotemporally overlap R, that is, for R to not be completely free of O. As Gilmore defines it, R is O’s path just in case R “has a subregion in common with all and only those regions at which the object is weakly located” (2008, p. 1228). Or, as Calosi and Fano describe, “the path of an object is simply the union of its exact locations” (2015, p. 287). For the perdurantist, O’s path will turn out to be the space-time region occupied by O’s space-time worm, whereas for the endurantist, O’s path will turn out to be the mereological sum of all the space-time regions that O occupies throughout O’s life.

Gilmore proceeds by employing the notion of an achronal surface (Gilmore 2008, p. 1228). In relativistic space-times, any space-time point A can bear one of three relations to any other space-time point B: A is light-like related to B just in case a signal moving at the speed of light can propagate from A to B or from B to A; A is time-like related to B just in case a signal moving slower than the speed of light can propagate from A to B or from B to A; and A is space-like related to B just in case a signal cannot propagate from A to B or from B to A. A surface is achronal just in case any two numerically distinct points on the surface are space-like related to each other (for a formal definition, see Calosi and Fano 2015, p. 287). Here, we use the term ‘surface’ in a fairly general way; an achronal surface can be a three-dimensional slice of the space-time block. In fact, achronal surfaces are, in some sense, the relativistic analogue of the three-dimensional spaces located at distinct times that are included in Newtonian and Galilean space-times. For Gilmore, O persists just in case O has a path that is not achronal. Put another way, O persists just in case O’s path includes two space-time points that are either time-like or light-like related.

We can now define an achronal part P of O such that “(i) P’s path is an achronal slice of O’s path, (ii) P and O are co-composed at P’s path, and (iii) P is a part of O at P’s path” (Gilmore 2008, p. 1228). (For an alternative definition in terms of the intersection of a Cauchy surface with O’s path, see Calosi and Fano 2015, pp. 287-288.) O can then be said to mereologically perdure just in case O has a “sufficiently full distribution” of achronal parts, e.g., “perhaps one for each achronal slice of its path, perhaps one for each continuous temporal ‘chunk’ of its path, perhaps even one for each sum of achronal slices of its path” (Gilmore 2008, p. 1228) or, as on Calosi and Fano’s version, one for each intersection of a Cauchy surface with O’s path. Moreover, O locationally perdures just in case O is exactly located at the space-time regions where O persists. We can also define mereological endurance, i.e., O mereologically endures just in case O persists but does not have such a distribution of achronal parts. (For the sake of intuition, compare what we’d say in Newtonian or Galilean space-time: O mereologically endures just in case O has all of O’s parts at each of the times that O exists instead of having a scattering of parts across space-time.) Likewise, we can define locational endurance, i.e., O locationally endures just in case all of O’s exact locations are achronal regions.

Intuitively, fully fleshed out and plausible versions of mereological and locational perdurance should be available. We can say that objects should be identified with their space-time worms and that they are exactly located at their paths. How to fully flesh out and render plausible either mereological or locational endurance is not as obvious. One can render perdurance or endurance consistent with the empirical adequacy of relativity by either (i) replacing perdurance or endurance with appropriate relativistic analogues, (ii) replacing relativity with an empirically equivalent theory that includes relations of objective simultaneity, or (iii) introducing additional structure into relativistic space-times in virtue of which one can define objective simultaneity relations. Each of these three possibilities faces formidable difficulties that often go underappreciated in philosophy of religion. For example, William Lane Craig (2017, 2008, 2001b, 2001c, 1999, 1990) has suggested replacing relativity with Neo-Lorentzianism, a theory supposedly empirically equivalent to Special Relativity. Neo-Lorentzianism restores three-dimensional spaces located at distinct absolute times. And if Neo-Lorentzianism is true, the endurance of objects is no more problematic than the endurance of objects in Galilean or Newtonian space-time.

However, if Neo-Lorentzianism is true, then space and time only appear, and do not actually, satisfy Lorentz invariance, a symmetry principle central to Special Relativity. Although Lorentz invariance itself cannot be directly experimentally or observationally verified, Lorentz invariance has motivated highly successful work in particle physics, including the development of quantum electrodynamics. Quantum electrodynamics has enjoyed incredible empirical success; as one example, quantum electrodynamics predicts a correction to the magnetic moment of the electron that has been experimentally verified to within around one part in a trillion (Gabrielse, et al, 2019; Hanneke, et al, 2008). Never mind precisely what Lorentz invariance or the magnetic moment are; our point is that quantum electrodynamics makes the most accurate prediction in the history of science. This degree of empirical success would be very surprising if Neo-Lorentzianism turned out to be true.

As part of their defense of the A-theory of time and an absolute time parameter, other authors (e.g., Mullins 2016b, p. 34; DeWeese 2004; Craig 2001a, 2001b) have pointed out that Special Relativity has been succeeded by General Relativity, that a specific class of General Relativistic space-times—Friedmann-Lemaitre-Robertson-Walker (FLRW) space-times—appears to pick out a preferred way to cut space-time into three-dimensional surfaces, and that the observable universe appears to at least approximate an FLRW space-time. Each of the three-dimensional surfaces can then (supposedly) be interpreted as a moment of absolute time and to pick out a collection of space-time points, all of which are (supposedly) objectively simultaneous. If this proposal succeeds, then endurance would again turn out to be no more mysterious than endurance was in Newtonian or Galilean space-time.

However, this proposal encounters a series of problems. For example, FLRW space-times require that the distribution of matter-energy within the universe is homogeneous and isotropic. The matter-energy density increases dramatically between points outside and points inside the Earth; clearly, our universe is not homogeneous and isotropic. Instead, our universe is only homogeneous and isotropic on average.[4] But endurance was supposed to be a theory about the persistence of material objects and not a theory that held true only for the large-scale statistical properties of the universe.

There are proposals for how non-homogeneous and non-isotropic space-time might be sliced into some preferred collection of three-dimensional surfaces; for example, there are some theoretical reasons for thinking that space-time should be sliced into surfaces of Constant Mean (extrinsic) Curvature (CMC) (Crisp 2012; Crisp 2008; Monton 2006; Pitts 2004; Saunders 2002, p. 290; Craig 2001b, p. 236; Valentini 1996, p. 60). Nevermind precisely what the CMC slicing amounts to (though, see Lockwood 2007, pp. 118-120 for a non-technical introduction); for our purposes, it suffices merely to note that there may be empirical reasons for denying that our universe can be sliced into CMC surfaces.[5] (For example, Michael Lockwood (2007, p. 152) argues that insofar as we have indirect evidence for black hole decay via Hawking radiation, we have evidence that space-time cannot be globally sliced into CMC surfaces.)

With taxonomic questions concerning persistence thus covered, we turn to our next set of taxonomic questions concerning modal register.

3.3 Modal Register

Here’s the first modal question: is the inertial thesis necessarily true if true at all? Explicit inclusion of the necessity operator is only found in Schmid’s EIT. And while I agree with Schmid (myself, lol) that “because existential inertia and existential expiration are (or would be) such broad, foundational metaphysical features of reality, it seems that they would necessarily obtain if they obtain at all” (2021, p. 202), I will here leave open the modal register of our thesis.

The second modal question is: do the temporal concrete objects within the domain of quantification actually persist without continual sustenance, or do they merely persist without requiring continual sustenance? The former entails the latter, but not vice versa. Recall that Schmid’s EIT only says that objects persist without requiring sustaining causes. Importantly, though, this is perfectly compatible with actually having sustaining causes. Schmid’s EIT therefore does not rule out (at least in principle) a scenario in which all temporal concrete objects are continuously conserved or sustained in being from without. In other words, the truth of Schmid’s EIT is compatible with nothing inertially persisting. This result may be unpalatable for inertialists who want their thesis—if true—to secure the actual inertial persistence of at least some temporal concreta. For this reason, my thesis here affirms that some objects actually inertially persist.

3.4 Dependence and Destruction

Still further taxonomic questions concern the kind of ontological dependence denied of inertially persistent objects. As I use it—and bearing in mind that this is a rough approximation—x ontologically depends on y just in case there is some extramental explanatory relation (e.g., causing, grounding, realizing) between y (the explanans) and x (the explanandum), such that x would not exist if x were not standing in this explanatory dependence relation to y (say, because y ceased to exist, or because y no longer engaged in the relevant activity, or whatever). Existential inertia debates are fundamentally concerned with ontological dependence, i.e., x’s ontologically depending on the concurrent or simultaneous existence or activity of y (at any moment at which x exists), such that x would instantly cease to exist the moment the explanatory dependence relation to y ceased to obtain. The relation, then, is a kind of sustenance or conservation. Note, finally, that I take ontological dependence to be a relation among concreta (and/or the activities, properties, parts, etc. of concreta).

The inertial theses of Schmid and Audi only deny the concurrent efficient causal dependence of inertially persistent objects on other things. But this seems too weak. Suppose all temporal concrete objects have a sustaining explanation not in terms of an efficient cause but instead in terms of a ground.[6] In that case, every temporal concrete object is concurrently dependent on some ground. And surely no temporal objects in this scenario enjoy existential inertia. Consequently, inertial theses should deny more than (concurrent) efficient causal dependence. One dependence question, then, is: what kind of ontological dependence is denied of inertially persistent objects?

The notion of concurrence is not obviously well-defined for relativistic space-times because, according to orthodoxy, relativistic space-times do not include objective simultaneity. But let’s put this worry aside for the moment and focus on Newtonian and Galilean space-times. Plausibly, inertial theses should affirm that, for each O within the scope of the inertial thesis, O’s existence does not concurrently ontologically depend on any O* that is not part of O. For, presumably, any composite O ontologically depends (at least in some sense) on O’s parts. In order to maintain any plausibility, then, inertial theses should allow constitutive dependence. What Newtonian and Galilean inertial theses should deny is that inertially persistent objects are either concurrently caused by, grounded in, or realized by things numerically distinct from and wholly outside them—that is, by non-parts.

Furthermore, inertial theses should plausibly remain neutral on concrete objects’ dependence on abstract objects. Suppose, for instance, that temporal concrete objects like people, particles, and platypuses have a range of essential properties, and suppose further that they have such properties in virtue of exemplifying or instantiating the relevant abstract universals. Finally, suppose that such abstract universals are not parts of the concrete objects in question. In the case at hand, one might think that temporal concrete objects ontologically depend on numerically distinct, non-constituent objects. Thus, in order to avoid entanglement with debates concerning abstract objects, I advise that inertial theses only deny that temporal concrete objects ontologically depend on numerically distinct, non-part concrete objects. And notice that once we put the requirement this way, we’ve no longer worded the requirement in terms of concurrence, and so the requirement is perfectly suited for inertial theses applicable to relativistic space-times.

What about destruction? What does it take to destroy inertially persistent things, to make them cease to exist? As with sustenance, Schmid’s and Audi’s inertial theses are cast in terms of causal destruction. But, plausibly, we should broaden this to include uncaused but grounded destruction.

Moreover, if absences can serve as causes, then Schmid’s EIT is perfectly compatible with every temporal concrete object being continuously sustained in existence (even though—ex hypothesi—they do not require sustenance in order to persist). For if absences can serve as causes, then the absence or withdrawal of (say) divine timeless conservation of O (for any O within EIT’s scope) might plausibly be a cause of O’s ceasing to exist, such that any actual withdrawal of such conserving activity causes the cessation of O. But in such a scenario, both (i) and (ii) of Schmid’s EIT are satisfied: (i) O persists without requiring sustenance (though it is in fact continuously sustained), and (ii) O ceases only if caused to cease (by a withdrawal of God’s sustaining activity, say). Thus, assuming that absences can serve as causes, Schmid’s EIT renders the following conjunction possible: (i) nothing inertially persists (i.e., every temporal concrete object is continuously conserved or sustained in being), and yet (ii) EIT is true. This seems implausible. By my lights, the truth of an inertial thesis should secure the actual inertial persistence of at least some temporal concrete objects.

This problem will likely afflict any inertial thesis that references causal destruction. To avoid controversial debates about causal relata, then, inertial theses should generally be cast in terms of positive destruction—some positive activity performed on O (either from within or from without) that brings about O’s cessation (i.e., that makes O cease to exist). This avoids altogether the question of whether absences can serve as causes and also allows for uncaused but grounded destruction. Now, positive destruction can be formulated in four ways. First, let’s define t-temporal part such that a t-temporal part is an object’s temporal part existing at time t. Then two such ways to define positive destruction are:

Positive destruction of enduring objects in Galilean/Newtonian space-time: to positively destroy O at time t is to actively bring about O’s cessation at t, such that O endures through [t’, t) but not [t’, t], where t’ < t.[7]

Positive destruction of perduring objects in Galilean/Newtonian space-time: to positively destroy a t-temporal part of O is to actively prevent this t-temporal part from existing (such that the t-temporal part would have existed had the active prevention not occurred). So: to destroy a t-temporal part of O is to act on one or more temporal parts of O in some interval [t’, t), where t’ < t, so as to actively prevent the existence of O’s t-temporal part.[8]

I have left two more ways of characterizing positive destruction: the positive destruction of enduring and perduring objects in relativistic space-times. I postpone characterizing these until Section 4 where I’ll introduce some additional formalism. For now, I turn to the penultimate taxonomic question.

3.5 Metaphysical Accounts

The central metaphysical question any inertial thesis must be supplemented with runs: in virtue of what does inertial persistence obtain (if it obtains at all)? Answers to this question are metaphysical accounts of existential inertia—coherent stories on which persistence is explained in inertialist-friendly terms (i.e., terms which include commitment to some temporal concrete objects persisting without external conservation—or, at least, terms which do not commit to the thesis that every temporal concrete object is continuously conserved from without.).

I’ll cover a variety of metaphysical accounts of EIT in Section 5. The purpose of the present section is simply to highlight the need for such accounts.


[1] Here, I mean ‘events’ as understood in the metaphysics literature. In relativistic physics, an ‘event’ is standardly defined as a space-time point. I shall also assume for simplicity that concrete objects are not themselves events (or mere collections thereof).

[2] Events as standardly understood within relativistic physics—that is, space-time points—are not good candidates for entities that persist over time either, because space-time points are not typically thought to either perdure or endure in Galilean or Minkowski space-time (Dasgupta 2015; Maudlin 2012, pp. 54-66; Gilmore 2008, p. 1226).

[3] While Feser and Adler solely quantify over contingent things, Audi, Schmid, and Beaudoin leave open (in principle) quantification over necessary things. Beaudoin holds that an object inertially persists when it continues in being “without being sustained in existence by any external agent” (2007, p. 86), while Schmid’s EIT quantifies only over temporal concrete objects as such. And Audi’s thesis, of course, is unrestricted in scope.

[4] This point against cosmic time as a candidate for absolute time has previously been made in, e.g., Saudek (2020, p. 56) and Lockwood (2007, pp. 117-118). As Gerald Whitrow describes, “cosmic time is essentially a statistical concept, like the temperature of a gas” (1961, p. 246).

[5] Since cosmic time depends on CMC slicing, note that evidence against the view that our universe can be globally sliced into CMC surfaces is also evidence against the view that cosmic time by can be adequately defined.

[6]  Pearce (2017), for instance, argues that God is the non-causal ground of the realm of non-God concrete objects.

[7] There may be a sorites paradox associated with positive destruction of material objects. For example, if a material object goes out of existence by losing constituent atoms, there might be no specific number of atoms that sharply divides existing from not existing and so no particular time at which the object goes out of existence. But, clearly, there is a time when a given material object has not gone out of existence and a time when the object has already gone out of existence. Therefore, instead of defining positive destruction as taking place at a specific time, positive destruction can be defined as having taken place somewhere in a range of times. In any case, we can ignore this complication for the purposes of this post.

[8] Note that this is a stipulative definition of ‘destruction’. We recognize that it may sound odd to call it ‘destruction’. (Then again, lots of our ordinary concepts sound odd when adapted or applied to perdurantism.) What matters for present purposes is that we have a clear and precise definition of the relevant concept. If the reader still demurs at our usage, simply replace the word with another, more fitting one.

4 A rigorous articulation of EIT*

  • Once again, thanks to DJ Linford for co-authoring this section. I will use ‘we’ and ‘our’ in this section to reflect this.

We consider two theories of persistence—perdurance and endurance—and three kinds of space-time—Newtonian, Galilean, and relativistic. We will use our answers to the taxonomic questions previously considered to build inertial theses for the various possible combinations of theories of persistence with kinds of space-time while avoiding the various problems afflicting the inertial theses of Audi, Schmid, and company. We will then define the Existential Inertia Thesis (EIT) as the disjunction of the various inertial theses.[1] We start by articulating our inertial thesis for endurance and Newtonian or Galilean space-times:

Newtonian/Galilean Endurantist Existential Inertia Thesis (NG-EEIT): For each member O of a (proper or improper) sub-set of temporal concrete objects and for each time t such that O exists at some time t* earlier than t, (i) at t, O does not ontologically depend on the existence or activity of some concrete object O*, where O* is not a (proper or improper) part of O, and (ii) if O is not positively destroyed within the temporal interval [t*, t], then O exists at t.

What about the relativistic analogue of NG-EEIT? First, let’s define ‘<<’ such that for achronal surfaces x and y, x << y if and only if every point in the achronal surface x is in the past light cone of some point in y, that is, x is in the absolute past of y. Second, note that on relativistic accounts of endurance, the analogue of times—importantly, the times at which O wholly exists—are a specific set of achronal surfaces; how that set of achronal surfaces is picked out will depend on the specific account of endurance.

For any point-like proper part p of O, there is a time-like curve traversed by p. Because O is spatio-temporally extended and has at least two space-like separated proper parts, there exists a collection of time-like curves traversing O’s path such that each curve in the collection is the trajectory of one point-like proper part of O. Call that collection O’s congruence and formally denote O’s congruence by Cong(O). Cong(O) is not necessarily restricted to O’s path because Cong(O) can, in principle, be extended beyond O’s path by continuing the trajectories of the point-like proper parts of O under restrictions imposed by the relevant physical laws. For example, suppose the existential inertia thesis is false and O stops existing when O is not sustained in existence by some entity E. In that case, the path of O could be cut prematurely short if E no longer sustains O, but we could still use physical laws to project where O’s path would have been had E continued to sustain O. That projected (but non-actual) path is a portion of Cong(O).

On relativistic endurantist accounts, O mereologically endures just in case O is wholly located at a specific collection of achronal surfaces intersecting O’s path, where the specific collection is determined by which relativistic endurantist account one chooses. (For example, Balashov (2000b; 2014) has considered an account on which O is wholly located at all of the space-like surfaces intersecting O’s path.) We leave open which endurantist account, if any, is correct and so leave open which collection of achronal surfaces is relevant. In any case, define SCong(O) to denote this collection of achronal surfaces. Because SCong(O) occupies the same space-time region as Cong(O), we can continue SCong(O) beyond O’s path the same way that we continued Cong(O) beyond O’s path. Consider, then, two achronal surfaces a ∈ SCong(O) and a* ∈ SCong(O) that bound a portion of Cong(O) and are such that (i) a* << a and (ii) O wholly exists on a*. In that case, if nothing positively destroys O along Cong(O) between a* and a, then, according to the existential inertialist, O also wholly exists on a. For a formal definition, we have:

Relativistic Endurantist Existential Inertia Thesis (R-EEIT): For each member O of a (proper or improper) sub-set of spatio-temporal concrete objects and for each surface a ∈ SCong(O) such that O exists on some achronal surface a* ∈ SCong(O) and a* << a, (i) on a, O does not ontologically depend on the existence or activity of some concrete object O*, where O* is not a (proper or improper) part of O, and (ii) if O is not positively destroyed on any surface s ∈ SCong(O) such that a* << s  << a, then O exists on a.

Recall that we’ve left the characterization of the positive destruction of enduring objects in relativistic space-times to this section. To positively destroy O on achronal surface s is to actively bring about O’s cessation on s, such that O endures through the portion of SCong(O) beginning at some achronal surface s’ ∈ SCong(O), such that s’ << s, but does not exist on s.

Now, a problem for positive destruction in Newtonian/Galilean space-times is that the end of the life of a material object may be vague. In some sense, this problem is exacerbated for the positive destruction of objects in relativistic space-times. To see why, first suppose that all of the atoms comprising O cease to exist at the same time according to reference frame F. In that case, there will exist another reference frame F’ in which the atoms comprising O do not go out of existence at the same time. Second, notice that, since the atoms comprising O will, in general, be in motion relative to each other, even if all of O’s atoms go out of existence at one time in the rest frame of one of the atoms, O’s atoms will not generally go out of existence at the same time in the rest frames of other atoms comprising O. And, third, there is some difficulty associated with defining the rest frame of a composite object in relativistic space-times.

These difficulties can be addressed with a solution parallel to the one that we sketched in the Newtonian/Galilean case. That is, positive destruction can be defined as having taken place within some space-time region sandwiched between two achronal surfaces. In any case, as with the Newtonian/Galilean case, we can ignore this complication for the purposes of this post.

Debates concerning existential inertia tend to be cast in terms of endurantism. And although endurantism and A-theory are often tied together in such debates, there has long been reason to think that objects can endure in Newtonian or Galilean space-time blocks by being multiply located at all the times at which they exist (van Inwagen 1990). Moreover, both NG-EEIT and R-EEIT make no use of tensed facts and so are consistent with both A-theory and B-theory. But we need not artificially restrict existential inertia to endurantism; thus, we proceed to sketch articulations of existential inertia available to perdurantists. Again, we first formulate an inertial thesis applicable to Newtonian or Galilean space-times. Consider the following:

Newtonian/Galilean Perdurantist Existential Inertia Thesis (NG-PEIT): For each member O of a (proper or improper) sub-set of temporal concrete objects and for each time t such that O has some t*-temporal part earlier than t, (i) O’s t-temporal part does not ontologically depend on the existence or activity of some concrete object (or temporal part of some object) O* numerically distinct from any temporal part of O, where O* is not a (proper or improper) part of O’s t-temporal part, and (ii) if none of O’s temporal parts within the interval [t*, t] are positively destroyed, then O’s t-temporal part exists.

In essence, NG-PEIT states that an object perdures at non-first times of O’s temporally extended life without ontological dependence on concrete non-parts of O so long as nothing positively destroys O’s temporal parts (which, recall, means that nothing actively prevents O from having such parts). Likewise, we can formulate an inertial thesis for perduring objects in relativistic space-times:

Relativistic Perdurantist Existential Inertia Thesis (R-PEIT): For each member O of a (proper or improper) sub-set of temporal concrete objects and for each achronal surface, a, such that O has some achronal part p* on a* where a* << a, (i) p does not ontologically depend on the existence or activity of some concrete object (or achronal part of some object) O* numerically distinct from any achronal part of O, where O* is not a (proper or improper) part of p, and (ii) if none of O’s achronal parts within the portion of Cong(O) connecting a* to a are positively destroyed, then O has an achronal part p on a.

Recall once more that we left the characterization of the positive destruction of a perduring object in a relativistic space-time to this section. To positively destroy an achronal part p of O on achronal surface s is to actively prevent p from existing on s, such that O perdures on the portion of SCong(O) to the absolute past of s, and in such a way that p would have existed on s had the active prevention not occurred.

Now we can re-define EIT as the disjunction of NG-EEIT, R-EEIT, NG-PEIT and R-PEIT.

Existential Inertia Thesis (EIT): NG-EEIT or R-EEIT or NG-PEIT or R-PEIT

Having defined our EIT, we can now turn to the final taxonomic question.[2]

4.1 Temporal Ontology

The intersection between EIT and temporal ontology raises one final taxonomic question that any inertial thesis must answer: what is the temporal ontology within which the inertial thesis is articulated and understood? Three issues are included under temporal ontology: (i) the ontological status of moments of time (presentism, eternalism, growing block, etc.), (ii) the objective un/reality of temporal becoming (A-, B-, C-theories, or something else altogether), and (iii) the manner in which objects persist (endurantism, perdurantism, etc.). We’ve already covered (iii), so let’s examine (i) and (ii).

Let’s turn, first, to the ontological status of moments of time. Non-relativistic presentist, eternalist, growing block, and moving spotlight accounts are easy to come by. Although Newtonian/Galilean space-times are typically characterized under the assumption that eternalism is true,[3] our discussion of Newtonian/Galilean space-times made no such assumption. (Moreover, nothing in our Newtonian/Galilean inertial theses assumes eternalism.) Therefore, our Newtonian/Galilean inertial theses are applicable in the context of Newtonian/Galilean versions of presentism, eternalism, growing block theory, and moving spotlight theory. Whether adequate relativistic versions of presentism, growing block theory, or moving spotlight theory can be constructed is more controversial, but to the extent that such accounts can be constructed, nothing in our relativistic inertial theses prevents their adoption. And so EIT is consistent with presentism, eternalism, growing block, and moving spotlight.

Now consider the objective un/reality of temporal becoming. As we’ve already noted, our formulation of EIT does not make use of tensed facts and so is consistent with both A- and B-theory. Our formulation of EIT is probably not compatible with C-theory, but we have difficulty understanding what ‘temporal persistence’ could mean if C-theory were true.

4.2 Taxonomic questions: Summary

I (really, we—DJ and I) have raised a dozen (or so) taxonomic questions for inertial theses. In order to appreciate the larger picture, I’ve listed each of the taxonomic questions below:

  1. Scope: Does the thesis quantify over temporal or non-temporal things (or both)?
  2. Scope: Does the thesis quantify over concreta or abstracta (or both)?

Assuming we only quantify over temporal concreta:

3. Scope: What kind of temporal concreta—objects, events, concrete states of affairs, etc.—inertially persist?

Assuming we only quantify over temporal concrete objects:

4. Scope: Do all temporal concrete objects inertially persist or only some proper subset thereof?
5. Persistence: Will the existential inertia thesis understand persistence in endurantist terms, perdurantist terms, or something else altogether?
6. Persistence: How does persistence (and, thus, existential inertia) relate to the findings of contemporary physics?
7. Modal Register: Is the inertial thesis necessarily true if true at all?
8. Modal Register: Do the temporal concrete objects within the domain of quantification actually persist without continual sustenance, or do they merely persist without requiring continual sustenance?
9. Dependence: What kind of ontological dependence is denied of inertially persistent objects?
10. Destruction: What does it take to destroy inertially persistent things, to make them cease to exist?
11. Metaphysical Account: In virtue of what does inertial persistence obtain (if it obtains at all)?
12. Temporal Ontology: What is the temporal ontology—encompassing (i) the ontological status of moments of time, (ii) the objective un/reality of temporal becoming, and (iii) the nature of persistence—within which the inertial thesis is articulated and understood?


[1] For present purposes, we will assume that material objects are composed of point-like proper parts and ignore the complications posed by quantum mechanics.

[2] Another important taxonomic question—one that we shan’t explore here—is whether we understand the structure of time continuously or discretely. For our purposes here, we leave this taxonomic question open.

[3] Newtonian/Galilean space-times are typically understood in what Gilmore, Costa, and Calosi (2016, po. 102-3) call unitist terms—that is, as being composed of space-time points (i.e., points that are neither fundamentally spatial nor temporal). Those points can be divided into equivalence classes that correspond to space at distinct times. (The points composing a relativistic space-time cannot be so divided.) In contrast, separatist accounts of space-time include two collections of points, one of which is fundamentally spatial and the other of which is fundamentally temporal. We have purposely phrased our characterization of Newtonian/Galilean space-times as neutral between unitism and separatism.

5 The metaphysics of EIT

My aim in this section is to categorize and develop metaphysical accounts of EIT—that is, explanations of persistence on which EIT holds (or can hold). This includes tendency-disposition accounts (§5.1), transtemporal accounts (§5.2), law-based accounts (§5.3), necessity accounts (§5.4), and no-change accounts (§5.5).

I do not take a stance on which of the accounts is best (or which, if any, is true). Instead, I aim to categorize, develop, and occasionally sketch motivations for such accounts. One reason such categorization and development is significant is that many objections to existential inertia simply fail to take into account inertialist-friendly explanations of persistence.[1] It is also significant because the sketched motivations for (some of) the accounts can potentially serve as arguments in favor of EIT.

5.1 Tendency-disposition Accounts

Tendency-disposition accounts explain persistent existence by appeal to some tendency or disposition of things to continue in existence. There are at least three closely-related tendency-disposition accounts in the literature: Beaudoin’s (2007, pp. 88-89), Benocci’s (2018, pp. 59-63), and Oderberg’s (2014, pp. 349-353).[2] I shall take these in order, beginning with Beaudoin.

It is worth quoting at length Beaudoin’s exposition of his account:

It is not part of [EIT] to suggest that the continuance of things is a brute fact. It is explained by reference to the facts (i) that the only power capable of annihilating the world’s fundamental material has so far gone unexercised, and (ii) that this material has no inherent tendency to just spontaneously disappear… Here again the analogy with mechanical inertia is illustrative: the continued uniform motion of a body through space is not to the physicist a mere surd. It is the outcome of the absence of any unbalanced force applied to the object, combined with its natural tendency to keep moving unless such a force is encountered. Of course, one may ask why the motion or existence of any object is characterized by inertia, and the inertialist in either context may or may not be able to provide an answer… But even if the existential inertialist cannot identify any deeper metaphysical basis for this form of inertia, this in no way invalidates [EIT] as an explanation of the world’s continuance; it is not a condition on legitimate explanation that a deeper explanation for every statement in the explanans always be ready to hand, or even that it exist at all. The inertialist may well run into a brute fact somewhere in his accounting for the world’s continuance, but it is far from clear that the proponent of DDC [the Doctrine of Divine Conservation] will fare better in this regard. (2007, pp. 88-89)

For Beaudoin, then, the explanation of why inertially persistent object O continues to exist is the following conjunction: (i) the only power capable of annihilating O has thus far been unexercised, and (ii) O lacks a tendency to spontaneously disappear. For if O lacks such a tendency, then O will not spontaneously disappear unless some sufficiently destructive or annihilating factor—whether intrinsic or extrinsic to O—comes along to destroy O. This is part and parcel of what tendencies involve: O has a tendency to manifest some outcome or undergo some process in conditions C if and only if O, when placed in C, manifests said outcome or undergoes said process.[3] Thus, if O lacks a tendency to spontaneously annihilate (disappear, cease to exist) in conditions C—say, when not subjected to sufficiently destructive or annihilating factors—then O, when placed in C, will not annihilate. And provided that O is in C—provided that there is an absence of sufficiently destructive or annihilating factors, as specified by condition (i) of the abovementioned conjunction—it follows that O will not annihilate (i.e., disappear or cease to exist) but will instead persist.

Thus, according to Beaudoin’s account, in order to explain O’s persistence, we need only cite the absence of a ‘spontaneous annihilation tendency’ in conjunction with the absence of sufficiently destructive factors operative throughout O’s life up until the present.

There is, of course, the further question as to why no such (sufficiently) destructive factors have been operative—why, in other words, the only power capable of annihilating O has thus far been unexercised—but this is a separate question from the metaphysical account of O’s persistence. And, plausibly, it won’t be all that difficult to provide plausible stories concerning the absence of such (sufficiently) destructive factors. Likewise, there is the further question as to why objects lack the spontaneous annihilation tendency. But it’s not clear why explaining this would be any more difficult than explaining why objects have or lack other tendencies. (And, to reiterate Beaudoin, it is not a condition on adequate metaphysical accounts of EIT that we have such further explanations ready to hand.)

A second tendency-disposition account is found in Benocci (2018), who emphasizes that “endurance theorists do not need any special metaphysical principle or postulate to account for existential inertia” (2018, p. 59). Benocci appeals to dispositions, which “consist in displaying a certain kind of manifestation under a certain kind of condition or stimulus” (2018, p. 60). He begins with an intuitive principle according to which changes an object undergoes reveal the presence of a corresponding disposition of which the relevant change is a manifestation. More precisely, “if an object a undergoes a change m in the circumstance c, then it has the disposition to undergo a change of kind M in circumstances of kind C, with m being a change of the kind M and c being a circumstance of the kind C” (2018, p. 60). He also defines a trivial disposition as a disposition that manifests in all possible circumstances.

With the aforementioned principle in hand, Benocci adds another principle to the mix:

Complementarity Principle: For any object x and any non-trivial disposition D to display a manifestation of the kind M in (and only in) the conditions of the kind C, if x has D then x also has the disposition D’ not to display a manifestation of the kind M in (and only in) the conditions other than those of the kind C. (2018, p. 62)

In simpler terms, any object with a disposition D similarly has D’s complementary disposition. Benocci argues that the principle is both intuitively plausible and free from ontological commitments. For although it may seem to require universals or tropes (since dispositions seem to be properties of some kind), the Complementarity Principle can be recast in terms that don’t quantify over dispositions as such but instead merely on objects’ being disposed in certain ways:

Complementarity Principle*: For any object x, if x is non-trivially disposed to display a manifestation of the kind M in (and only in) the conditions of the kind C, then x is also disposed not to display a manifestation of the kind M in (and only in) the conditions other than those of the kind C. (2018, p. 62)

From these principles, we have a recipe—so Benocci argues—for an account of existential inertia. For according to the Complementary Principle, any object with a disposition to destruct in (and only in) circumstances C likewise has a disposition to remain in existence in (and only in) non-C circumstances. And it is this disposition of an object to remain in existence that, according to Benocci, is “aptly called the existential inertia of that object” (2018, p. 62). All that’s left to add is that the relevant circumstances of kind C are those in which the object is subject to sufficiently destructive or annihilating factors. Benocci concludes:

[T]here is nothing spooky in attributing an existential inertia to ordinary objects. To talk about the existential inertia of an object that is enduring is not different from talking about the ‘malleability inertia’ of a piece of iron that is not undergoing a compressive stress: in both the cases we have just an object that is not manifesting a certain disposition, and which on the other hand is manifesting the complementary disposition to remain in a certain state. With this in mind, the perturbation that makes an object pass away is simply a stimulus of the kind that makes it display its destructive disposition. (2018, p. 63)

This, then, is the essence of Benocci’s strategy. It seems to represent a prima facie defensible, coherent, and ontologically lightweight metaphysical account of EIT.

The third and final tendency-disposition account I’ll consider is Oderberg’s.[4] According to Oderberg, “things tend to continue to exist” (2014, p. 349). This tendency to persist is “so basic to a thing” that it “manifest[s] not only for as long as the thing exists but because it exists” (ibid, p. 350).

One might wonder whether this account is circular. Isn’t Oderberg committed to both (i) objects enjoy a tendency to persist because they exist, and (ii) objects exist because they have a tendency to persist? More poignantly, Oderberg takes the tendency to persist to be a kind of “property—one that holds for as long as the object exists and precisely because it exists.” But aren’t properties ‘ontologically parasitic’ on their bearers, in the sense that properties are posterior to and dependent upon the (ontologically) prior existence of their bearers? And if properties are posterior to the existence of their bearers, how could such properties explain their existence? We could also frame this circularity worry diachronically. If, for any property F, F posterior to its bearer O, then O’s continued possession of F merely presupposes O’s continued existence. But then S’s continued possession of F cannot explain O’s continued existence.

This is one of two general problems that any account of EIT must avoid. In particular, any account of EIT must avoid both (i) vicious circularity (i.e., presupposing the (prior) persistence of the concrete object(s) in question) and (ii) relocating the quest for an explanation of persistence (i.e., explaining the persistence of objects of type T by implicitly or explicitly assuming the persistence of objects of type T*). We will see throughout many of the following subsections how the various accounts circumvent such worries. For now, let’s consider how Oderberg replies to the circularity worry.

According to Oderberg, the charge of circularity is misguided. An object has the tendency to persist “because it exists as a certain kind of thing. … [A]ll concrete objects… have it because of the kind to which they all belong, namely, concrete object. So concrete objects have the tendency to continue to exist because they exist as concrete objects” (ibid). Oderberg explains that such objects “do not exist as concrete objects because they have the tendency to continue to exist”, thus avoiding vicious circularity (ibid). Answering the question of why they exist as concrete objects would appeal to some “fundamental metaphysical analysis” rather than “some property or other of the object” (ibid).

On Oderberg’s view, while “it is logically impossible for a thing to have the tendency to continue to exist without existing in the first place,” this doesn’t entail that “its mere existence explains the tendency. So we do not have any explanatory or other circle here” (ibid). In short, the explanation of why concrete objects persist in existence is that they have a tendency to do so; and they have a tendency to do so not because they exist or persist, but rather because of the kind of things they are—they have such a tendency by their very natures. (There is of course the further question of why they essentially have such a tendency. But now we are asking a separate explanatory question from what explains their persistence. Moreover, it’s not clear why this would be any more mysterious than why objects have other tendencies or dispositions by nature.)

Oderberg’s response, though, might seem unsatisfying. For while neither the mere existence nor mere continued existence of O explains O’s tendency to continue to exist, Oderberg seems to grant that O’s existence is prior to O’s tendency. (For instance, Oderberg says O cannot have the tendency without existing in the first place.) And this is surely all we need to get the circularity worry up and running. For if O’s tendency to persist presupposes O’s existence, then surely the continued possession and manifestation of said tendency likewise presupposes O’s continued existence. And in that case, the tendency cannot explain or account for O’s continued existence. Consider: even though my having a phone (plausibly) doesn’t explain why I received a call from my friend today, it is nevertheless prior to and presupposed by the latter. And in that case, the latter cannot explain why the former obtains.

But perhaps Oderberg can avoid this circularity, too. For Oderberg can grant that at time t, O’s tendency to persist cannot explain O’s existence at time t, precisely because the former presupposes the latter. But that is not the proposal. The proposal for explaining O’s existence at some non-first time t at which O exists is, rather, something like: (i) O existed at some time t* earlier than t; (ii) O possessed, at t*, the tendency to persist in existence unless subjected to sufficiently destructive factors; (iii) O has this tendency in virtue of the kind of thing O is (i.e., in virtue of O’s nature); and (iv) O was not subjected to sufficiently destructive factors between t* and t. The explanans here is neither explained by nor presupposes the prior obtaining of the explanandum (O’s existence at t). Circularity, then, seems to be avoided after all.[5]

Before turning to transtemporal accounts, I want to highlight Oderberg’s motivation for belief in the tendency to persist. He points out, first, that “[i]t is not as though objects cease to exist under such a variety of circumstances that we cannot, even in principle, find anything common to them” (ibid, p. 351). Instead, there is little mystery about the conditions under which objects cease to exist: “They cease to exist when and only when forces act upon them” (ibid). This is not some brute fact, “any more than if it were the case that all objects ceased to exist when and only when in the vicinity of objects twice their size” (ibid). We need some account of why there is such commonality among ceasing to exist. According to Oderberg, this is explained in terms of a tendency to cease to exist, a tendency whose manifestation conditions involve “destruction by the application of forces” (ibid). This tendency to cease, moreover, “depends ontologically on the tendency to continue to exist in the absence of those forces. Nothing can cease to exist in certain conditions unless it already continues to exist prior to those conditions” (ibid). Oderberg is here appealing to some version of the complementary principle adduced by Benocci.

We could put Oderberg’s argument for the tendency to persist as follows. We observe temporal concrete objects ceasing to exist when and only when subjected to sufficiently destructive factors/forces. But this universal phenomenon cries out for some explanation, just as objects’ ceasing to exist when and only when next to objects twice their size does. The best or only explanation for this phenomenon is that such objects possess some tendency to cease to exist when and only when subject to such sufficiently destructive factors.[6] But tendencies imply their complements; in other words, the possession of a tendency to cease to exist in certain conditions—“a specific disposition activated by a relatively limited kind and range of stimuli”—implies the possession of “a tendency to continue to exist absent those stimuli” (ibid). Hence, temporal concrete objects have a tendency to persist in existence.[7]

The inertialist, moreover, can extend Oderberg’s argument: given that there is such a tendency, and given that, in conjunction with other conditions adumbrated earlier, this tendency explains the persistence of temporal concrete objects, a continuously operative sustaining cause or ground of the moment-by-moment existence of temporal concreta is explanatorily otiose. ‘Shave it off’, sayeth Occam’s razor and the existential inertialist.

Onward we march, then, to a second family of inertialist-friendly explanations of persistence.

5.2 Transtemporal Accounts

Transtemporal accounts appeal to some kind of transtemporal relation(s) to explain persistence. There are at least two such accounts in the literature, found in Schmid (2021) and Mackie (1974). Let us consider each in turn, beginning with the former.

Here is how Schmid articulates the account:

For concrete object O and times t-1 and t (where t-1 is immediately temporally prior to t), the existence of O-at-t is explained by the conjunction of (i) the state and existence of O-at-t-1 and (ii) the absence of any sufficiently causally destructive factors acting on O-at-t-1 and through t. (2021, p. 205)

Schmid’s account is explicitly written in endurantist terms and is not compatible with the orthodox (i.e., Minkowskian) interpretation of relativity. Nonetheless, Schmid’s account can be easily rendered consistent with the relativistic and perdurantist accounts of existential inertia we discussed in section 2. For example, for globally hyperbolic relativistic space-times, times  t-1 and t can be replaced by appropriately related Cauchy surfaces; meanwhile, instead of considering O at distinct times or on distinct Cauchy surfaces, we can consider proper parts of O located at distinct times or on distinct Cauchy surfaces.

Schmid leaves open the precise nature of the relevant transtemporal explanatory relation, allowing the relation to be either causal or non-causal. According to Schmid’s account, then, O’s persistence is explained by (i) the absence of sufficiently causally destructive factors operative on O, plus (ii) transtemporal explanatory relations (causal or otherwise) obtaining between the temporally successive states of O’s life (so to speak).

Schmid points out that explanations of present things’ existence in terms of past things seem not only legitimate but often indispensable. Present allelic frequencies in biological populations are explained (at least in part) in terms of past selection pressures and past reproductive behavior; discursive reasoning processes plausibly require not only a justificatory or reasons-based link between past consideration of the premises (on the one hand) and acceptance of the conclusion (on the other) but also a causal or explanatory link between one’s previous consideration of the premises and one’s present acceptance of the conclusion; and so on. In principle, argues Schmid, there seems to be nothing debarring past things from explaining (causally or otherwise) the existence of present things.[8]

One might think that Schmid’s account presupposes that for each time there is an immediately temporally prior time. But if time is continuous as opposed to discrete (such that for any two distinct times, there is a time between them), then it is simply false that there exists an immediately temporally prior time. By way of response, note first that even if Schmid’s account works only under discrete time, it would still be significant if there were a workable metaphysical account of existential inertia assuming discrete time.[9] Second, according to Schmid, the account is not (after all) essentially tied to discrete time. If time is continuous, we can let the temporal state immediately prior to t be some suitably small (perhaps infinitesimally small) non-zero interval of time with t as its later-than bound.

The continuity or discreteness of time is something deserving much more attention in debates concerning existential inertia. Pursuing this in the requisite depth, however, would take me too far afield given my present purpose of categorizing, developing, and sketching motivations for different metaphysical accounts.[10] Before turning to Mackie’s transtemporal account, though, I’ll Schmid’s (i.e., my) transtemporal account avoids the vicious circularity worry.

Importantly, Schmid’s transtemporal account does not merely presuppose that O persists from the previous time to the succeeding time; instead, the account provides an explanatory means by which such persistence obtains.  O remains in being at the succeeding time (or on, e.g., the succeeding Cauchy surface) precisely in virtue of the state and existence of O at the prior time (or on, e.g., the prior Cauchy surface) in conjunction with the absence of sufficiently destructive factors. The explanans secures or accounts for (instead of presupposing the ontologically prior reality of) the explanandum (viz. the moment-by-moment existence of concrete objects, for non-first moments of their lives).

I simply see no circularity in the following explanatory schema. For the sake of simplicity, suppose that the relevant transtemporal explanatory relation is causation and suppose that we are considering a Newtonian or Galilean space-time with discrete time. The explanandum is O’s existence at t. Schmid’s explanans, under the aforementioned assumptions, is: (i) There is an absence of sufficiently causally destructive factors operative on O from t-1 to t (where t-1 is the time immediately prior to t), and (ii) the state and/or existence of a temporal concrete object (or, at least, one within EIT’s quantificational domain) at a given time at which it exists causally produces its existence at the next moment provided that no sufficiently causally destructive factors are operative.

I find the explanation proffered here both non-circular and illuminating. There is nothing in the explanans that presupposes the prior reality or obtaining of the explanandum. And the explanation is illuminating in that it cites facts that remove mystery as to why the explanandum obtains—the explanandum was simply derived from the explanatory facts cited.

To be sure, there might be the further question of why some of those explanatory facts themselves obtain. For instance, there might be the question as to why reality is so constituted that the successive stages in an object’s life are related by causal relations.[11] But this is a separate question from why O exists at t. And, plausibly, it won’t be all that difficult to provide plausible stories for the former question. (Indeed, it’s not clear why explaining it would be any more difficult than explaining why reality is so constituted so as to have any causal relations at all (ever), or to have causal relations other than those relating the successive stages of an object’s life, or what have you.) Finally, to quote Beaudoin, “it is not a condition on legitimate explanation that a deeper explanation for every statement in the explanans always be ready to hand, or even that it exist at all” (2007, p. 89).

Now, one might worry that the above transtemporal account does not explain persistence simpliciter but instead only explains O’s existence at t or on a specific Cauchy surface. This worry, however, is misguided. An explanation of O’s persistence simpliciter arises with the conjunction of all the applications of the explanatory schema outlined in Schmid’s account to every non-first time or every non-first Cauchy surface at which O exists.

So much for Schmid’s account—let’s turn to Mackie’s. Unlike Schmid, Mackie specifies that the transtemporal explanation at play is causation. In particular, Mackie suggests treating a physical object as “a self-maintaining process or cluster of such processes” and holds that “[t]he earlier phase of [such] a self-maintaining process surely brings about, or helps to bring about, the later phase” (1974, p. 156). According to Mackie’s transtemporal account, then, earlier phases of temporal concrete object O cause (bring about, produce) later phases of O.

Mackie’s account raises questions concerning the relata of the transtemporal explanatory relation. The relata for Mackie are the ‘phases’ of an object—presumably a stage or state of the object at (speaking loosely) a single moment of its life. But transtemporal accounts need not share this part of Mackie’s account. They can, instead, opt for other views of the relata. Here are some options, where, for the sake of simplicity, we assume a Newtonian or Galilean space-time and that time is discrete:

  1. Where t-1 is the time immediately prior to t and where t is a non-first time at which O exists: the event of O’s existing at t-1 causes—in the absence of (sufficiently) destructive factors—the event of O’s existing at t; or
  1. the state of affairs involving O’s existing at t-1 causes—in the absence of (sufficiently) destructive factors—the state of affairs involving O’s existing at t; or
  1. the conjunctive proposition that O exists at t-1 and that no sufficiently destructive factors are operative explains the proposition that O exists at t; or
  1. the successive states (or ‘phases’) in the life of O itself could be the relata; and so on.[12]

One question for causal transtemporal accounts is: when is the causing taking place? In response, we note that this question is ambiguous between (i) ‘when is the cause causing?’, and (ii) ‘when is the effect effecting (i.e. being effected or brought about)?’. The answer to the former is the immediately prior moment, while the answer to the latter is the immediately posterior moment. And there is no tertium quid; we need not (and, we suggest, should not) reify the ‘event’ of the cause’s causing the effect. There’s just the cause and the effect, the former of which is immediately temporally prior to the latter.

  • Note: Thanks to DJ Linford for primarily authoring the following paragraphs ending the section on transtemporal accounts!

Some might worry that (causal) transtemporal accounts violate the simultaneity of causation. Causes do not temporally precede but are simultaneous with their effects, or so the worry goes.[13] Imagine, for instance, that a foot has (from eternity past) been impressing upon some sand. We are supposed to have the intuition that the imprint in the sand is caused by the foot’s pushing down on the sand. But the latter does not temporally precede the former. There is simply a relation of concurrent dependence. The reader should note that the objection from the simultaneity of causation only works if causes must be simultaneous (or concurrent) with their effects; otherwise, if causes can temporally precede their effects, then defenders of transtemporal accounts can say that temporal objects are the transtemporal cause of their own persistence.

Although a foot causing an impression in some sand from eternity past is a common example of a cause that does not precede its effect, we do not find the example convincing. According to accounts on which causes are substances that produce their effects, why think that the foot must have produced the impression in the sand? Consider, for example, a situation in which there already existed an impression in some sand and, at some subsequent time, a foot exactly matching the impression comes to rest in the impression. Although one could argue that some cause needs to be posited to explain the shape of the impression or to explain why the impression exactly matches the foot, the foot would not be posited as the mediate cause of the impression. We can consider moving the time that the foot comes to rest in the impression arbitrarily far into the past. Points at positive or negative infinity are not standardly included on the real line. Therefore, when we move the event of the foot coming to rest in the impression infinitely far into the past, we remove that event from the timeline altogether. Since, by construction, the production of the impression occurred prior to the foot coming to rest in the impression, we would also have removed the production of the impression from the timeline. There would be no event involving the creation of the impression. So, where some have argued that the foot is the eternal cause of the impression, we can instead imagine that the foot and impression have simply co-existed for all eternity without positing that one is the cause of the other.[14] Whatever one might make of this argument, we think that there is a deeper reason to deny simultaneous causation.

The trouble is that simultaneous causation is not obviously consistent with an orthodox understanding of relativity. As we’ve discussed, orthodox interpretations of relativity entail that objective simultaneity does not exist. Friends of simultaneous causation would be right to point out that they use the term ‘simultaneous’ differently from how that term is deployed by physicists (or even as ‘simultaneous’ has been deployed in much of this post). For physicists, to say that two numerically distinct space-time points p1  and p2 are simultaneous is just to say that p1  and p2  occur at numerically the same absolute time, that is, that p1  and p2  are instantaneous. Friends of simultaneous causation should not be interpreted as endorsing the view that causes are instantaneous with their effects. For that reason, simultaneous causation cannot be ruled out merely on the grounds that relativity rules out objective instantaneity. However, friends of simultaneous causation move too quickly if they conclude that simultaneous causation runs into no problems at all from relativity. The trouble is that the failure of objective instantaneity makes objective concurrence much more difficult to understand. And without objective concurrence, standard accounts of how causes can be simultaneous with their effects break down.

Consider one of Mumford and Anjum’s stock examples of causation. Sugar, water, and stirring are jointly sufficient for producing a sugar water solution. Therefore, according to Mumford and Anjum, sugar, water, and stirring are the joint cause of the sugar water solution. According to Mumford and Anjum, the sugar is stirred into the water over a period of time, say, from t1  to  t2. They argue that the simultaneity of the effect and cause consists in the concurrence of all three components throughout the process from t1  to  t2. However, according to relativistic orthodoxy, there is no such process. Each of the microphysical constituents comprising the sugar and water carries its own clock and, in general, no clock can be constructed that objectively synthesizes all of those microphysical clocks. And since no clock can be constructed that synthesizes all of the microphysical clocks, there is no time at which the sugar objectively begins to be stirred into the water and no time at which the sugar objectively finishes being stirred into the water. There is no objective fact about when the process starts or about when the process ends. Moreover, since each of the microphysical constituents carries its own clock, and no sugar molecule is ever co-located with a water molecule, we have a difficult time understanding how the water molecules could be said to objectively concur with the sugar molecules.

One could try to rescue Mumford and Anjum’s account by assuming that space-time is globally hyperbolic and then defining concurrence in terms of co-existence on the appropriate Cauchy surfaces. But we have trouble seeing how this will ultimately work out for another reason. The points on a Cauchy surface are space-like related. For that reason, no signal can propagate between numerically distinct points on a Cauchy surface. And since causes presumably make some difference to their effects, a signal must propagate from cause to effect. (Here, we mean ‘signal’ to be general enough that the cause, itself, might be the signal that reaches the effect.) For the sake of simplicity, suppose that O is a point-like particle and that t is O’s proper time. The only entities that can transmit a signal to O are located in O’s past light cone. Supposing that O-at-t does not cause O’s persistence at t, whatever substance causes O to persist at t must be located in O’s past light cone. Therefore, if the cause of O’s persistence has a location in O’s proper time and we disregard the possibility of retrocausation, the cause of O’s persistence must be located in O’s absolute past. This argument can be generalized to cases where O is not a point-like object by considering O as a collection of point-like objects.

Friends of simultaneous causation might attempt to argue that causes and effects can concur for another reason. Consider that the interactions between the water and sugar molecules can ultimately be understood in terms of the interactions between electric charges and that the interactions between electric charges are ultimately mediated by photons. For example, when momentum is exchanged between two electrons, we might imagine that one electron transmits a photon, carrying the momentum, and the other electron absorbs that photon, thereby receiving the momentum. When an electron absorbs a photon, we might imagine that the electron and photon coincide at a single space-time point. In the Feynman diagram that illustrates the absorption of a photon by an electron, the point where the electron and photon coincide is called a vertex. One might suggest that since the electron and the photon occupy numerically one space-time point at the vertex, the electron and the photon — and so cause and effect — come to be instantaneous by co-occuring at the same space-time point.  This attractive picture is presented in popular science documentaries, where one billiard ball — representing the electron — somehow absorbs another billiard ball — representing the photon. Unfortunately, as attractive as this picture is, the picture glosses over a tremendous amount of complexity.

While many philosophers outside of philosophy of physics may be tempted to interpret Feynman diagrams (or popular science documentaries) literally, Feynman diagrams are merely a useful heuristic for representing terms in an infinitely long sum, where each term in the sum represents an increasingly complex process, involving larger numbers of vertices and particles (or virtual particles). As Oliver Passon describes the view of Feynman diagrams popular among philosophers of physics, Feynman diagrams (emphasis his) “visualize formulae and not physical processes” (2019, pg. 1; see Passon’s paper, and references therein, for additional arguments against the naive literal reading of Feynman diagrams).  And as Friebe, et al, write, the “intuitively attractive picture of ‘particle exchange’” is a “metaphor” that “occurs only in the Feynman graphs for perturbation calculations” that should not be understood as a description of a real process unfolding in space-time (Friebe, et al, 2018, pp. 241-2). More generally, Feynman diagrams “do not have the task of imaging spatiotemporal processes. They cannot play that role for reasons of principle” (ibid, pg. 241).

Instead of interpreting Feynman diagrams literally, realists about quantum field theory should say that, to the degree that the sum should be interpreted realistically at all, reality is represented by the full sum and not by particular terms in the sum. Moreover, while electrons and photons are represented in Feynman diagrams as discrete particles, they should more accurately be understood as quantized vibrations in their respective fields. The “trajectories” in distinct Feynman diagrams have questionable physical significance, both because the “trajectories” can interfere with each other and because topologically equivalent diagrams featuring distinct intermediate particles correspond to numerically the same terms in the sum (Passon 2019, pp. 11-2). Realistically interpreted, the full sum represents (perhaps) some complex, knotted, spatio-temporally extended and difficult (perhaps impossible) to imagine mess of vibrations in the electron and photon fields. Whether a localized event occurs in which two discrete particles momentarily occupy a common vertex is at least unclear. Thus, whether the absorption of a photon by an electron can legitimately be said to be an example of a state of affairs in which a cause comes to be instantaneous with its effect by occupying numerically one space-time point is at least unclear, if not completely unfounded.

Perhaps friends of simultaneous causation could instead object by appealing to empirically indistinguishable alternatives to relativity that reinstate absolute time. If absolute time can be reinstated, then the intuitive notion of causes concurring with their effects can be reinstated. For our purposes in this post, we make two observations. First, alternatives to relativity remain deeply controversial. So long as relativistic orthodoxy remains a live option, the possibility that causes do not concur with their effects remains a live option. And, for that reason, transtemporal accounts of existential inertia remain live. Second, the objection concerns alternatives to relativity that are empirically indistinguishable from relativity, at least within some domain. In any case in relativity’s domain of application where we can detect a relationship between a cause and an effect, theories that are empirically indistinguishable from relativity must have the same observational consequences as relativity. Thus, within relativity’s domain of application, insofar as relativity suggests that causes temporally precede their effects, any observationally indistinguishable alternative to relativity will suggest that causes temporally precede their effects.

But why might someone accept transtemporal accounts? We’ve already canvassed some reasons (e.g., plausibly, appealing to past things to explain present things is often both legitimate and indispensable). Schmid offers (what can naturally be interpreted as) an argument from diachronic identity. In particular, Schmid is concerned with “adequately accounting for the distinction between instantaneous replacement and genuine persistence” (2021, p. 215). Schmid argues that transtemporal causal relations plausibly provide necessary conditions for adequately distinguishing “between a successive series of numerically distinct but qualitatively similar simulacra (on the one hand) and genuine persistence (on the other)” (ibid). According to Schmid, relevant causal continuity among the stages of an object’s life—where “the later existence of the object at least partly causally or explanatorily depends on the earlier state(s) and existence of the object”—is a necessary condition and partial ground of diachronic identity (ibid). And if we already have such transtemporal relations in our ontology, why posit anything further to account for persistence? “Shave it off”, sayeth Ockham’s Razor and the existential inertialist.

Onward we march to law-based accounts.

5.3 Law-based Accounts

  • I thank DJ Linford for writing this section. It is not 100% complete — it is still under construction. But once it is complete, this blog post will be updated.

Law-based accounts represent another family of inertialist-friendly explanations of persistence. Such accounts grant laws of nature a special role in explaining (inertial) persistence. Philosophers have developed several distinct accounts of laws of nature. For each account of laws of nature, we develop an associated account of inertial persistence.

Hume famously argued that there are no relations of necessary connection between distinct existences. That is, in instances where we have the expectation that, given the occurrence of an F, an instance of a G will follow, we have the expectation because our minds have formed an association between instances of Fs and instances of Gs and not because there is any necessity linking Fs to Gs. (Or so goes the standard interpretation of Hume provided in undergraduate classes; addressing whether this is the correct interpretation of Hume is beyond the scope of this post.) David Lewis developed a metaphysical system — that we will call Humean Supervenience — according to which all that ultimately exists are (1) a collection of space-time points and (2) an assignment of qualities to space-time points. Everything else supervenes on the distribution of qualities. Lewis’s program takes inspiration from Hume in that there are no necessary relations between distinct existences, that is, between distinct space-time regions. Causation, laws, and chance are all to be explained in terms of how the distribution of qualities covaries between possible worlds. For example, according to the Best System Account (BSA) of laws, although the distribution of qualities can be systematized or summarized in various ways,  the laws are the axioms of the system that captures the best trade off between simplicity and generality.

The doctrine of Humean Supervenience can be generalized. Instead of understanding the distribution of qualities as an assignment of qualities to space-time points, we can instead understand the distribution of qualities as an assignment of qualities to the fundamental arena, whatever the fundamental arena turns out to be. For example, if, as David Albert, Barry Loewer, and Alyssa Ney have argued, we should understand the wavefunction as a distribution of qualities on a high dimensional configuration space, then Humean Supervenience can be understood as the doctrine that there is, ultimately, nothing but a distribution of qualities on a high dimensional configuration space.

According to friends of Humean Supervenience, what explains the persistence of objects? Ordinary objects supervene on the distribution of qualities on the fundamental arena. Consequently, what explains the persistence of objects is just the fact that a specific pattern of qualities obtains on the fundamental arena. Consider that, for friends of Humean Supervenience, the truth conditions for modal claims — and so the truth conditions for counterfactuals — are provided in terms of how the distribution of qualities covaries from the actual world to nearby worlds. So, why does an object persist from one time to another or from one Cauchy surface to another or whatever?

For friends of Humean Supervenience, this is either a question about why a specific collection of counterfactual statements have the truth values that they have, e.g., why does O persists when O could have ceased to exist?, or a question about a specific regularity, e.g., why does O persist unless positively destroyed from without? Again, friends of Humean Supervenience will say that the truth conditions for counterfactuals or for laws are provided by the distribution of qualities on the fundamental arena and how that distribution covaries across possible worlds. The counterfactual question can then be answered by noting that at the closest possible worlds where O does not persist, O is positively destroyed from without. The regularity question can be answered by noting that in each of the instances where an object ceases to exist, the object was positively destroyed from without. This drives home a lesson for all law-based accounts of inertial persistence and not only those committed to Humean Supervenience: the fact that there is a specific sort of regularity concerning the persistence of objects should be explained in the same way that regularities are explained in general and the fact that objects persist unless specific conditions obtain should be explained in the way that counterfactual facts are explained in general.

However, we may worry about the cogency of strategies that explain inertial persistence in Humean terms. We can think about explanations of inertial persistence in two distinct contexts. First, there is the task of providing an account of inertial persistence regardless of whatever alternatives there might be to existential inertia. Second, we can think about providing an account of inertial persistence that can compete against alternatives to existential inertia. Classical theists provide an alternative to existential inertia, i.e., that a timeless God sustains temporal objects in existence, and that they would object that Humean Supervenience fails to provide an ultimate explanation of the persistence of objects. Humean Supervenience fails to provide an ultimate explanation of persistence, according to classical theists, because, they argue, a hierarchical series that ends in some metaphysically brute object (or objects) cannot provide an explanation for why the entire hierarchy exists.

Whether this objection to the Neo-Humean explanation of inertial persistence succeeds is at least unclear. While objects that occupy space-time regions are temporal, we should not understand the space-time block as temporal; times are explained by the space-time block and not vice versa. (Consider: when is the space-time block located?) Whether we should consider space-time points, or a distribution assigned over space-time points, as temporal is, at least, unclear; certainly, space-time points do not persist. Since friends of Humean Supervenience could understand the space-time block, the space-time points comprising the block, and/or the distribution of qualities as fundamental, friends of Humean Supervenience can secure an explanation of the persistence of temporal objects in terms of non-temporal or non-persistent objects. A fortiori, if the fundamental arena turns out to be non-spatio-temporal, the persistence of temporal objects could again be explained in terms of non-temporal objects. Classical theists may still argue that an explanation has not been provided for why the space-time block, space-time points, or the distribution of qualities exist, but, in that case, the classical theist has foregone an argument from persistence and retreated to an argument from contingency.

While friends of Humean Supervenience would say that the classical theist’s standard of metaphysical explanation is alien to their project and is, at any rate, inappropriate, there is another account of laws for which the classical theist’s objection cannot be applied. That is, there are anti-Humean accounts of laws, according to which there are necessary connections between distinct existences. Anti-Humean accounts of laws are united not only by their denial of Humean Supervenience, but also by the modal status that they assign to laws. Lewis held to the free recombination principle. The free recombination principle is a principle about how the space of metaphysically possible worlds can be formally constructed. Roughly, Lewis held that, given two numerically distinct possible worlds W1 and W2, a third world W3 can be generated by conjoining space-time regions from W1 and W2. (Caveats need to be deployed about the two regions having, e.g., geometrically and topologically compatible boundaries.) If, contrary to Lewis, we accept that there are necessary connections between distinct existences, then the necessary connections will constrain both the space of metaphysically possible worlds (e.g., perhaps, given the nature of fire, there is no possible world where fire exists and fire fails to burn) and the relations between worlds (e.g., given the elevated status that laws enjoy on anti-Humean accounts, nearby possible worlds will be those where the laws are satisfied).

Instead of spelling out the rest of this section in detail, here is a bullet-point list of some of the main takeaways:

  • Law-based accounts: 
    • Necessary connections between distinct existences
      • Maudlin’s primitivist account
        • According to Maudlin, there is a sui generis class of statements that report laws. Unlike Neo-Humeans—who reduce laws to patterns of qualities—or friends of platonic accounts—such as Dretske, Armstrong, and Tooley, who reduce laws to relations between universals— or Aristotelians—who reduce laws to the essences or natures of objects—Maudlin maintains that laws are irreducible and so cannot be explained in terms of anything else.
        • For Maudlin, laws explain causation, counterfactuals, and persistence.
        • Moreover, Maudlin maintains that the passage of time and the persistence of objects from one Cauchy surface to another is explained by the nomic necessities between the two Cauchy surfaces. In some sense, laws produce one Cauchy surface from another and provide constraints both on what does happen and what can happen.
          • Here’s a possible objection: just as the the classical theist objects to the Neo-Humean that the Humean mosaic is brute and so cannot deliver on an explanation of a hierarchical series, why wouldn’t the classical theist say that Maudlin’s primitive laws are yet another collection of brute facts that cannot deliver on an explanation of hierarchical series?
          • Here’s a reply. What Maudlin’s account does accomplish is an explanation of the distribution of items in the mosaic. And when it comes to explaining persistence, this is precisely what’s required. Maudlin’s primitive laws are not further items within space or time and so do not persist in time at all. Our aim in this post is to examine alternative accounts of persistence. And while classical theists may say that they can provide a good account of laws, whether the classical theist’s God is required to explain laws—once conceived of as non-spatio-temporal items—is beyond the scope of this post.
      • Dretske-Tooley-Armstrong (DTA) Account
        • Laws are explained by necessitation relations between abstracta
          • For example, on Armstrong’s version, the nomic regularity that all Fs are Gs is explained by the truth of the statement that N(F, G), where N is a sui generis relation between the universals F and G. [Here’s something that I just learned about Armstrong’s account: for Armstrong, universals are parts of states of affairs and states of affairs are concreta. So, while we could conceive of a view where the necessitation relation is a relation between abstracta, that’s not how Armstrong’s view works. Apparently, Armstrong discusses this issue in his book A World of States of Affairs.]
        • Here’s a possible objection: just as the the classical theist objects to the Neo-Humean that the Humean mosaic is brute and so cannot deliver on an explanation of a hierarchical series, why wouldn’t the classical theist say that the DTA account’s necessitation relations are yet another collection of brute facts that cannot deliver on an explanation of hierarchical series?
          • Here’s a reply. Just as the primitivist can say that their nomological primitives are outside of space-time and so do not persist in time, so, too, friends of the DTA account can say that their necessitation relations are outside of space-time and so do not persist in time. As we’ve said, our aim in this post is to examine alternative accounts of persistence; our aim is not to rebut every possible argument for classical theism. Whether or not classical theism can offer a superior account of the necessitation relation is, again, beyond the scope of this post.
      • Essentialist or Aristoteliean Accounts
        • Laws are explained by the natures or essences had by the objects that fall under the scope of nomic regularities. The chief defender of this view in contemporary philosophy is Brian Ellis—need to get some good citations for Ellis’s work. Note, also, that, at least according to Thomas Kuhn, this is the view that comes closest to that deployed in contemporary physics, albeit without the Aristotelian notion of fundamental (or irreducible) efficient causation.
          • Here’s how Walter Ott explains this sort of account of laws of nature: “Contrast a ‘bottom-up’ view, which holds that the course of nature is fixed by the properties of created beings. On this position, fire’s tendency to burn dry wood under standard conditions is a function of the powers of the fire and wood; to create a world in which all of the conditions are right and yet the wood remains unsinged, God would have to create a world in which neither fire nor wood exist. Once the relevant properties are instantiated, nature takes the course it does simply in virtue of the kinds of things that make it up. On this view, the properties of objects, not the will of God, play the fundamental role. ‘Laws of nature’ will then be nothing more than convenient ways to state relations among these properties.” (Ott 2013, pg 6)
          • This was the view endorsed by early modern French atheists, such as Julien Offray de la Mettrie, Denis Diderot, and Paul Henri Baron d’Holbach. As theologian Michael J. Buckley describes d’Holbach’s view, “Motion as a result or inherent attribute of matter gives natural philosophy its own enclosed world, its own principle, and eliminates the natural theologies of either the religious believer or the deist. Matter carries the attributes of god. It is the necessary being. It is contradictory, inconceivable, to imagine a moment when it did not exist. And since motion is a necessary property of matter, it is coeval with matter” (Buckley 1987, pg 282). Here’s a quote Buckley provides from d’Holbach that I should find in the Systeme: “Motion is a manner [of being], which matter derives from its own proper existence”.
          • From page 300 in Buckley’s 1987: “Clarke’s first three theorems demonstrate three successive predicates about reality. There is a being that is (1) eternal, (2) immutable and independent, and (3) self-existent or necessary. D’Holbach concurs in these attributes, but counters that they inhere in nature or matter. Clarke is right about his predicates; he is wrong about their subject.” Page 301: “That matter is necessary, that is, indestructible, becomes the ground for asserting that it is eternal.”
        • Here’s a possible objection: just as the the classical theist objects to the Neo-Humean that the Humean mosaic is brute and so cannot deliver an explanation of a hierarchical series, why wouldn’t the classical theist say that the Aristotelian account leaves unanswered why the specific collection of essences or natures that are exemplified are exemplified in the first place?
          • Here’s a reply. We can distinguish between two items that are in need of metaphysical explanation. On one hand, why there exist objects that exemplify specific natures or essences is one item in need of metaphysical explanation. A second is why those objects persist in existence once they already do exist. There are two possibilities:
            • At least some physical objects that exist never began to exist. But if that’s so, then nothing is required to explain their beginning to exist; instead, what’s required is an explanation as to why those objects continue to exist. And, given an Aristotelian account of laws, we can explain why those physical objects continue to exist.
            • Every physical object began to exist. In that case, theists may argue that an account should be given as to why physical objects began to exist at all. But even so, our concern in this post is with the persistence of objects. And once a given physical object O exists, the Aristotelian account can explain why O persists.

So much, then, for law-based accounts. We turn next to necessity accounts.

5.4 Necessity Accounts

Necessity accounts explain persistence by appeal to metaphysical necessity.[15] I divide such accounts into two basic kinds: propositional and objectual necessity accounts.

Propositional necessity accounts explain the truth of EIT in terms of its necessary truth. Such accounts therefore adduce the necessary truth of the proposition reporting EIT as an explanation of the proposition’s truth. Why does O (for each O within EIT’s quantificational domain) persist, according to such accounts? Simply because (i) it is a metaphysically necessary truth that, if O exists, O persists unless and until positively destroyed, and (ii) O has not (yet) been subjected to sufficiently destructive factors. Objectual necessity accounts, by contrast, are more metaphysically heavyweight. Such accounts explain persistence ultimately in terms of the necessary existence of one or more temporal concrete objects. Below, I will articulate Schmid’s (2021) propositional necessity account, after which I will develop a novel objectual account.

According to Schmid’s propositional necessity account, inertial persistence is simply “a basic, primitive, foundational feature of reality”, by which he means that it neither analyzes into nor “obtain[s] in virtue of more fundamental/basic facts” (2021, p. 209). By ‘necessary feature of reality’, it is clear from context that Schmid means that EIT itself—that is, the existential inertia thesis—is simply a metaphysically necessary truth.[16] The truth of EIT, then, is explained by the metaphysical necessity thereof; and its metaphysical necessity is not explained in terms of any more basic or fundamental truths.

Does this account accrue a theoretical cost by ending in a primitive or unexplained necessity? It’s not clear that it does. The proposition for which we seek an explanation is that temporal concrete objects persist. Let this proposition be p. Now, the chain of explanations of p is either finite or infinite. If it’s infinite, then it’s unclear that we actually have an adequate explanation of p, after all. If p obtains in virtue of q1, and q1 obtains in virtue of q2, and so on ad infinitum, it seems that we have simply infinitely deferred an adequate explanation of p. This kind of infinite dependence regress, in many philosophers’ eyes, is vicious. So suppose that the chain of explanations of p is finite. In that case, the chain ends in something that is not further explained—that is, it ends in something primitive.[17] In that case, though, it is surely no mark against Schmid’s propositional necessity account that it, too, ends in something primitive, since every explanation of p—if the aforementioned reasoning holds—must likewise end in something primitive.[18]

But perhaps someone will object that Schmid’s propositional necessity account stops the explanatory buck too early. For there are further explanations available—ones that are viable and illuminating. All else being equal, if we have a viable, readily-available, illuminating explanation for x, then we shouldn’t adopt a view on which x is simply brute. We should, instead, explain things as far as we can. Schmid’s propositional necessity account violates this. It explains persistence ultimately by appeal to brute metaphysical necessity. But there is a viable, readily-available, illuminating explanation of persistence in terms of the sustaining activity of something timeless. Like Schmid’s account, this explanation will eventually bottom out in some primitive metaphysical necessity; but unlike Schmid’s account, it doesn’t fall afoul of stopping the explanatory buck too early. What to make of this objection?

I have two responses. First, even if the alternative account adducing timeless sustenance has an explanatory advantage over Schmid’s propositional necessity account, the latter is far more parsimonious than the former in terms of both quantitative simplicity (the number of entities postulated) and qualitative simplicity (the number of irreducible kinds of entities postulated). For the former posits not only more entities (namely, timeless concreta in addition to temporal concreta, as well as relations obtaining between the two) but also more kinds of entities (namely, the new category of timeless concreta). Thus, even granting an explanatory advantage to the timeless sustenance view, it’s not at all clear that it is superior to Schmid’s account all things considered.

Second, it’s not at all clear that the timeless sustenance view enjoys an explanatory advantage. For it only seems to multiply rather than reduce mystery. What, for instance, does the timeless-to-temporal explanatory relation consist in? How can something timeless cause (or ground, or realize, or whatever) temporal things? If a dynamic view of time is correct, wouldn’t the timeless cause change at least in its relational properties, thereby entailing succession in its life (and, hence, temporality)?[19] Moreover, why does the timeless cause seem to make a concerted effort to ensure that objects only cease to exist once they are positively destroyed? This harkens back to Oderberg’s observation: we witness things ceasing to exist when and only when they are subjected to destructive forces. But this seems wholly mysterious if there is an altogether separate way for such objects to cease to exist (namely, a withdrawal of timeless sustaining activity). If the timeless sustainer genuinely could remove its sustaining activity at any moment of an object’s life, it becomes a mystery why this never seems to occur for objects (except when and only when the objects are subject to destructive forces—but even in these cases, it is surely the destructive factors, not the withdrawal of timeless sustenance, that explain the object’s cessation).[20]

Obviously, none of the aforementioned questions amount to objections. My purpose in posing them is to illustrate that the timeless sustenance account seems merely to raise more questions than it answers. It only seems to multiply rather than reduce mystery when it comes to explaining persistence. This is why I say it’s not at all clear that the timeless sustenance account enjoys an explanatory advantage over Schmid’s propositional necessity account.

Having defended Schmid’s propositional necessity account, I now develop a new, objectual necessity account:

Objectual Necessity Account (ONA): There exist one or more concrete objects (call them ‘N’[21]) such that (i) N necessarily exists, (ii) N is temporal, (iii) N is fundamental (i.e. N’s existence is not caused, grounded, or realized by any more fundamental object), and (iv) N’s existence and/or activity directly or indirectly explains (whether by causation, grounding, constitution, or realization) the existence of every non-N temporal concrete object at any moment at which it exists.

ONA is inertialist-friendly because N is one or more temporal concrete objects that persist in existence in the absence of both destruction and conservation from without. Six notes, though, are in order.

First, ONA leaves open the explanatory relation that obtains between N and non-N objects. It allows the relation to be one of efficient causal sustenance, or of grounding, or of constitution, or of realization.

Second, ONA differs from Schmid’s propositional necessity account, since the latter aims to explain the truth of the inertial thesis in terms of the metaphysical necessity thereof, whereas ONA posits one or more necessarily existent temporal concrete objects to explain inertial persistence. ONA is therefore more metaphysically heavyweight (as it were) than Schmid’s account, though also potentially more explanatory for that very reason. (More on this below.)

Third, recall the fundamental question with which we’re concerned: what explains temporal objects’ moment-by-moment existence? ONA provides a satisfying explanation: any non-N temporal object, at any moment at which it exists, is either caused or realized by, grounded in, or constituted by some fundamental thing. Persistence of non-N objects, then, is explained in terms of some kind of conservation or sustenance.[22] All that’s left to explain, then, is the persistence of N. And surely there’s an explanation of that at hand. Why does N persist in existence? Because (i) N is temporal (and so liable to persist), and (ii) N is necessarily existent—it cannot fail to exist and, a fortiori, cannot cease to exist.

Of course, there’s the further question of what explains N’s necessity; but notice that we are now no longer concerned with explaining persistence per se but instead some feature of N. And even if we cannot answer this further question, it doesn’t mitigate the explanatory efficacy of N’s necessity with respect to N’s persistence. The explanation is straightforward and illuminating, even if we don’t have further illumination concerning N’s necessary status. And, moreover, there are a number of (seemingly) viable proposals for explaining N’s necessity. A theistic solution might explain N’s necessity in terms of perfection (Byerly 2019); a non-theistic solution might explain it in terms of (i) an Aristotelian account of modality and (ii) the non-existence of causal powers capable of destroying N or bringing it about that N never existed (Lo 2020); or (if self-explanation is coherent) the necessity might explain itself; and so on.

Fourth, there’s the question of N’s intrinsic nature or character: what might N be? I wish to remain neutral on this. Perhaps it’s a collection of fundamental particles, or mereological simples, or physical simples, or superstrings.[23] Or perhaps it’s some foundational quantum field, or the universal wave function, or some neutral monist substance, or the priority monist universe. Or perhaps it’s the panentheist or neo-classical theist temporal God.[24] Options abound for the inertialist.

Fifth, if there’s a necessarily existent temporal concrete object, doesn’t this entail that the past is beginningless? It does not. Following Craig and Sinclair (2009, p. 189), N may exist timelessly sans the beginning of time and temporally at and subsequent to time’s beginning; or, following Mullins (2014, pp. 164-167) and Swinburne (1993, pp. 208-209), N may pre-exist the beginning of metric time in a non-metric, amorphous time; or N (along with time itself) may begin to exist in all possible worlds (thereby securing both the finitude of the past and N’s necessity); or whatever.[25]

Sixth, what if there cannot be necessarily existent concrete objects? I’m not convinced that there couldn’t be.[26] But suppose necessarily existent concrete objects are impossible. A fortiori, the foremost view that denies EIT (viz. the view that every temporal concrete object is sustained by a necessarily existent, timeless divine cause) is false. Dialectically, then, this is not a strategy open to the foremost detractors of EIT. Moreover, there’s a swift modification of ONA that doesn’t commit to any necessarily existent concrete objects: simply modify ONA to say (roughly) that necessarily, some foundational temporal concrete object or other exists, even though no particular foundational object is such that it necessarily exists.

Once again, I don’t claim that ONA is true. My purpose here is to develop a new, inertialist-friendly explanation of persistence and to situate it within the existential inertia debate at large.

5.5 No-change Accounts

A fifth family of accounts could be termed no-change accounts. These accounts view persistence as an absence of change and take this fact to be central to their inertialist-friendly explanation of persistence. No-change accounts are developed in Rundle (2004), Oppy (Forthcoming, p. 4), and Schmid (2021; Forthcoming).[27] Because Oppy does not develop the account in much detail, we will focus on Rundle’s and Schmid’s renditions.

Rundle (2004, pp. 88-92) holds that persistence is not itself a change and concludes, on that basis, that no continuously operative sustaining cause is needed to explain it. Thus, Rundle writes: “When, we may ask, is a sustaining cause needed? Most obviously, when there is a disintegrating factor to be countered or inhibited, as when a structure will collapse unless it is supported. However, if there is no such threat there is no such need” (ibid, p. 88). He continues:

[B]y revealing certain conditions as not changes in the relevant sense… we rebut the need for a sustaining cause. … [A]s a change, the cessation of a state requires a cause, and this in two possible ways: a cause acts so as to terminate a state, or a sustaining cause ceases its activity. With the continuation of a state, a cause is needed if the state is of the latter kind, involving certain kinds of change within it, or if a cause which would terminate the state needs to be inhibited. This leaves persisting states, as with simply existing, or being oblong, seemingly without need, in general, of any causes of their persistence, the most obvious reason being that persistence does not involve change. If something is still around after many years, this may well be remarkable, but that will be because it has somehow, against the odds, survived threats to its integrity. (Ibid, p. 91)

The above passage indicates a general structure of Rundle’s explanation of persistence. Suppose some temporal concrete object O within EIT’s quantificational domain persists from t* to t, where t* < t. Here is Rundle’s explanation of O persistence through this interval.

Once brought into existence, O could only cease to be in state S if either (i) some cause C positively brings O out of S, (ii) whatever is preventing or inhibiting C ceases its preventative activity, or (iii) the maintenance of S involves continuous changes but the causal source of such changes ceases its activity. But no such cause as C acted to bring O out of existence from t* to t; and no such cause as C was ‘waiting in the wings’, as it were, for some preventative force to cease its prevention and allow C to destroy O; and, finally, the mere maintenance of the state of existence does not involve continuous changes. Hence, none of (i)-(iii) are met—in which case, it follows by modus tollens that O will not cease to exist.

But what explains O’s duration or continuance through time? For, in principle, there are two ways for O to ‘not cease to exist’: O could either exist timelessly or else O could persist. Thankfully, Rundle provides resources for explaining why we have a case of persistence:

[T]o explain why something gets older, where this means no more than ‘continues in being’, is not to give a causal explanation, but to explain what conditions have to be satisfied for a description in these terms to be warranted: you can talk of ‘continuing in being’ because the changes necessary to make sense of the passage of time have been taking place. We are not to read real changes into the object on the strength of these relations… There is not, merely by dint of having endured, any feature of the object that needs accounting for; that need arises only with the real changes which might give the measure of its persistence. The object itself is merely co-present with certain of these. Not a change in it, but a mere change in a relation. (2004, pp. 91-92)

Rundle, then, offers a fleshed-out, illuminating, and prima facie defensible no-change account. Objects persist without causal sustenance because (i) only changes of state require causes, (ii) mere persistence is a state characterized by the absence of change, (iii) no sufficiently destructive causal factors have thus-far impinged on them, and (iv) the objects changes in its relations to constant, repeated changes elsewhere. Even if one disagrees with one or more theses here, (i) if they were true, they would surely remove mystery as to why things persist, and (ii) they are at least rationally defensible.

Schmid’s no-change account is closely related to Rundle’s. The account—adumbrated in Schmid (2021, fn. 19) and developed more fully in Schmid (Forthcoming)—makes two core claims: (i) existence is a state or condition of stasis or unchangingness, and (ii) states of stasis by their very nature diverge to another state (or no state at all) only when positively disrupted.

The account, as Schmid (Forthcoming) develop it, takes its cue from one of Edward Feser’s foremost ways of reconciling mechanical inertia with the Aristotelian-Thomist causal principle (CP), according to which whatever changes is changed (actualized) by something else in a state of actuality. Prima facie, CP is incompatible with mechanical inertia, according to which an object can uniformly move from point A to point B without any distinct actualizing cause of said change. One of Feser’s principal ways of reconciling the two treats uniform spatial motion as stasis or unchangingness rather than as the transition from potency to act. Feser writes:

[P]recisely because the principle of inertia treats uniform local motion as a ‘state,’ it treats it thereby as the absence of change. … In this case, the question of how the principle of motion and the principle of inertia relate to one another does not even arise, for there just is no motion (in the relevant, Aristotelian sense) going on in the first place when all an object is doing is ‘moving’ inertially in the Newtonian sense. To be sure, acceleration would in this case involve motion in the Aristotelian sense, but as we have seen, since Newtonian physics itself requires a cause for accelerated motion, there is not even a prima facie conflict with the Aristotelian principle of motion. (2013, pp. 239, 250-251)

As Feser points out, it seems entirely justifiable to understand uniform spatial motion as a state of stasis or unchangingness. “But just as we can understand uniform spatial motion as stasis or unchangingness,” writes Schmid, “it seems we can equally justifiably understand persistence in existence as an absence of change” (Forthcoming). “Indeed,” continues Schmid, “persisting in existence is commonly thought not to involve change but rather the maintenance of a state of actuality. … [W]e tend to think only that deviations from something’s state of non-existence or existence count as changes (i.e. either coming into or passing out of being)” (Forthcoming).[28]

Schmid’s account seems to employ a kind of Principle of Sufficient Reason (PSR) according to which dynamic changes of state require explanation in terms of some extrinsic cause or actualization of such a change, whereas maintenance or non-disruption of a state does not require an explanation in terms of some extrinsic cause which keeps the state non-disrupted or maintained. Instead, the maintenance or non-disruption of a state is explained in terms of the very nature of states (stable, unchanging actual conditions which are retained unless positively disrupted) in conjunction with there being no such disrupting factors operative.[29]

It is worth emphasizing that this view decidedly rejects there being any brute facts concerning the moment-by-moment existence of things. There is an explanation of (inertial) persistence; it is in terms of (i) an understanding of existential persistence as a state of actuality as opposed to a dynamic transition from non-being to being at any given moment, (ii) the very nature of states as a kind of stasis or unchangingness (á la Feser’s account of spatial motion), and (iii) there being (or having been) no sufficiently destructive factors that induce deviations from the state in question (viz. persistent existence).

In summary form, Schmid’s no-change account says that persistent existence is a state of unchangingness or stasis.[30] By their very nature, states of unchangingness deviate from their actual condition only if there is some positive disruption of their condition.

A third no-change account is as follows. For simplicity, I’ll assume endurantism and Newtonian or Galilean space-time. Nothing hangs on this, though—an exactly similar account can be cast in perdurantist and/or Relativistic space-time terms mutatis mutandis. (Again, my goal is not to prove or positively justify the account; instead, I’m simply articulating an account that—if true—furnishes an inertialist-friendly explanation of persistence. I only need the account to be rationally defensible.)

The account is as follows. For temporal concrete object O to fail to exist at t despite existing from [t*, t), where t* < t, is for some change to occur.[31] But, plausibly, a change occurs only if some factor causally induces said change. It’s not as though a raging tiger, or thousands of photons, or whatever could just spring uncausedly into existence in this room right now; that would require some cause.[32] Moreover, it’s not like New York, or my house, or this article could pop out of existence uncausedly; such changes would require some cause. From this observation it follows that if no factor causally induces a change, then the change won’t occur. Thus, if no factor causally induces O to fail to exist at t despite existing from [t*, t), then O exists at t. Once we add that nothing came along to causally induce this—that is, once we add that nothing came along to destroy O from t* to t—it simply follows that O exists at t. Here, we seem to have a perfectly respectable, perfectly legitimate explanation of O’s existence at t, and this explanation does not presuppose the prior obtaining of the explanandum (i.e., O’s existence at t).

The account above, as it stands, needs to be supplemented. For starters, it is (so far) perfectly compatible with EIT’s being false. For suppose that O’s persistence requires continuous sustenance from without. Now suppose that the absence of sustenance from without can cause O’s cessation. In this case, it is false that the absence of sufficiently positively destructive factors operative is by itself sufficient for there being no cause of O’s cessation at t, and hence the latter claim is insufficient to conjoin with the fact that <there is no cause of O’s cessation at t, and O fails to exist at t despite existing from [t*, t) only if there is a cause of O’s cessation at t> in order to explain the fact that <O exists at t>. To avert this problem, I add to the account that absences are not causes. By my lights, there are no such things as absences, and hence there are no such things as absences that are also causes.

Another supplement is in order. For while one might grant that the absence of sustenance from without is not a candidate cause of O’s cessation, perhaps the event of the withdrawal of sustenance from without could play that role. This will threaten the account in the same way that the consideration about absences did. There are several ways I might supplement the account to avert this problem. I might add that while (some) events exist, withdrawals are not among them. (Do we really want to countenance withdrawals into our ontology?) Or I might add that there are no such things as events. Yes, there are objects, and such objects have, gain, and lose various properties.[33] But are there such things as events corresponding to such havings, gains, and losses, over and above the objects and properties? It’s not obvious, and my account can be supplemented with a negative answer. There may be other supplements besides, but that suffices for present purposes.[34]

Finally, I will address the following objection to no-change accounts: No-change accounts are inapplicable to substances that are undergoing real (intrinsic) continuous change. If the substance is continuously intrinsically changing, then there is no ‘state of existence’ that can persist from one moment to later ones. Surely, the state of existence of a substance includes all of its real features. If this state is continuously changing, something external to the substance must then be responsible for the substance’s persistence through such a series of distinct real states.

In response, I note that no-change accounts do not hold that the total set of an object’s real features comprises the object’s ‘state of existence’. Instead, it only holds that the very being (existence, actuality) of the object as such is a state or condition of stasis. The object’s very being or existence is distinct from its continuously changing (accidental) real features, since the object can retain its existence (i.e. persist as one and the same substance) despite the gain or loss of such real (accidental) features. Hence, the very existence or actuality of the object is not the same as the total set of the object’s real features. (NB: I’m not here committing to the thesis that there is such a thing as an object’s being or existence.) And it is this very substantial existence or actuality that the accounts in question treat as a state or condition of unchangingness.

Here’s another way to think about it. Consider again the case of mechanical inertia. One of Feser’s reconciliations of CP with mechanical inertia is to treat uniform rectilinear spatial motion as a mere state or condition of unchangingness. In this way, it does not involve the kind of dynamism that CP says requires a distinct actualizing cause. But when an object is uniformly rectilinearly moving through space, it may be undergoing other, continuous processes of real change. These would—granting the truth of CP—require a continuously operative actualizing cause. But this does not negate or nullify the fact that the object is still in a state or condition of unchangingness with respect to its uniform rectilinear motion. And so, in this respect, the object does not require a continuously operative cause, though the other respects in which the object is really changing do (or would). And this is precisely the proposal of no-change accounts: although the object may be undergoing a host of real accidental changes (and thus may require continuous causes of such changes), the object is not undergoing the kind of real dynamism or change in respect of its actuality or existence that would require a continuously operative sustaining cause.

Thus, saying that such continuously occurring real (accidental) changes require causes is crucially different from saying that the very substantial being of the object as such requires a continuously operative efficient sustaining cause. Those continuous real changes only demand continuously operative causes that merely modify the accidental features of the object. But this is wholly separate from the need for an efficient sustaining cause of the entire object’s existence at every moment at which it exists (or at every time at which it has a temporal part).[35]


[1] Schmid (2021) makes this point in connection with the criticisms of Feser (2011). I’ve also made the point in my interactions with Feser. Cf. the two posts here and here.

[2] Unlike Beaudoin and Benocci, Oderberg doesn’t explicitly defend existential inertia. Nevertheless, Oderberg defends a tendency of concrete temporal objects to persist or continue in existence, and this can quite clearly be employed in the context of existential inertia.

[3] We can set aside the vast literature on masks and finkish tendencies/dispositions (cf. Martin 1994), since I’m not concerned here with an analysis of tendencies/dispositions. For present purposes, we can either (i) simply build into the account that for all normal cases, persistence and destruction don’t succumb to the presence of finks, masks, and the like (which is a plausible assumption, since the examples of finks and masks are often outlandish or non-ordinary); or (ii) restrict the domain of quantification of the inertial thesis (and, consequently, the inertial tendency/disposition espoused in tendency-disposition accounts) to cases where finks, masks, and the like are absent; or (iii) simply include within C the specification that no masks/finks/etc. are operative; or etc.

[4] Note that Oderberg uses ‘tendency’ and ‘disposition’ interchangeably (2014, p. 349).

[5] Here is another way to think about it. Suppose that O is a point-like particle whose life begins at time t*. In that case, there are two things we need to explain: (i) O’s beginning to exist at t* and (ii) O’s continuing to exist to (say) t. For (i)—that is, O’s beginning to exist—whatever explains O will not be O itself. Presumably, the explanation will be something that brought O into existence. But we are interested here in explaining persistence, not beginning. So, let’s turn to (ii). What explains the fact that O continues to exist to t? Plausibly, it could be explained by some fact about O at one or more times earlier than t (in conjunction, of course, with a few other conditions, to wit, those adumbrated in the main text).

[6] Schmid (2021, p. 212) makes a similar point: “EIT also explains why, in our experience, we only see objects cease to exist when some causal factor impinges on them in a destructive way. This is precisely what EIT predicts: objects must continually persist in existence and can only cease to exist when some causal factor destroys them.”

[7] Also of note is Oderberg’s (2014, pp. 351-353) defense of the tendency to persist in light of an objection from radioactive decay. Schmid (2021, p. 204) likewise addresses this objection. I’ll address it in Section 7 of this post.

[8] Indeed, those unsympathetic to existential inertia—that is, those who hold that all temporal concreta are sustained or conserved in existence from without—are entirely free to recognize the legitimacy of transtemporal explanations. Consider, for instance, Pruss: “Plausibly, your existence at earlier times causes your existence at later times” (2018, p. 167). For further discussion/justification of transtemporal explanatory relations, see Schmid (2021) and Swinburne (1994, pp. 81-90). Note, though, that transtemporal accounts do not say that all causation is non-simultaneous. (Indeed, they are not necessarily wedded to viewing the relevant transtemporal explanatory relation as causal, as we’ve seen.) Final note: one objection to transtemporal accounts derives from the conjunction of presentism and the thesis that explanatory relations are existence-entailing. I respond to this objection in Schmid (2021).

[9] Some recent developments that may favor discrete time are based on causal finitism, the view that every event, state, or substance has a finite causal history. (For sustained arguments for causal finitism, see Pruss (2018) and Koons (2014)). For if time were continuous, then presumably infinitely many distinct states could causally influence (whether directly or indirectly) a final state (e.g., the location of a particle at time t would be dependent upon or causally influenced/explained by the particle’s location at each of infinitely many distinct previous times; or my existence at each of infinitely many distinct earlier times causally contributes (directly or indirectly) to my present existence; etc.). Pruss (2018, ch. 8) argues along different lines that while causal finitism does not automatically entail discrete time, it still provides evidence for it.

[10] For an extended treatment of issues pertaining to the continuity of time, divine conservation, and explaining persistence, see Miller (2007). Miller (2007, ch. 3), in particular, criticizes the ability of transtemporal causal relations to explain persistence if time is continuous. If successful, his arguments could be applied to the present context (mutatis mutandis) to show that transtemporal accounts of inertial persistence require discrete time. Miller’s case is in some ways modeled after the arguments for the impossibility of self-sustenance and for the impossibility of a diachronic causal nexus found in Kvanvig and McCann (1988) and McCann and Kvanvig (1991).

[11] I use ‘successive stages of an object’s life’ as neutral between endurantism and perdurantism.

[12] The ‘causation’ in question, moreover, can be variously understood. We might, for instance, follow D. H. Mellor’s discussion of the causation of stasis in Mellor (1998, ch. 9). Mellor distinguishes between factual and particular causation, arguing that something’s remaining unchanged between times t and t’ is merely an instance of factual causation. See also Saudek (2020, p. 92) on this point.

[13] For a defense of the simultaneity of causation, see Mumford and Anjum (2011, ch. 5) and Ingthorsson (2021, chs. 4 and 5). DJ Linford has written the follwoing reply to the rejoinder from simultaneous causation, not Joe.

[14] The argument that the foot is the eternal cause of the impression is worse off when we consider, e.g., counterfactual accounts of causation. At the possible worlds closest to the world containing the co-eternal feet and impression where the foot is absent, the impression remains. And so, on the counterfactual analysis of causation, the foot is not the cause of the impression.

[15] Hereafter, by ‘necessity’, ‘possibility’, etc. I mean metaphysical necessity, possibility, etc.

[16] Schmid describes his account, for instance, as one according to which EIT is “a primitive necessity” (2021, p. 210). The account is thus clearly a propositional necessity account.

[17] One might think that the terminus of the chain of explanations—while not explained by any further proposition—is nevertheless self-explanatory. I have two responses. First, I find it plausible that nothing can explain or account for why it itself is true (or obtains, or exists) at all. For in order to have any explanatory power in the first place, it would ‘already’ (as it were) have to be true (obtain, exist). (‘Already’ expresses not temporal but ontological priority.) Plausibly, one cannot merely presuppose the very thing for which one sought illumination. (Also, just imagine: if I ask why q is true, it’s no use responding ‘because q is true’. Surely, I say, this explains nothing!) But suppose I’m wrong about the impossibility of self-explanation. Suppose, in other words, that some propositions can explain their own truth. Presumably, this will involve the proposition having a kind of ‘intrinsic intelligibility’—once the proposition is grasped, there is no mystery concerning why it is true. This brings us to my second response: plausibly, no explanation of persistence will be self-explanatory—in which case, each explanation of persistence will end in an unexplained primitive. And in that case, the fact that Schmid’s propositional necessity account ends in an unexplained primitive is not a mark against it as an explanation of persistence. Consider explanations of persistence that deny EIT. Such explanations either adduce one or more atemporal concrete objects that sustain temporal objects or else an infinitely descending chain of more fundamental temporal concrete objects sustaining or conserving less fundamental temporal concrete objects (of which the former aren’t parts). But surely neither the proposition <there is such an atemporal concrete object> nor <there is such an infinitely descending chain> are self-explanatory.

[18] I could go further: this primitive is either contingent or necessary. But contingent propositions—precisely because they genuinely could have been false—seem to call out for further explanations of why they are true. (Cf. Pruss (2006) for an extended defense of the thesis that every contingent proposition has an explanation.) Plausibly, then, the primitive must be necessary. But then any explanation of p terminates in a primitive necessity, and so surely it’s no mark against Schmid’s propositional necessity account that its explanation of p terminates in a primitive necessity.

[19] For example, assuming that x is causally sustained by the timeless cause, the timeless cause will arguably acquire a relational (causal) property once x begins to exist and lose a relational (causal) property once x ceases to exist. To be sure, such changes need not be intrinsic to the timeless thing. There are plenty of cases where a subject gains or loses some relational property without undergoing intrinsic change. (A father might become shorter than his son solely because his son has grown.) But even in such cases, the subject of the extrinsically relational change is temporal, since it can only gain or lose the relational property if there are distinct moments m and m* of its life at which the relational property is had and then lacked (or vice versa).

[20] One might say that the timeless sustainer is necessitated to engage in such sustaining activity. But why would that be? Is this just a brute or primitive necessity? Going this route will similarly multiply mysteries, it seems. (Detractors of EIT of a theist stripe will likely resist this response, too, as it seems to remove the timeless God’s freedom.)

[21] I will use ‘N’ as a singular noun, but I stress that it stands for one or more objects of the kind articulated in ONA.

[22] If the relation between N and non-N objects is one involving constitution, then N will not be an outside or external sustaining or conserving explanation of the persistence of non-N objects—in which case, given our articulation of EIT, both N and non-N objects would inertially persist. But if the relation is one that doesn’t involve N constituting non-N objects, then only N will inertially persist. In either case, though, EIT is true.

[23] For a recent defense of an existential inertialist view on which the foundational temporal reality is a collection of enduring mereological simples, see Benocci (2018).

[24] For an extended defense of divine temporality, see Mullins (2016).

[25] Note that while the proposals of Craig and Sinclair, Mullins, Swinburne, and co. are theistic, there is nothing inherently theistic about them; they can serve as models of N’s relation to time even if N is one of the non-theistic options articulated earlier (e.g. one or more fundamental quantum fields).

[26] For a variety of arguments for at least one necessarily existent concrete object, see Pruss and Rasmussen (2018). See especially ch. 9, wherein arguments for the impossibility of necessary concreta are addressed.

[27] Rundle does say that “in the absence of forces which would bring them to an end, [objects’] continuation from moment to moment is in no need of explanation” (2004, p. 93). Never mind Rundle’s precise view; we are concerned with what resources Rundle’s reflections provide in explaining persistence. For simplicity, we use ‘Rundle’s explanation’ rather than ‘an explanation that can be gleaned from Rundle’s reflections’.

[28] Two notes. First, in this context, ‘something’s state of non-existence’ does not mean that there is an x such that x is in a state of non-existence. Instead, it means only that there does not exist an x such that x is in a state of existence. So, it would be more accurate to say ‘non-state of existence’ instead of ‘state of non-existence’. For ease of exposition, I stick with the latter. Second, I mean ‘change’ broadly construed so as to include coming into and going out of existence.

[29] Schmid echoes these commitments in his [or, rather, my…] account of one necessary condition for the need for a sustaining cause: A “sustaining cause C is required for substance S’s being in condition or outcome O only if (i) there is some causal or explanatory factor or force F—intrinsic or extrinsic to S—acting on S to bring S toward some condition or outcome ~O; (ii) F is a net factor or force in the absence of C’s causal operation; and (iii) S (or some state of affairs involving S) is in condition or outcome O distinct from ~O” (Forthcoming). According to Schmid, an “intrinsic causal factor or force would be something like a natural tendency, inclination, or disposition inherent to a thing”, whereas “an extrinsic one would be something like the effect of gravity, friction, and so on” (Forthcoming). And by ‘net force’, Schmid means “a causal factor or group of causal factors whose overall causal contribution is like a vector quantity insofar as it contributes toward a definite end state or outcome and is not counterbalanced by some other (group of) causal factor(s)” (Forthcoming).

[30] This is not to deny objective temporal passage, just as treating uniform local motion as a state of unchangingness or stasis does not deny that things genuinely go from one spatial position to another spatial location.

[31] Of course, it’s not as though O undergoes some alteration in this process. But still, it is obvious that there is some kind of change here, whether in the ontological inventory of what there is, or in the incorporation of what were previously O’s parts into parts of something else, or in the passing away of a state (e.g., the state of O’s existence), or whatever. If the reader still demurs at my use of ‘change’, I can alter the account to speak of changes of state (i.e., cases where some new state comes to be or some old state passes away).

[32] The cause might be indeterministic, as when (under some interpretations of quantum mechanics) virtual particles nondeterministically arise from prior, law-governed quantum states.

[33] Alternatively, they satisfy, begin satisfying, and cease satisfying various predicates.

[34] Furthermore, if the external sustainer is God (classical conceived), an immediate problem for the proposal that there is an event of God’s withdrawal of sustenance from O. For under classical theism, anything distinct from God is created and sustained by God (Rogers (1996, p. 167), Bergmann and Brower (2006, p. 361), Grant (2019, ch. 1), and Kerr (2019, pp. 15 and 44)). (And God, after all, is surely not identical to a withdrawal!) But surely whatever x that God creates and sustains is such that God is free to withdraw his sustaining activity of x. So, God could withdraw his sustaining power from the relevant event. But then there would be the event of God’s withdrawing this sustaining power from the event of God’s withdrawing his sustaining power of O. But then we could run the exact same reasoning ad infinitum. So, there could be an infinitely descending chain of divine-withdrawal events. But that seems absurd.

[35] Here are two further responses. First response: The continuous processes of change are only to the object’s accidental features, and thus only require causes that simply modify the object in various ways. But through such accidental changes, the object’s essential features remain unchanged and intact. And, surely it is this unchangingness that matters for existential inertia, since the object persists so long as its essential properties remain unchanged and intact! Second response: Essentially, I could grant that no-change accounts are inapplicable to objects undergoing the kind of continuous intrinsic change specified by the objection. But this does not undermine EIT, since EIT only applies to temporal concrete objects or some subset thereof. Thus, so long as some subset of temporal concrete objects (at the foundational layer of temporal reality, say) do not undergo the kind of continuous intrinsic change that this objection specifies, then no-change accounts of EIT can avoid the objection in question. Whether or not there are such temporal concrete objects is a separate question, of course. But it is significant in its own right if we have a viable metaphysical account of inertial persistence that applies (or would apply) to such objects.

6 Motivating EIT

In the following sub-sections, I sketch various motivations favoring EIT.

6.1 Theoretical Virtues

Before considering EIT’s theoretical virtues, let’s clarify the rival hypothesis in this dialectical context:

Classical Theistic Sustenance Thesis (CTST): All temporal concrete objects (would) immediately annihilate absent sustenance or conservation from without, such that a necessary condition for their persistence is the moment-by-moment sustaining or conserving activity of a timeless God.

I think EIT—or, more accurately, various metaphysical accounts of EIT—enjoys a number of theoretical virtues (at least in relation to CTST). Before considering the first virtue of EIT, however, let’s consider the following question: why does anything exist at all? Why isn’t reality just blank? Here is a simple answer: something exists rather than nothing because it is metaphysically necessary that something exists. This answer nicely explains why there is something rather than nothing: it is simply metaphysically impossible for there to be nothing.

A similar question arises with respect to concrete objects’ persistence. Why do objects, once in existence, persist in existence instead of being instantly annihilated or annihilated at random, arbitrary points during their existence? If one affixes a necessity operator in front of the EIT, EIT provides a simple answer: objects persist rather than succumb to instant or random annihilation because it is metaphysically necessary that they do so (absent causally destructive factors). (Even if one doesn’t so affix a necessity operator, certain metaphysical accounts (e.g., objectual and propositional necessity accounts) themselves provide metaphysical necessity as an explanatory avenue.) This answer, analogous to the one concerning existence simpliciter, nicely explains why objects persist rather than chaotically being annihilated: the latter is simply metaphysically impossible.

Note that I’m here seeking to explain a contrastive fact: why do objects, once in existence, persist in existence instead of being instantly annihilated or annihilated at seemingly random, arbitrary points during their existence? What explains why objects persist rather than chaotically being annihilated? We are thus contrasting two situations: one in which things uniformly and reliably persist absent sufficiently destructive causal factors, and another in which things annihilate at seemingly chaotic or random points in their lives without some sufficiently destructive causal factor operative. We claim that the necessary truth of EIT nicely explains why we see the former rather than the latter.

And I also claim that this is a clearly satisfactory explanation of this contrastive fact. Consider that in debates surrounding the Principle of Sufficient Reason (PSR) and various causal principles (like those adduced in cosmological-type arguments), defenders of such principles will say something like ‘why is it that we see a uniform, regular, intelligible reality rather than things just chaotically popping into and out of existence uncausedly and inexplicably? Why isn’t there chaos instead? Why doesn’t Beethoven, or a raging tiger, or whole swathes of photons just spring into existence without an explanation or cause? Here’s a nice explanation: these don’t happen because such unexplained, uncaused occurrences are simply metaphysically impossible. And this is just to say that the (necessary truth of the) PSR or the causal principle (variously construed) provides a satisfying explanation of the data.’[1]

I find this to be a perfectly kosher explanation. And yet the explanation is structurally identical to our EIT-based explanation of the aforementioned contrastive fact. (Just replace ‘PSR/causal principle’ with ‘EIT’ and ‘chaotically ceasing and beginning to exist’ with ‘chaotically ceasing to exist’.)

Indeed, things get even more interesting when we do a Bayesian comparison between CTST and the necessary truth of EIT (hereafter, n-EIT) when it comes to explaining the contrastive fact above. Let that fact be F. Now, recall F: temporal objects, once in existence, uniformly and reliably persist absent sufficiently destructive causal factors as opposed to being annihilated at other points in their lives without being subject to causal destruction. For under the n-EIT, the probability of F is 1. In other words, P(F|n-EIT) = 1. But the probability of F is significantly less than 1 under CTTS. In other words, P(F|CTST) << 1. And hence F strongly confirms n-EIT vis-à-vis CTST.

Why do I say P(F|CTST) << 1? Consider that for an object to persist for about 116 days is for it to persist approximately 10,000,000 seconds. And consider that there are about 1044 Planck times in one second, and hence to persist for about 116 days is to persist for about 1051 Planck times. Now, for each (possible) Plank time in an object’s life, God is free to sustain the object for that Planck time or refrain from so sustaining it, and this is true even if God has sustained the object through previous Planck times. (It’s not as though God suddenly becomes forced or compelled to continue sustaining an object, and it’s not as though some moral obligation is newly imposed on him, or whatever. These are unacceptable by the classical theist’s lights.) This entails that for a given world w in which an object persists for (at least) 116 days and is then causally destroyed by something, there are about 1051 worlds in which that object ceases to exist at some point earlier than w due to God’s withdrawal of sustenance. Moreover, under classical theism, God remains utterly invariant across all worlds, and hence there is no difference or variance on God’s end to which one could point in any such world to explain why it came about as opposed to some other world. Thus, conditional on CTST, the expectation that w obtains—that is, that the single world in which the object persists for 116 days (upon which it is finally causally destroyed) obtains rather than any of the other 1051 worlds—is very small. And hence the data of F, as applied to this 116-day-lived object, is very unlikely conditional on CTST, whereas its probability is 1 conditional on EIT. And hence EIT garners very strong evidential confirmation from F vis-à-vis CTST. Much more can be said on this line of reasoning, but that suffices for a brief sketch (and, hence, it suffices for the purposes of this section.)

From the above, I’m left to conclude that EIT not only provides an illuminating explanation of F (just as the PSR/causal principle does with respect to the absence of chaotic beginnings and cessations) but also gains substantial evidential confirmation from F vis-à-vis CTST.

EIT also explains why, in our experience, we only see objects cease to exist when some causal factor impinges on them in a destructive way. This is precisely what EIT predicts: objects continually persist in existence and only cease to exist when some causal factor destroys them. On the other hand, if objects genuinely could be instantly (or randomly) annihilated, then we have a puzzle: why does nothing in reality behave this way? Why don’t such apparently chaotic ceasings-to-exist pervade reality?

What’s more, EIT seems to provide the best explanation for why we don’t observe sustaining causes in the world around us—or if we do, why they are so rare. Consider your present existence. Do you have a sustaining efficient cause?

Perhaps you think oxygen, heat, air, and so on are present sustaining causes of your existence, or that the fact that your computer is three feet above ground is presently causally dependent on the table. “But,” notes J.H. Sobel,

I am dependent on these things only eventually for my future existence… Take away oxygen and I am dead, not now, however, but only shortly. Take away heat from my environment, plunge it to absolute zero, and I am gone more quickly, but again not immediately. Take away the sun, and the heat, most of it hereabouts continues for eight minutes or so, so the sun is no part of its efficient sustaining cause. Oxygen and the like are at best not sustaining, but perpetuating, and so not necessarily concurrent efficient causes of people. (Sobel 2003, p. 177)

If oxygen, heat, and so on were sustaining causes of your existence, then you would instantly die in their absence. But upon removing oxygen, heat, and so on, you do not die instantly but only after a short period of time. Therefore, oxygen, heat, and so on are not sustaining causes of your existence. (Keep in mind, moreover, that EIT only denies (for objects within its quantificational domain) conservation or sustenance from without. Hence, the fact that you are dependent on certain components of you (e.g., your brain, your bones, the blood and sugar and lipids in your body, the oxygen in your lungs, etc.) is neither here nor there.) In fact, the only reason why removing oxygen from the external environment will eventually result in your death is that it facilitates the absence of oxygen within you, which in turn facilitates cell death within you. And it is this that kills you. But we have already seen why your dependence on factors within you is perfectly compatible with EIT.[2]

In fact, even if oxygen were such a sustaining cause, it’s not at all clear that there is anything external to oxygen that causally sustains it in being. (Note that the covalent bond holding the two oxygen atoms together is within the oxygen molecule and consists in the sharing of electrons therein.) Hence, even if we could pinpoint a few examples of sustaining causes of existence, they are rare. Finally, the causal activity of the table is nothing to the very existence of your laptop. The table does not sustain your laptop as an efficient cause of its very being. This example, then, also fails. Overall, then, it seems we lack good experiential reason to affirm the existence of sustaining causes of existence—precisely what we would expect if EIT were true (at least assuming a version of EIT with a wide quantificational domain).

EIT also makes good sense of our ordinary, everyday explanatory reasoning as well as our scientific reasoning. Why did the banana remain on the table (and in existence) from yesterday to today? Because (i) it was on the table yesterday, (ii) no one moved it off the table, and (iii) nothing came along to destroy it. This seems like a perfectly adequate explanation, and it seems to have as a (tacit) component that the banana persists in existence so long as nothing destroys it. We rarely cite (or even see any need of citing) sustaining causes of existence to explain why objects persist. EIT seems to make the best sense of such explanatory practices. For if objects genuinely would utterly annihilate absent existential causal sustenance, any sufficient explanation would plausibly have to take such causal sustenance into account—and the fact that we seem to provide sufficient explanations without adducing existential causal sustenance is defeasible evidence that we do, in fact, provide such explanations and hence is defeasible evidence in favor of EIT vis-à-vis CTST.

Furthermore, a key determinant of prior or intrinsic probability is simplicity. And, indeed, EIT enjoys both quantitative and qualitative simplicity (and enjoys them far better than CTST). For one thing, CTST requires a categorically different kind of causation in our ontology (namely, sustaining causation). EIT does not require this. For another thing, CTST requires a categorically different kind of being in our ontology. In particular, CTST requires at least one timeless being that sustains temporal things in existence. It is thus committed not only to all the temporal entities EIT is committed to, but it also includes more entities (a timeless entity), more kinds of causation (sustenance or conservation from without, as well as timeless-to-temporal causation), and more fundamental (i.e., irreducible) kinds of entities—reality is fundamentally divided at least into timeless/immutable concrete things and temporal/immutable concrete things. This makes CTST much more complex than EIT and hence much less intrinsically probable.[3]

Overall, EIT explains and unifies a whole host of disparate phenomena: (i) why objects persist instead of being annihilated at seemingly chaotic points, (ii) why objects cease to exist only when causally destroyed, (iii) the dearth of observational evidence of sustaining causes, and (iv) the legitimacy of ordinary and scientific explanation. It also enjoys both quantitative and qualitative simplicity. This leads me to conclude that EIT is superior to CTST both on account of explanatory breadth and depth as well as simplicity.

6.2 Aristotelian proof argument

Edward Feser provides a wide variety of arguments and considerations in favor of his Aristotelian proof’s causal principle (CP). According to CP (captured in premise (4)), no potential can be actualized unless something already actual causally actualizes it. Suppose CP is true, and suppose object O exists at time t. Now, for O to go out of existence at some arbitrary t’ (such that t’ > t) is for a change to occur—it is for O’s potential to cease to exist to become actual. But since a potential can only become actual if something already actual causally actualizes it, it follows that O could only go out of existence at t’ if something already actual causally actualizes this. So, if nothing already actual casually actualizes O’s going out of existence at t’, O will not go out of existence at t’.

But remember that t’ represents any arbitrary time later than t. Hence, if nothing already actual casually actualizes O’s going out of existence from [t, t’] for any t’ > t, O will not go out of existence from [t, t’]. Hence, if nothing already actual casually actualizes O’s going out of existence from [t, t’] for any t’ > t, O will persist from [t, t’].

But absences are not actual things—they are precisely the absence or non-existence of actual things. A fortiori, an absence of sustenance or conservation from without is not an actual thing. Hence, an absence of sustenance from without cannot causally actualize O’s going out of existence at t’, precisely because only something already actual can causally actualize potentials (per CP), whereas absences are not actual.

Thus, if no positive reality[4] causally destroys O from [t, t’] for any t’ > t, O will persist from [t, t’]. But this is just to say that EIT is true. According to EIT, temporal objects persist without external sustenance so long as they aren’t positively destroyed. We already saw that the mere absence of sustenance couldn’t be an ‘already actual cause’ of O cessation at any time t’ later than t. Thus, so long as no positive reality comes along and positively destroys O, O will persist to t’ even in the absence of sustenance from without. So, CP seems straightforwardly to entail EIT.

We thus have an argument for EIT:

  • If CP is true, EIT is true.
  • CP is true.
  • Therefore, EIT is true. (1, 2)

Let’s consider some objections.

Objection. Recall one of the central claims of my argument above: for O to go out of existence at some arbitrary t’ (such that t’ > t) is for a change to occur—it is for O’s potential to cease to exist to become actual. But surely if change is the movement from potentiality to actuality, then ceasing to exist cannot be a change—non-existence is not a state of actuality. That is, a ‘potential to cease to exist’ is impossible since ceasing to exist (or non-being) is not some actuality to which O’s potency could point. Ceasing to exist simpliciter is not a change in the relevant respect (a potentiality becoming actual).

This, in essence, is the same (or at least a similar) problem Schmid (Forthcoming) argues afflicts Feser’s Aristotelian proof. Prior to O’s existing, it cannot have any potentials, including a potential for existence. Naturally, then, its potential for existence cannot be caused to reduce from a state of potency to a state of actuality. But by the same token, posterior to O’s existing, it cannot have any potentialities or actualities, including a potentiality for or actuality of non-existence. Naturally, then, its potential for non-existence cannot be actualized.

Reply. I have three replies. First, there are ways to avoid the objection at hand. For instance, perhaps we can run the argument in terms of propositions. In particular, it seems plausible that all the motivations for Feser’s CP equally well support the following principle: whenever a proposition P changes in its truth value, there is an explanation (for that change) reporting either (i) that something already actual causally brings about P’s worldly correlate (in the case of P’s changing from false to true) or (ii) that something already actual causally removes P’s worldly correlate. Let’s call this principle the Propositional Causal Principle (PCP). PCP is extremely similar to Feser’s CP, since PCP says that changes in propositions’ truth values require causal explanations that either bring about or remove that to which the proposition corresponds.

Now let P be the proposition <O exists> and suppose P is true at t. By PCP, P can only be false at any t’ (where t’ > t) if something already actual causally removes P’s worldly correlate. But P’s worldly correlate is O (or O’s existing). So, P can only be false at any t’ if something already actual causally removes or destroys O. Hence, if nothing already actual causally removes or destroys O between t and t’, P remains true from t to t’. But the proposition <P remains true from t to t’> entails that O persists from t to t’. So, if nothing already actual causally removes or destroys O between t and t’, O persists from t to t’. And that, it seems, delivers EIT. (Bearing in mind, of course, the points about absences not being actual.)

Second, it seems plausible that we could use ‘possibility’ instead of ‘potentiality’ to run the argument. For even if there is no potentiality for O’s existence/non-existence in the sense of some disposition pointed towards an outcome that can be manifested or elicited when exposed to a relevant stimulus, it seems plausible that we can nevertheless legitimately speak of the possibility that O ceases to exist. And, plausibly, the same motivations favoring CP equally favor the principle that when a possibility is not actual from [t*, t) but is actual at t (such that t* < t), something already actual causally brings about the possibility’s actuality at t. This principle will equally facilitate the inference to EIT.

Third, it seems plausible that we can infer EIT employing the notion of ‘Cambridge change’, which (as I use it) refers to a change in the predicates satisfied by S without any corresponding gain or loss of S’s properties (whether intrinsic or extrinsic properties). In particular, we can focus on the predicates ‘is such that O exists’ and ‘is such that O doesn’t exist’. We can then consider changes to the extensions of such predicates. O ceases to exist just in case the potential for the first predicate’s extension to be empty becomes actual (else: the potential for the second predicate’s extension to be everything becomes actual). Since—per CP—every potential that is actualized requires something already actual to cause this actualization, and since the only plausible way to cause the relevant change to the predicate(s) is to cause O to cease to exist (i.e., to positively destroy O), we can infer the truth of EIT just as before.

I conclude, then, that the argument for EIT under consideration survives the objection at hand.

Objection. While absences cannot serve as an ‘already actual causal actualizer’, perhaps the event of God’s withdrawing his sustaining activity can serve this role.

Reply. This is a valuable objection. I have at least three replies. First, this requires that there actually are such things as events—a controversial assumption which significantly weakens the Aristotelian proof. Yes, there are objects, and such objects have, gain, and lose various properties. But are there such things as events corresponding to such havings, gains, and losses, over and above the objects and properties? It’s not obvious. (Also, even if we countenance some events within our ontology, do we seriously want to countenance withdrawals therein, too? Might as well believe in shadows, cold fronts, laps, and other occult entities belched from the land of voodoo metaphysics.) Second, the assumption seems to be rather un-Aristotelian. At least according to standard Aristotelian stories, it is substances, not events, that are the causal relata. This is significant given the underlying Aristotelian metaphysical framework within which the Aristotelian proof and CP is cast (and within which various of Feser’s replies to objections are articulated).

Third, it’s not clear what this event consists in. Is it timeless, given that (under classical theism) all of God’s causal activities are timeless? Could there be a timeless event (i.e., happening or occurrence)? Setting this aside, there is a deeper worry. For the event in question is distinct from God.[5] But under classical theism, anything distinct from God is created and sustained by God (Rogers (1996, p. 167), Bergmann and Brower (2006, p. 361), Grant (2019, ch. 1), and Kerr (2019, pp. 15 and 44)). But surely whatever x that God creates and sustains is such that God is free to withdraw his sustaining activity of x. So, God could withdraw his sustaining power from the relevant event. But then there would be the event of God’s withdrawing this sustaining power from the event of God’s withdrawing his sustaining power of O. But then we could run the exact same reasoning ad infinitum. So, there could be an infinitely descending chain of divine-withdrawal events. But that seems absurd.

I conclude that the objection at hand—while valuable and suggestive—fails to avoid my argument for EIT.

6.3 Divine Temporality

Yet another motivation for EIT derives from divine temporality. For if God is temporal, then there is at least one temporal concrete object that persists in existence without sustenance or conservation from without. I don’t have space here for a full defense or justification of EIT. I simply note that the various arguments in favor of divine temporality are ipso facto arguments for EIT (cf. Mullins 2016).[6] I will, though, sketch an argument for divine temporality. I’ve developed this argument at length and responded to objections elsewhere.

I’ll just dive right into the premises:

  1. There is change.

Justification: it’s manifest and evident to the senses. Many in your audience will get that reference.

  • If there is change, then some things go from being true to being false (or vice versa).

Why? Well, just think about it. If there is change—e.g., water’s going from a liquid state to a solid state—then <that the water is liquid> goes from being true to being false. And <that the water is solid> goes from being false to being true.

  • If some things go from being true to being false, then God goes from knowing them to not knowing them.

This follows from the factivity of knowledge. Knowledge is factive—in other words, one can only know that p is true if p is, in fact, true. It cannot be the case that one knows p but p is false. E.g., It’s true that 1+1 is 2, and so it makes no sense to say that I know 1+1 isn’t 2.

This entails that if p genuinely goes from being true to being false, then it can’t remain the case that one knows p. For if it remained the case that one knows p, then since knowing p presupposes p’s truth, it would follow that p itself remained true. But ex hypothesi, p went from being true to being false and hence didn’t remain true. Thus, if some things go from being true to being false, then God goes from knowing them to not knowing them.

(As an aside: In the case at hand, God would go from knowing p to knowing the negation of p, because he’s omniscient. I’m not saying that there are truths God doesn’t know.)

  • If God goes from knowing them to not knowing them, then there is succession in God’s life [and, moreover, God has potential that becomes actual—viz. potential to acquire knowledge]

This is just what succession means—to go from one thing to another. It’s also just what potential means—God has the potential, here, to acquire knowledge.

  • If there is succession in God’s life [and, moreover, if God has potential that becomes actual], then God is temporal.

This is obviously true.

  • If God is temporal, EIT is true.

Once again, this is obviously true. (Moreover, I have already explained why it’s true earlier in this post.)

  • Therefore, EIT is true.

6.4 Bayesian argument

I explained the argument in Section 6.1, but it’s also a distinct argument in its own right. I’ll include it here so that it serves as a standalone argument. Note that this is a brief sketch—much more can be said on its behalf.

We’ll do a Bayesian comparison between CTST and the necessary truth of EIT (hereafter, n-EIT) when it comes to explaining the contrastive fact from Section 6.1. Let that fact be F. Now, recall F: temporal objects, once in existence, uniformly and reliably persist absent sufficiently destructive causal factors as opposed to being annihilated at other points in their lives without being subject to causal destruction. For under the n-EIT, the probability of F is 1. In other words, P(F|n-EIT) = 1. But the probability of F is significantly less than 1 under CTTS. In other words, P(F|CTST) << 1. And hence F strongly confirms n-EIT vis-à-vis CTST.

Why do I say P(F|CTST) << 1? Consider that for an object to persist for about 116 days is for it to persist approximately 10,000,000 seconds. And consider that there are about 1044 Planck times in one second, and hence to persist for about 116 days is to persist for about 1051 Planck times. Now, for each (possible) Plank time in an object’s life, God is free to sustain the object for that Planck time or refrain from so sustaining it, and this is true even if God has sustained the object through previous Planck times. (It’s not as though God suddenly becomes forced or compelled to continue sustaining an object, and it’s not as though some moral obligation is newly imposed on him, or whatever. These are unacceptable by the classical theist’s lights.) This entails that for a given world w in which an object persists for (at least) 116 days and is then causally destroyed by something, there are about 1051 worlds in which that object ceases to exist at some point earlier than w due to God’s withdrawal of sustenance. Moreover, under classical theism, God remains utterly invariant across all worlds, and hence there is no difference or variance on God’s end to which one could point in any such world to explain why it came about as opposed to some other world. Thus, conditional on CTST, the expectation that w obtains—that is, that the single world in which the object persists for 116 days (upon which it is finally causally destroyed) obtains rather than any of the other 1051 worlds—is very small. And hence the data of F, as applied to this 116-day-lived object, is very unlikely conditional on CTST, whereas its probability is 1 conditional on EIT. And hence EIT garners very strong evidential confirmation from F vis-à-vis CTST. Much more can be said on this line of reasoning, but that suffices for a brief sketch.

6.5 Moorean argument

The Moorean argument for EIT essentially begins with the observation that the principal ‘rival’ to EIT is classical theism. But — so the argument goes — there are many, many reasons to think classical theism is false. Hence, since the main rival to EIT is false, EIT follows. (That’s a super-duper rough-and-ready sketch.)

If you want to explore a small sub-set of the abovementioned reasons, you can see, inter alia:

  • My video, “Arguments Against Classical Theism | Part 1/3” and the forthcoming Parts 2 and 3
  • The two problems articulated at the end of my recently-published IJPR article, “The fruitful death of modal collapse arguments”. (Note: I develop both of these problems in much more detail in another paper currently under review.)
  • My article on the aloneness argument

There are more arguments besides, but what I’ve sketched suffices for now. Let’s now consider arguments against EIT.


[1] Pruss and Rasmussen (2018), for instance, offer an argument along these lines for one of their principles.

[2] Another way to see this is to imagine that while all the oxygen in your surrounding environment is removed, there is nevertheless some continual source of oxygen within you. (Perhaps your lungs acquire a new ability to generate oxygen anew.) In this scenario, you simply wouldn’t die despite the utter absence of oxygen in your environment. This shows that it is the status of factors within you upon which you are dependent in this case, not factors outside of (external to, extrinsic to) you.

[3] Note that when I speak of fundamental kinds of things in the context of theoretical virtue comparison, I simply mean kinds of things that are not analyzable in terms of or reducible to other kinds of things. This is a separate issue from one such fundamental kind of thing (e.g., temporal things) standing in a causal dependence relation to another such fundamental kind of thing (e.g., a timeless thing or things).

[4] Positive reality =def anything that is not an absence.

[5] God, after all, is not identical to a withdrawal!

[6] Another argument for EIT—one I shan’t pursue beyond this footnote—is that the denial that temporal concrete objects (or some subset thereof) persist inertially will ultimately have to admit a timeless sustaining cause of temporal reality or some portion thereof. But—so the argument continues—a sustaining causal relation cannot obtain between a timeless, immutable being and the temporal world. This latter claim has been defended by several able philosophers—see, for instance, Fales (1997), Rundle (2004, ch. 4), Mullins (2016, ch. 5), and Craig (2001, ch. 3).


7 Arguments against EIT

My goal in this section is to survey and assess the principal criticisms of EIT in the literature (and elsewhere), beginning with alleged counterexamples thereof.

7.1 Alleged Counterexamples

It is necessary to alleviate some immediate, potential counterexamples to EIT before considering more substantive criticisms of EIT found in the literature. Consider first a violinist’s musical sounds. If the violinist stops playing, presumably the musical sounds stop existing as well. Isn’t this a counterexample to EIT’s claim that things don’t have sustaining or conserving causes?

Note first that EIT only claims that temporal concrete objects or some subset thereof persist in the absence of both sustenance and destruction. Thus, even if this were a case of a concrete object being sustained in existence (or requiring sustenance in order to exist), it wouldn’t serve as a counterexample to EIT.

Crucially, though, this is not even a case of a concrete object being sustained in existence. For this case is a process of playing music and hence is not a concrete object. Moreover, the removal of the violinist does not cause the sound to cease, since the mechanical sound waves continue to exist as compressions and rarefactions of air molecules that can be heard after the violinist ceases to exist.

At best, then, this is a case of a continually perpetuating, rather than sustaining, cause—a cause such that its instantaneous removal (i) does not necessitate the instantaneous removal of the effect but (ii) contributes to the removal of the effect after a (sometimes short) duration of time.

Another purported counterexample to EIT is the laboratory-synthesized heavy elements that exist for a very short amount of time. Isn’t it false that they persist in existence in the absence of sustenance and destruction? Don’t they instead simply spontaneously cease to exist?

The first response to this proposal is that their cessation of existence is not the robust kind required by CTST, since the heavy elements do not, along with their parts, instantaneously annihilate. Second, recall that EIT only quantifies over temporal concrete objects or some subset thereof. Hence, even if spontaneously decaying particles do not inertially persist, this does nothing to refute EIT. Third, the facts concerning the cessation of existence of such elements are fully consistent with EIT. This is because such quick cessations of existence are not utterly inexplicable (for that would destroy the possibility of their scientific study!). Rather, they are caused to cease to exist, either by their intrinsic nature/character or by environmental conditions unconducive to their perpetuation. But that means such elements are not counterexamples to EIT, since they could easily be such that they persist of their own accord for that short span of time and are interrupted in such persistence by sufficiently destructive internal or external causal factors.

See also Oderberg (2014, pp. 350-353) on the compatibility of radioactive decay and a tendency to persist in existence. I’ll include the important paragraphs here:

Although EIT is immune to immediate, alleged counterexamples, it remains to be seen whether or not the principal criticisms of EIT in the literature succeed. It is to these that I turn next.

7.2 Red Chairs

I’ll consider Oppy’s reflection on existential inertia and the criticisms thereof from McNabb and DeVito (2020). To begin, it’s worth quoting Oppy at length:

Yesterday, throughout the entire day, there was a red chair in my room. Pick some time t

around noon yesterday. At t, the chair existed, and the chair was red. Moreover, at t, the chair had the potential to exist, and to be red, at t + ε, where ε is some relatively short time interval (say, a millionth of second). Do we need to postulate the existence of some distinct thing that exists through (t, t + ε) that actualizes at t+ε the potential that the chair had at t to both exist and be red at t+ε? I do not think so. Given that, at t, the chair has the potential to exist and to be red at t + ε, all that is required for the realization of this potential is that nothing intervenes to bring it about, either that the chair does not exist, or that the chair is not red, at t + ε. Potentials to remain unchanged do not require distinct actualizers; all they require is the absence of any preventers of the actualization of those potentials. In particular, things that have the potential to go on existing go on existing unless there are preventers—internal or external—that cause those things to cease to exist. (Forthcoming, p. 4)

The idea here is that states or conditions of unchangingness don’t require distinct actualizers to ‘keep’ or ‘maintain’ them in such states or conditions; only deviations from such states or conditions require a distinct actualizing cause. Stated differently, actuality behaves inertially insofar as states simply continue in their state of actuality unless changed to some different state of actuality (becoming green as opposed to red, say, or being destroyed). McNabb and DeVito, however, are not impressed by Oppy’s point. They respond:

[I]n order for the chair to remain red at t + ε, the chemical microstructure of the chair will have to continue being in a certain way. If the microchemical structure of the chair were different, the chair would no longer be red. So, it isn’t the case that the red chair can remain red at t + ε as long as nothing intervenes. Rather… in order to retain ‘the redness,’ something outside of ‘the redness’ needs to be in place. (2020, p. 7)

But this response to Oppy is inadequate on multiple fronts. In order to simplify things, let’s simply focus on the existence of the chair. There are three reasons for this simplification. First, Feser’s Aristotelian proof (and Persistence Arguments at large) doesn’t primarily concern inherent properties of things (like redness) or states of affairs (like the chairs’ exemplifying redness at some time) but instead concerns the very being or existence of substances themselves. Second, the response to Oppy raised by McNabb and DeVito carries over to the chair’s existence itself (viz. the chair’s existence being in some sense dependent on the microstructure of the chair’s wood, etc.). Third, Feser himself argues that the existence of (e.g.) water is ‘actualized’ by its atomic constituents, which in turn are actualized by the subatomic constituents, and so on (2017, pp. 26-27). So, my criticisms don’t merely apply to McNabb and DeVito but also to Feser.

The first problem with the response is that the chemical microstructure of the chair is a material cause of the chair, not an efficient cause that brings the chair into and efficiently sustains it in being. Hence, the example doesn’t constitute a counterexample to Oppy’s claim, since Oppy is fundamentally and ultimately concerned with existential inertia, i.e., things’ continuing in existence not without a material cause but instead without a continuously sustaining efficient cause (or ground, etc.).

The second problem with their response is that appealing to microstructure as a ‘cause’ upon which a chair depends simply undermines their ultimate intended inference to an unsustained sustainer of the existence of everything apart from itself. For the microstructure of the chair is a component of the chair, and only ‘actualizes’ the chair in the sense that something ‘depends’ (in some sense) on its components. But upon tracing this causal chain of ‘dependence’ down to a first member, all we’re entitled to infer about such a member is that it is an uncomposed component, not an unactualized actualizer of the very being or existence of the things in question. McNabb and DeVito have not pointed to something outside the composite object as an actualizer of its very being; instead, they’ve merely sent us off on a regress of more fundamental component parts of something on which that thing ‘depends’ (in the sense of whole-to-part dependence). Their response, then, ends up doing more harm than good.[1]

The third problem with their response is that (i) the microstructural elements of the chair are parts of the chair, but (ii) the parts of something plausibly cannot efficiently cause (i.e., actualize the existence of) their whole. Here are, briefly, three reasons why this is so. First, if we allow that parts of something can efficiently causally sustain their whole, then sustaining-cause-arguments could only get us to an unsustained, uncomposed component that sustains wholes—not a radically transcendent God who is decidedly not a part of creation. Second, parts of wholes seem in some sense less fundamental than the wholes of which they are parts; plausibly, they’re only intelligible with reference to the substances they compose. Their existence qua the things they are, then, presupposes the (ontologically) prior existence of the substance and hence cannot causally explain its existence. It seems that your arm cannot be the efficient cause of you, since its being your arm in the first place presupposes your existence as a substance. Third, it’s a component of Thomistic-Aristotelianism that parts of substances exist merely virtually (and hence only in potency); but per the Aristotelian causal principle, only things existing in actuality can causally actualize the existence of something else. (Cf. Feser (2014, p. 197) for the point about parts and virtual containment.) And while our final two points in this paragraph depend on Aristotelian views of substances, there is still value in bringing to light incompatibility between McNabb and DeVito’s rejoinder (on the one hand) and such Aristotelian views (on the other hand). For not only does Feser accept such views, but they are also rationally defensible—and to the extent that one finds them plausible, one has pro tanto reason to reject McNabb and DeVito’s rejoinder.

Overall, then, McNabb and DeVito have not succeeded in providing a counterexample to Oppy’s point. The final criticism McNabb and DeVito level towards Oppy’s point is:

[T]here is still a fundamental question that Oppy would have to engage. What is it that actualizes the potential of chair to possess such a nature that it can continue being red from T1 to T2 without any external interference? Oppy will be hard pressed to find an explanation that doesn’t bottom out in an uncaused caused. (2020, p. 8)

I have two points to make in response. First, while it’s true that Oppy would have to engage this fundamental question, it is also true that the onus of justification is on Feser and defenders of sustaining-cause-arguments to demonstrate that any temporal thing concurrently depends on an efficient sustaining cause at any moment which it exists. They are, after all, in the business of giving a positive argument for God’s existence. So, even if Oppy does not answer the question concerning what explains such beings’ existence at a given (non-first) moment of their existence, that would by no means vindicate such arguments.

Second, there exist a variety of ways that Oppy could engage this very question. This was the entire project of the earlier sections on metaphysical accounts of EIT. Oppy could make use of any of those metaphysical accounts of EIT to explain persistence and answer the question McNabb and DeVito here pose. Nothing in what McNabb and DeVito say here (and, moreover, nothing in any of Feser’s proofs) rules out these alternative explanations of the persistence (and, hence, existence-at-any-given-non-first-moment) of temporal concrete objects.

7.3 Hsiao and Sanders

Hsiao and Sanders (Forthcoming) provide four (or so) lines of argument against EIT. I will address each in turn.

7.3.1 Counterexamples to EIT

In speaking of beings for which a concurrent cause is needed for their moment-by-moment existence, Hsiao and Sanders write:

Upon reflection, we see that the world is filled with such beings. For example, an ice sculpture is a dependent being in the sense that its continued existence depends on external conditions like temperature level. If the temperature were to suddenly rise to 10,000 degrees, then the sculpture would immediately cease to exist. Cold temperature functions as a sustaining cause that keeps ice sculptures in existence. (Forthcoming)

I make several points in response. First, note that temperature is (something like) mean molecular kinetic energy. In other words, it’s the average energy content associated with the movement or jiggling of molecules. But what molecules are we talking about? It won’t matter for the continued existence of the sculpture if the mean molecular kinetic energy to which we’re referring is that of molecules wholly outside the sculpture. For if every single molecule within the sculpture (including the one’s on its exact boundaries) retained a mean molecular kinetic energy with a measured value below 32 degrees Fahrenheit, then it will persist in existence. Thus, the kinetic energy of the molecules wholly outside the sculpture are not sustaining causes here. Thus, the kinetic energy of the molecules to which Hsiao and Sanders label as a sustaining cause must be that of the molecules within the sculpture. But in that case, the molecules and activities to which we’ve referred are merely proper parts of the sculpture, not efficient sustaining causes of its existence. They constitute the sculpture; they don’t continuously efficiently cause its very being.[2]

Second, the appeal to parts will not target EIT as I’ve articulated it. For my EIT only debars inertially persistent objects from dependence on non-parts of them. Thus, my EIT already takes into account any dependence something might have on its parts, even if the relevant parts in question are efficient sustaining causes.

Third, as with the appeal to microstructural elements of the chair as purported actualizers of its existence, an appeal to microstructural elements of the sculpture (e.g., its molecules, their bonds, the internal operations of laws, etc.) as a ‘cause’ upon which it depends plausibly undermines any intended inference to the classical theistic God. For once again, we’ve been sent on a regress that need only terminate in a not-further-composed component of something. But the God of classical theism, of course, isn’t a part of anything.

But perhaps their other examples of sustenance will avoid the abovementioned objections. They write:

Likewise, you and I are dependent beings in the sense that our continued existence depends upon a myriad of physical conditions, including temperature, pressure, entropy, the strong nuclear forces holding our atoms together, and the laws of physics. Additionally, we depend on biological processes such as blood flow, breathing, cell replication, energy production, etc. in order to exist. If these conditions or processes were removed or significantly altered, we would cease to exist at that moment. (Forthcoming)

But again, Hsiao and Sanders have merely pointed to component parts (or component processes) of things, and hence such examples do not avoid the variety of problems articulated above. Simply re-run the points raised in response to the sculpture example here. In particular, upon distinguishing between the intrinsic versus extrinsic temperature, pressure, entropy, etc. as applied to humans, we submit that we can once again see that these are actually mere component parts and processes of humans (just as we did in the case of the sculpture’s temperature). Hsiao and Sanders also write that “[i]maginary beings like those in your dreams would cease to exist the moment you cease thinking about them” (Forthcoming). But dreams and their contents are processes, not concrete objects in their own right, and hence this will not target EIT as we articulate it.

Yet another point worth emphasizing here is that EIT only quantifies over temporal concrete objects or some subset thereof. Thus, even if sculptures, people, and other macroscopic physical objects require outside sustaining causes of their existence, this would in no way invalidate EIT. For this says nothing about (say) the foundational layer of reality (whether it be naturalist-friendly (e.g., one or more fields, or the universal wavefunction, or a collection of fundamental particles, or a collection of mereological simples, or whatever) or theist-friendly (e.g., the neo-classical or panentheistic God)).

7.3.2 First Metaphysical Argument

From what I can make out, three distinct metaphysical arguments against EIT are proffered in Hsiao and Sanders (Forthcoming), to which I turn now. Their first metaphysical argument runs as follows:

Another reason to think dependent beings require a continual explanation distinct from themselves is that existence is not built into their nature. For example, there is nothing in the nature of what it is to be a dinosaur that requires it to exist. It is possible for dinosaurs to exist or not to exist. Their existence is neither impossible nor required. Because of this, the “default state” of a dinosaur’s nature is non-existence. (Forthcoming)

I think it would be more accurate, however, to say that such objects simply have no default state. And as I’ve explained in my Sophia paper and my response to Feser, this fact can plausibly motivate inertial persistence: just as a cup does not occupy location L by default, it doesn’t follow that the cup needs some concurrent causal ‘keeper’ of the cup to maintain it in location L. (Indeed, we have good reason to think this is false—compare the astronauts on the ISS.) Indeed, the cup has no default location whatsoever—neither L nor non-L. But what this means is that, plausibly, once the cup is placed in a given location, then—provided that there are no ‘net causal factors’ operative—the cup will simply retain its state of being in L. And this without needing something to ‘hold’ or ‘keep’ it there. The first metaphysical argument, then, doesn’t count against EIT.

A second thing to note is that EIT is actually compatible with non-existence being the default state for everything contingent and the further claim that anything for which non-existence is the default state require concurrent sustenance in order to exist moment-by-moment. For EIT, again, quantifies only over temporal concrete objects or some subset thereof, and hence EIT might be false for contingent temporal concrete objects while true for some necessary temporal concrete object(s).

7.3.3 Second Metaphysical Argument

The second metaphysical argument they provide runs: “Second, if dinosaurs did have existence built into their nature in the same way that being a reptile is built into their nature, then they could never cease to exist and would have to exist forever” (Hsiao and Sanders, Forthcoming). Hsiao and Sanders do not define what something’s being ‘built into’ a nature means. We will understand F’s being ‘built into’ the nature of S to mean that F is either an essential property of S or else an essential part of S.

But this is a non-sequitur. Suppose we think x is F essentially if and only if necessarily, if x exists, then x is F. This is a typical formulation of what is at least logically equivalent to (though probably not an informative analysis of) essentiality. In this case, though, existence is essential to everything, since it’s trivial that, necessarily, if x exists, then x exists.

In this understanding, then, existence is ‘built into’ everything’s nature in the sense that existence is at least predicated essentially of everything. (And, if we think existence is a property, then existence is an essential property of everything within this understanding). But this doesn’t mean that things automatically or necessarily exist. It doesn’t mean that x exists because existence is essential to x. An x would only be necessarily existent, under this view, if necessary existence were an essential property of x (or if it were essentially predicated (truthfully) of x). But that isn’t the proposal in question. And merely from the fact that existence is essential to x, it doesn’t follow that x cannot cease to exist. All one needs to do is cause any one of x’s essential properties to go out of existence in order to cause x to go out of existence.

Finally, one might simply deny the reifying of existence. Is there such a thing as something’s existence or being? It’s not obvious, and Hsiao and Sanders need to justify this assumption if they want inertialists to accept their argument.

7.3.4 Third Metaphysical Argument

The third metaphysical argument in Hsiao and Sanders (Forthcoming), as I formalize it, runs:

  1. If EIT is true for some contingent x, then existence becomes part of x’s nature upon x’s coming to be.
  2. If existence becomes part of x’s nature, then existence becomes either an essential or non-essential feature of x.
  3. If existence becomes an essential feature of x, then it is impossible for x to go out of existence.
  4. But it is possible for x to go out of existence (at least if caused to do so, as EIT says).
  5. If existence becomes a non-essential feature of x, then the existence of x is both prior to and posterior to itself.
  6. But it is not possible for the existence of x to be both prior to and posterior to itself.
  7. Therefore, EIT is not true for any contingent x. [1-6]

On behalf of premise (1), they write that the “defender of EI [Existential Inertia] would have to claim that existence is not built into its nature but becomes a part of its nature… after it is caused to exist. This would allow a being to explain its own continual existence without something distinct from itself” (Ibid). I make four points in response this.

First, Hsiao and Sanders provide no justification as to why the mere denial that existence is built into the nature of some contingent, inertially persistent x would entail that existence must become part of its nature. Thus, their argument rests on a claim for which no justification is proffered.

Second, EIT does not maintain that a being ‘explains its own continual existence without something distinct from itself’. As we saw earlier, there are many different metaphysical accounts that adduce explanatory facts distinct from O at t to explain O at t.

Third, it is false that inertial persistence would somehow require existence to become built into something’s nature. By analogy, consider inertial location. In particular, consider that spatial location is almost always (perhaps always) numerically distinct from the essence of the thing occupying said location. A cup, for instance, is distinct from a particular spatial location, since the cup can exist without being located at such a position. (This shows, moreover, that such a spatial location is not even essential to the cup, and nor does it ever ‘become built into the nature’ of the cup.) But the mere fact that the cup’s particular spatial location is distinct from its essence (and not even included in its essence as one of its essential properties, and not even ‘built into it’) doesn’t entail that the cup’s being in this particular spatial location (at any moment at which it is in such a location) requires some kind of continuously concurrent external causal ‘keeper’ or ‘sustainer’ of the cup’s being in said location. Indeed, the opposite seems to be the case: we have good reason to think that no such continuously concurrent sustenance is required in order for the cup to simply retain its spatial position.

What this shows is that inertial maintenance of location L does not require that ‘being in L’ somehow ‘become built into the essence of the cup’. But by exactly parallel reasoning to what Hsiao and Sanders provide (in the case of existence), it seems that such inertial maintenance would require this. At the very least, nothing Hsiao and Sanders say here gives us any reason to think that existence is relevantly dissimilar to being in location L. And that’s precisely what they would need to show for their argument to work.

Fourth, the claim under consideration fails to consider the various metaphysical accounts of EIT. Tendency-disposition accounts, for instance, neither require nor entail that existence becomes part of something’s nature. They only need a tendency or disposition in conjunction with certain manifestation conditions. Similarly, no-change accounts do not require that existence becomes part of something’s nature. They only need an understanding of existence as a state/condition of unchangingness in conjunction with a claim about the nature of such states/conditions. Propositional necessity accounts say or entail nothing about existence’s becoming built into the nature of things; all they require is a primitive commitment to the metaphysical necessity of EIT. Objectual necessity accounts, likewise, have nothing to do with existence’s becoming ‘built into’ natures of things. Instead, necessary existence is (in some sense) built into one or more foundational concrete objects, and these objects go on to continuously explain (whether by grounding, realization, causation, or constitution) non-foundational temporal things. Transtemporal accounts similarly have nothing to do with existence’s becoming built into things’ natures. Rather, they say that transtemporal explanatory relations, in conjunction with the fact that there are no sufficiently destructive factors operative, adequately explain persistence. And so on.

I conclude, then, that premise (1) is both unjustified and false. But I wish to continue my appraisal of the argument, since insights can be gleaned from what Hsiao and Sanders say on behalf of its other premises. We need not consider premises (2) and (6), since they seem relatively innocuous. And we can likewise grant premise (4). That leaves premises (3) and (5) to consider.

On behalf of premise (3), Hsiao and Sanders write that if some feature is essential to S, then S “could not possibly lose that feature. For example, being a living organism is essential to a human. If it lost that feature, it would stop being a human. So if existence is essential, then it would be impossible for this being to go out of existence” (Forthcoming).

But this confuses de re necessity with de dicto necessity. If F is essential to S, this only means (inter alia) that S cannot exist without being F. Importantly, it does not mean that necessarily, S is F. The former kind of necessity is de re necessity. The latter kind of necessity is de dicto necessity.

Now suppose, along with Hsiao and Sanders, that existence is essential to S. All that follows from this is that S couldn’t exist without S existing. And this is true. What does not follow is that ‘necessarily, S exists’ or ‘it is impossible for S to go out of existence’. These are de dicto necessities that are not entailed by the de re necessity of existence’s being essentially had by S. It is thus a non-sequitur to infer the impossibility of something’s going out of existence from the mere fact that existence is essential to it. All we can infer from existence’s being essential to S is that necessarily, if S exists, then S exists. Nothing follows, however, about the impossibility of S’s going out of existence.

Here’s another way to think about it. If F is essential to S, then all we can infer is that it is impossible for S to lose F while remaining in existence. Thus, it is impossible for Sophia, a human, to lose the feature ‘being a living organism’ while remaining in existence. But this doesn’t entail that it’s impossible for Sophia ‘lose’ this feature (in the sense that it is no longer true that Sophia has this feature), since Sophia can ‘lose’ the feature (in the aforementioned sense) by going out of existence. In such a case, it is no longer true that ‘Sophia is a living organism’, whereas it was true earlier. What this shows us is that all we can infer from S’s essentially being F is that it is impossible for S to lose F while remaining in existence. But applying this to existence’s being essential to S, all we can infer here is that it is impossible for S to lose existence while remaining in existence. It does not follow that it is impossible for S to lose existence (in the sense of ‘S exists’ going from true to false), just as it does not follow that it is impossible for Sophia to lose the feature ‘being a living organism’ (since, again, Sophia can ‘lose’ it in the sense of going out of existence and hence losing all her properties). I conclude, then, that premise (3) is false or, at least, unjustified.

On behalf of premise (5), Hsiao and Sanders point out that S’s having a non-essential feature F presupposes that S’s nature or essence already exist, in which case F cannot be S’s existence. For then S’s existence would be both prior and posterior to itself. As they write, one “cannot attach a case (non-essential feature) to a phone unless the phone already exists” (Ibid). And from the denial that existence is either an essential or non-essential feature, they conclude that “[e]xistence must be an activity in the same way that change is an activity” (Ibid).

         We make two points in response. First, it seems that the inference to existence’s being an activity is a non-sequitur. Merely from the fact that existence is neither an essential nor non-essential feature of S, it doesn’t follow that existence is an activity of S. (For instance: perhaps there is simply no such thing as existence—perhaps, in other words, there is no positive ontological item that is existence.) Second, the exact same reasoning Hsiao and Sanders employ against existence’s being a property or a non-essential feature of S equally applies to their claim that existence is an activity of S. For in order for S to act in a given way—to perform some activity like walking, say—S must already exist. Something must first exist in order to engage in various activities (i.e., in order to act, in order to have activity). Non-existent things, after all, surely don’t act or engage in activities. Thus, existence cannot be an activity of S. This reasoning, by our lights, is exactly parallel to the reasoning Hsiao and Sanders use to dismiss existence’s being a property and a non-essential feature (respectively).

Finally, even if—contrary to what I’ve argued—their argument succeeds, the conclusion, (7), is perfectly compatible with EIT, since EIT quantifies only over temporal concrete objects or some subset thereof, and the subset could very well include only non-contingent (i.e., necessary) things.

Overall, this third metaphysical argument is fraught with problems. Let’s move on.

7.4 Proportionate Causality

Feser’s central argument against EIT derives from the Principle of Proportionate Causality (PPC), according to which a total cause cannot give to an effect what it does not have to give in the first place. More precisely, the PPC holds that whatever exists in an effect E must exist in the total cause of E in some manner (formally, virtually, or eminently). With this principle in hand, Feser argues:

  1. A cause cannot give what it does not have to give.
  2. A material substance is a composite of prime matter and substantial form.
  3. Something has existential inertia if and only if it has of itself a tendency to persist in existence once it exists.
  4. But prime matter by itself and apart from substantial form is pure potency, and thus has of itself no tendency to persist in existence.
  5. And substantial form by itself and apart from prime matter is a mere abstraction, and thus of itself also has no tendency to persist in existence.
  6. So neither prime matter as the material cause of a material substance, nor substantial form as its formal cause, can impart to the material substance they compose a tendency to persist in existence.
  7. But there are no other internal principles from which such a substance might derive such a tendency.
  8. So no material substance has a tendency of itself to persist in existence once it exists.
  9. So no material substance has existential inertia. (2011, p. 258)

There are at least five problems with this argument.

First, it presupposes a controversial metaphysical account of the nature of (material) substances. Indeed, the argument—if successful (something we wouldn’t grant)—establishes merely that “whether [hylemorphism] is correct depends in part on whether things have existential inertia in the first place,” for if they have existential inertia, then (per Feser’s argument) hylemorphism is false (Audi 2019, p. 7). And note that in this dialectical context, the onus is not on me to provide positive reasons to think that hylemorphism is false. Rather, the present dialectical context is one wherein Feser is attempting to provide positive reason to think that EIT is false. Thus, I need only provide undercutting defeaters of the argument. And thus it is a perfectly acceptable dialectical move to pinpoint an assumption of the argument that is left unjustified in the dialectical context at hand. (If Feser wants to provide the inertialist with some reason to change their mind, then it’s no use presenting them arguments with premises that they simply reject or find unmotivated.)

Second, premise (3) is false. Premise (3) says: “Something has existential inertia if and only if it has of itself [i.e., intrinsically or internally[3]] a tendency to persist in existence once it exists.” The left-to-right side of the biconditional here says that an existential inertial tendency intrinsic to S is a necessary condition for S’s persisting inertially. But this is false. It is not a necessary condition for S’s inertially persisting that there is some ‘inertial persistence tendency’ intrinsic to S. For there are whole swathes of explanations of an object’s persistence—that is, metaphysical accounts of EIT—that do not posit such a tendency but instead adduce facts extrinsic to S at m to explain S at m.

Consider one of my no-change accounts. For S to fail to exist at m despite existing from [m*, m), m* < m, is for some change to occur.[4] But a change occurs only if some factor causally induces said change. Hence, if no factor causally induces a change, then the change won’t occur. Thus, if no factor causally induces S to fail to exist at m despite existing from [m*, m), then S exists at m. Once we add that nothing came along to causally induce this—that is, once we add that nothing came along to destroy S from m* to m—it simply follows that S exists at m. Here, we seem to have a perfectly respectable, perfectly legitimate explanation of S’s existence at m, and this explanation adduces facts outside of or extrinsic to S at m. And the same is true of the other no-change accounts I surveyed.

Or consider inertialist-friendly explanations based on laws of nature or transtemporal explanatory relations, each of which adduces facts extrinsic to S at m to explain S at m. An explanatory appeal to the de dicto necessity of the inertial thesis (à la propositional necessity accounts) likewise cites facts extrinsic to S at m. And so on. Premise (3), then, is false.

To drive home this point further, consider the following Explanandum:

Explanandum: S’s existence at m.

Now consider the following Transtemporal Explanans and No-change Explanans:

Transtemporal Explanans: (i) There is an absence of sufficiently causally destructive factors operative on S from m-1 to m (where m-1 is the moment immediately prior to m), and (ii) the state and/or existence of temporal concrete objects (or, at least, those within EIT’s quantificational domain) at a given moment at which they exist causally produces their existence at the next moment provided that no sufficiently causally destructive factors are operative.[5]

No-change Explanans: (i) S existed immediately before m (i.e., at m-1); (ii) if S existed immediately before m but fails to exist at m, then S’s cessation is (or involves, or entails) a change; (iii) nothing causally induces S’s cessation at m-1 or m—that is, nothing destroyed S from the immediately prior moment, m-1, through m; and (iv) a change occurs only if some factor causally induces said change.

Now, neither Transtemporal Explanans nor No-change Explanans cite facts that explain Explanandum (S’s existence at m) that are intrinsic to Explanandum (i.e., that are intrinsic to S at m or S’s existence at m). And yet both scenarios are ones in which S inertially persists. And hence something can inertially persist (i.e., persist in the absence of both destruction and continuously and concurrently operative sustenance from without) without this inertial persistence deriving from (resulting from, being explained by) some intrinsic principle or tendency. Instead, for each non-first moment m of S’s life, the explanation of S at m is extrinsic to S at m but not a concurrently operative extrinsic sustaining efficient cause.[6] Thus, premise (3) is false.

One might object that neither Transtemporal Explanans nor No-change Explanans sufficiently explains Explanandum. I have two responses. First, the present dialectical context is one wherein Feser is proffering a positive argument that requires the explanatory insufficient of Transtemporal Explanans and No-change Explanans. All I need to do, then, is point out that nothing in the argument or what’s said on its behalf gives those who do think Transtemporal Explanans and No-change Explanans sufficiently explain Explanandum sufficient reason to abandon their view. Second, the explanatory facts adduced in each explanans are sufficient. If by ‘sufficient explanation’ we mean an explanation citing facts that remove mystery as to why the explanandum obtains, then I confess that—by my lights—each explanans removes mystery as to why and how S exists at m. The explanandum was simply derived from the explanatory facts cited, and they illuminate precisely why S exists at m.

To be sure, there might be the further question of why some of those explanatory facts themselves obtain. For instance, there might be the question as to why reality is so constituted that the successive stages in an object’s life are related by causal relations. But this is a separate question from why S exists at m. And, plausibly, it won’t be all that difficult to provide plausible stories for the former question. (Indeed, it’s not clear why explaining it would be any more difficult than explaining why reality is so constituted so as to have any causal relations at all (ever), or to have causal relations other than those relating the successive stages of an object’s life, or what have you.) Finally, to quote Beaudoin, “it is not a condition on legitimate explanation that a deeper explanation for every statement in the explanans always be ready to hand, or even that it exist at all” (2007, p. 89).

Here’s my third response to the argument. Consider chemical reactions in which two reactant species are each (individually) necessarily and essentially colorless, but when mixed together produce a vibrant red color. Although each individual thing within the total cause of the vibrant red color is essentially colorless, the combination of the individual things within the total cause nevertheless produces a vibrant red color.

This is not a proposed counterexample to the PPC. Instead, it reveals that there are ways that features can be present in total causes that Feser’s argument neglects. In particular, features can be present within total causes in a way I shall term conditional potencies. O possesses a conditional potency for F provided that O, when conjoined with some other condition or thing O*, gives rise to a system (O-O*) that manifests F. The red was not actually or formally present in the total cause (the two chemical species), but it was nevertheless present in the total cause as a conditional potency of each reactant species. The first reactant species had the conditional potency, when combined with the second species, to produce red; likewise with the second species.

We can now apply this to Feser’s argument. Specifically, merely from the fact that neither prime matter nor substantial form (of themselves) can have a tendency to persist in existence, it does not follow that their composition cannot have a tendency to persist in existence—any more than the fact that neither of the two reactant species can (of themselves) manifest redness entails that their composition cannot manifest redness. In the case of the chemical species, their composition can manifest redness precisely because each component has the conditional potency to manifest—when combined with the other—redness. Similarly, Feser’s argument neglects the fact that a form-matter composition may be able to manifest a tendency to persist in existence because each component has the conditional potency to, when combined with the other component, manifest such a tendency. By illegitimately presupposing that neither form nor matter could have such a conditional potency, Feser’s argument fails.

Now, Feser responds to the above rejoinder. Here’s what he writes:

One problem with this is that, just left at that, it doesn’t really amount to much of an objection. For in the case of the chemical constituents, there are chemical facts we can point to that explain exactly why they will together generate something red. But Schmid does not tell us exactly what it is about prime matter and substantial form that would (or indeed could), when they are combined, generate a tendency to persist in existence. (2021)

In response, this reply simply misses the dialectical context at hand. My objection is an undercutting defeater. I don’t need to positively justify or spell out exactly what it is about prime matter and substantial form that would, when combined, result in such a tendency. I need only point out that nothing Feser says in his argument gives any reason for ruling this out, and that ruling this out is what he would need to do for his argument to succeed. In other words, I need only point out that Feser has not given the hylemorphically-inclined existential inertialist who does think that the form and matter, when combined, result in such a tendency sufficient reason to abandon their position. And so it is irrelevant that I don’t specify exactly what it is about form and matter that, when combined, results in such a tendency.

Imagine we live in 500 B.C. without any knowledge of the underlying chemistry of chemical species A, B, and C. I have gone blind recently, but you haven’t. Before going blind, I interacted a lot with A and B and know that they are colorless liquids. But neither you nor I have combined A and B together to see what results. Today, though, that’s going to change. We combine A and B together to produce C. You can see that C is a vibrant red liquid. I, of course, cannot. You tell me it’s vibrant red. I scoff, responding with the following:

“But that’s not possible. The only things from which C could derive such a feature would be A and B, and neither of them have the feature to grant to C. They are both colorless. And so there is simply nowhere from which C could derive this vibrant color. You must therefore either be lying to me, or playing a trick on me, or your visual apparatus is malfunctioning.”

You respond:

“But your argument there assumes that there is nothing about A and B such that, when combined, they’re able to manifest this further feature not found in either of them individually. Think of it like wine. Wine has a flavor none of its various constituents individually do.”

Imagine I respond with the following:

“That doesn’t really amount to much of an objection. In the case of wine, there are facts we can point to that explain exactly why they will together generate something with a distinctively wine-like flavor. But you do not tell us exactly what it is about A and B that would (or indeed could), when they are combined, generate this vibrant red color.”

This response is confused. It is irrelevant whether you can point to facts about A and B that illuminate why they generate redness when combined. For I was the one offering a positive argument that A and B cannot generate redness. And my argument assumed that there couldn’t be anything about A and B that can combinedly manifest some feature not individually found in either. You then pointed out that my argument needs to assume this in order to succeed, but that absolutely no reason was provided for such an assumption. And you pointed to a case wherein precisely this ‘manifesting of a feature not individually found in any of the constituents’ occurs in order to bolster the in-principle legitimacy of this kind of phenomenon. My argument does nothing to rule this out in the case of A, B, and C, and yet ruling this out is precisely what needs to be done for my argument to work. So, my argument fails. And for the same reason, Feser’s argument fails. (The case of A, B, and C is structurally identical to the dialectic with Feser’s PPC argument.)

Fourth, depending on how we understand ‘principle’, premise seven is arguably question-begging in this dialectical context. (Unfortunately, Feser does not define or explicate the notion—indeed, it seems to be a conceptual primitive in the Aristotelian framework). For whether there is a ‘principle’ of material substances (namely, something that accounts for their (inertial) persistence, however we spell this out (cf. the metaphysical accounts discussed in Section 5)) is precisely what is at issue. It is precisely the question at hand whether form and matter are the sole principles of material substances, since it is precisely the question at hand whether there is an additional principle (or additional fact about reality or temporal concrete objects) which accounts for the persistence of things. Hence, Feser’s argument seems question-begging.

Feser responds to the above rejoinder as follows:

To see the problem with this objection, consider an EIT-rejecting reductive naturalist who argues as follows:

The physical world consists of nothing more than fermions and bosons and the laws that govern them. But there is nothing in the nature of fermions and bosons or the laws that govern them that entails that they have existential inertia. Hence, there is no such feature in the physical world.

Whatever you think of such an argument, would it beg the question? Not if the speaker has independent grounds for being a reductive naturalist. Hence, in response to such a reductive naturalist, a defender of EIT would either have to give some argument against reductive naturalism, or show that the conclusion does not follow from the premises. It would not be enough merely to accuse the speaker of begging the question. But by the same token, my argument does not beg the question if I have independent grounds for being a hylemorphist, which I do. Hence, even if Schmid had other good reasons to reject the argument, accusing step 7 of begging the question is not a good one. (2021)

But this reply is mistaken. Suppose I’m a theist, and suppose I argue as follows:

  1. God exists.
  2. Therefore, God exists.

Now suppose an atheist charges me with question-begging. Following Feser’s rejoinder above, suppose I respond:

“Whatever you think of such an argument, would it beg the question? Not if I have independent grounds for being a theist. Hence, in response to such an argument, an atheist would either have to give some argument against theism, or show that the conclusion does not follow from the premises. It would not be enough merely to accuse me of begging the question.”

This response is confused. And yet it is precisely the response Feser has given to my charge of question-begging. It is irrelevant whether the proponent of the argument has independent reasons for accepting one of the premises. What matters, instead, is whether the argument—its premise(s) and that which is said on their behalf—gives those who don’t already accept it sufficient reason to abandon their position. The job of the detractor of the argument is not to give some argument against the premise, i.e., some argument positively justifying its denial. Just as the onus is not on the atheist to give some positive argument against theism in the above dialectical context, the onus is not on me to give some positive argument against hylemorphism in the context of Feser’s PPC-based argument. Thus, Feser’s response to the charge of question-begging fails.

Here’s my second reply to Feser’s response to the charge of question-begging. Even if Feser has independent reasons for the thesis that material substances are compounds of substantial form and prime matter, what matters is whether Feser has independent reasons for the thesis that material substances are compounds only of substantial form and prime matter. Many of the arguments for hylemorphism—if successful—only establish that material substances are composed of substantial form and prime matter.[7] But it is an entirely separate question whether these are the only principles of material substances.

My fifth and final response to Feser’s PPC-based argument is that even if successful, the argument isn’t an argument against EIT as such. Recall that EIT quantifies over temporal objects or some subset thereof. But in order for Feser’s argument to show that no temporal objects persist inertially, he would have to show that every temporal object is a material object. For only then can he infer the denial of EIT from his conclusion (that ‘no material substance has existential inertia’). But why should we accept this claim? Why should we accept that any temporal object is a material object? It seems eminently plausible that there could, at least in principle, be a non-material temporal thing. (Consider, for instance, the neo-classical theistic or panentheistic God. Or consider neutral monism.)

Feser’s argument against existential inertia, then, fails.

7.5 Form-matter Interdependence

Feser’s second criticism of EIT is form-matter interdependence. “For since in purely material substances matter depends on form and form depends on matter,” writes Feser, “we would have a vicious explanatory circle unless there was something outside the form/matter composite which accounts for its existence” (Feser, 2011, pp. 247-248). What to make of this argument?

I have five criticisms. First, the argument is quite dialectically limited insofar as it rests on a hylemorphic account of temporal, material objects. Audi’s point about hylemorphism and existential inertia is as true here as it was in response to the first argument. (Assuming, that is, that Feser’s argument here succeeds—something we wouldn’t grant.)

Second, even if it is true that, because they depend on one another for their actual existence at each time t at which they are conjoined, form and matter at t require an explanation outside themselves for their actual compositional existence, this is compatible with all or nearly all of the accounts of the metaphysics of existential inertia articulated earlier. Consider transtemporal accounts according to which O-at-t-1 (in conjunction with a few other facts) explains the existence of O-at-t. In this case, we avoid vicious explanatory circularity, since we are not explaining the form of O-at-t by the matter of O-at-t (or vice versa); instead, we are explaining O-at-t by O-at-t-1, which amounts neither to self-causation, nor to self-explanation, nor to vicious explanatory circularity. Or consider propositional necessity accounts: the conjoined-ness of the form and matter at t is explained by something outside of themselves, to wit, the metaphysical necessity of their continued conjoined-ness in the absence of sufficiently destructive factors operator. And so on down the list of metaphysical accounts—in each case, vicious circularity will be avoided, since each of them adduces factors beyond O at t (and beyond O’s form at t and matter at t) to explain O at t (for each non-first time t of O’s life).

Third, vicious causal/explanatory dependence for existence is metaphysically impossible regardless of whether there is something extrinsic that accounts for the viciously intertwined things. If x causes/explains the existence of y, and y causes/explains the existence of x, then x is both prior to y (on account of causing/explaining y’s existence) and posterior to y (on account of being caused/explained by y), which is a contradiction. But contradictions are impossible irrespective of something extrinsic that allegedly grounds their obtaining. This is not an argument against EIT, then; it simply imputes to material objects an impossibility from the get-go.

Fourth, consider again conditional potencies. Arguably, although form and matter may interdepend with respect to the beginning of a substance’s existence (i.e., with respect to the origination of the form-matter composition), it does not follow that they thereby interdepend at every moment at which they exist. This is because—for all Feser has shown—there may very well be a conditional potency within each that accounts for why, when combined with the other, they are able to manifest some further ‘feature’ (namely, a tendency to persist).[8] Consider again the case of the chemical species. The first species will not manifest redness unless the second is present, while the second will not manifest redness unless the first is present. But all this demands is an explanation for why they combined in the first place, since that original composition is what actualized the conditional potencies to transition into a state of actuality. And once the composition’s components have their conditional potencies actualized in the first place, they remain in a state of actuality unless separated by (say) some chemical or physical process.

“But,” one may object, “surely that is the very question at issue—namely, whether conditional potencies, once actualized, remain in a state of actuality with respect to one another.” This is true. But this shows that we cannot (in a non-question-begging manner) assume from the get-go an answer either way. In particular, we would beg the question if we assumed from the get-go that conditional potencies, once actualized, do not remain in a state of actuality. But such a presupposition is precisely what Feser needs for his form-matter interdependence argument to succeed. For if form and matter interdepend but also (individually) have the conditional potency to persist in existence once combined, and if conditional potencies, once actualized, remain in a state of actuality, then it is simply false that vicious circularity ensues in our explanation of the present existence of some substance. This is because the explanation of the present existence of a substance would not be in terms of form’s dependence on matter and matter’s corresponding dependence on form. Rather, the explanation would be in terms of (i) the cause of the origination into existence of the substance (and thereby the composition of the matter and form), (ii) the actualization of the requisite conditional potency within form and matter, (iii) the nature of conditional potencies (namely, to remain in a state of actuality once actualized[9]), and (iv) there being no sufficiently destructive causal factors operative.

Finally, as with the previous objection, this objection doesn’t constitute an argument against EIT as such. For this objection, like the previous one, only applies to material objects. But we’ve already seen that inertialists need not be wedded to materialism.

7.6 Contingent Natures

Feser offers another argument against EIT from contingent natures. He begins with the following illustration:

To take an example I have often used, suppose you explain, to someone who has never heard of them before (a young child, say), the nature or essence of a lion, of a Tyrannosaurus Rex, and of a unicorn. Then you tell him that, of these three animals, one exists, one used to exist but has gone extinct, and the other never existed and is fictional. You ask him to tell you, based on his new knowledge of the essences of each, which is which. Naturally, he couldn’t tell you. For there is nothing in the essence or nature of these things that could, by itself, tell you whether or not it exists. Existence is something additional to the essence of a contingent thing.  It doesn’t follow from such a thing’s essence. (2021)

But suppose I grant all this. All the child should conclude is that—precisely because there is nothing about a contingent thing (or its nature) that tells us whether it exists—there must be some other factor that explains why the contingent thing exists. In other words, we need some reason why the contingent thing is in reality at all. But this, of course, is an entirely separate question from why, once in existence, the thing continues to exist. And, indeed, we would argue that the child would recognize the plausibility of existential inertia as applied to such continued existence. Consider the following dialogue between me and the child.

Joe: Suppose something S exists immediately before a given moment m. Now, for S to fail to exist at m despite existing immediately before m is for some kind of change to occur. Of course, it’s not as though S undergoes some alteration in this process, since S doesn’t become something different. But still, there is some kind of change here, whether in the ontological inventory of what there is, or in the incorporation of what were previously S’s parts into parts of something else, or in the passing away of a state, or whatever.[10]

Child: That seems reasonable to me.

Joe: But changes of state (i.e., cases where some new state comes to be or some old state passes away) plausibly require some cause. It’s not as though a raging tiger could just spring into existence in this room right now; that would require some cause.

Child: Yeah, changes of state seem to require causes.

Joe: So, if there is no cause that induces the relevant change of state, then there won’t be such a change.

Child: That follows.

Joe: So, if there is no cause that induces S to cease to exist at m—that is, if there is nothing that comes along to destroy S—then S will not cease to exist at m. And in that case, S will persist to m. For you granted earlier that S’s failing to exist m despite existing before m constitutes some kind of change. In particular, it’s a change of state in the sense of an old state passing away. And in that case, we get the conclusion that so long as nothing destroys S from immediately before m through m, then S will exist at m. We derived this in a manner that removes mystery as to why and how S exists at m.

Child: That makes sense.

In this (obviously gerrymandered) conversation, we have a seemingly perfectly illuminating inertialist-friendly explanation of why S exists at m once S is in existence. The explanation tells us precisely how and why S exists at m. And whether or not existence follows from the essence of a contingent thing is entirely irrelevant to this point.

Feser continues with his objection: “The point for the moment is this. If nothing about the essence or nature of a thing entails that it exists at all in the first place, then it is hard to see how anything about its essence or nature could entail that will persist in existence once it does exist” (2021). But—by way of response—nothing in the exchange above assumes that it was something about the essence or nature of the contingent thing that explains why the object persists. Totally separate explanatory facts were cited. And so Feser’s point doesn’t support the denial of EIT.

And my points here don’t merely apply to the no-change account proffered in the above dialogue. The same holds true of other metaphysical accounts of EIT. Even those who accept a tendency-disposition account on which things have an inertial tendency by nature should not be convinced by what Feser says. They will simply retort that if you leave out a tendency to persist in your description of their essences, then your description is simply incomplete.

I make two final points in response. First, Feser’s argument rests on a non-sequitur. From the fact that <nothing about the nature of a contingent object entails that it exists>, it simply doesn’t follow that <nothing about the nature of a contingent object entails that, once in existence, it will persist in the absence of both destruction and external sustenance>. Indeed, instances of this inference schema are clearly false. Consider that—by Feser’s own lights—nothing about the nature of a contingent object entails that it exists. But it is nevertheless false—again, by Feser’s own lights—that nothing about the nature of a contingent object entails that, once in existence, it will lapse into non-being unless sustained by God. Thus, merely from the fact that <nothing about the nature of a contingent object entails that it exists>, it does not follow that <nothing about the nature of a contingent object entails that, once in existence, the object is (or will be) F>.

Second, suppose—contrary to what I’ve argued—that Feser did show or render plausible the claim that contingent things do not enjoy existential inertia. This does not entail that EIT is false. For EIT quantifies over a subset of temporal concrete objects. The inertialist could simply hold that while contingent things do not inertially persist, there is nevertheless some foundational necessary temporal concrete object(s) that inertially persists. (Theist-friendly examples include the neo-classical or panentheistic temporal God, while non-theist-friendly examples include one or more foundational quantum fields, or a spatiotemporal wavefunction, or a collection of fundamental particles, or whatever.) In this case, it is false that nothing about the necessary foundation demands its existence or its persistence; indeed, the opposite is true. Hence, even if—contrary to what I’ve argued—Feser’s argument works, EIT is not threatened. (To be sure, Feser might try to adduce other arguments claiming that only the classical theistic God could be necessarily existent. But that is a separate argument from the one under consideration, and my sole purpose here is to point out that the argument under consideration need not move an inertialist to abandon their position.)

7.7 Vicious Circularity

Feser also charges that EIT is viciously circular. Applying his objection to an example of a contingent substance (viz. water), he writes:

Existential inertia would be a property or power of the water. So, the water’s persistence from t – 1 to t would, on this account, depend on this property or power. But properties and powers depend for their reality on the substances that possess them. So, we seem to have a situation where the water’s persistence depends on that of a property or power which in turn depends on the persistence of the water. (2021)

I have several responses. First, few (if any) of the metaphysical accounts of EIT developed in Section 5 treat inertial persistence as a property or power of substances. We saw that tendency-disposition accounts can be cast in metaphysically lightweight terms that commit to the existence of neither a property nor power corresponding to inertial persistence. In fact, Beaudoin’s tendency-disposition account merely cited the absence of a tendency to expire in conjunction with the non-exercise of potentially destructive powers. Similarly, transtemporal accounts like the Transtemporal Explanans adduced earlier do not postulate a property or power of a substance that explains its persistence. Instead, what explains persistence is transtemporal explanatory (e.g., causal) connections that relate the successive phases of objects’ lives. Law-based accounts cite laws of nature, and many such accounts do not treat laws as properties or powers of substances. Clearly neither objectual nor propositional necessity accounts treat inertial persistence as a property or power of substances. Finally, no-change accounts like the No-Change Explanans adduced earlier make no appeal to properties or powers of substances. Thus, Feser’s criticism here has no teeth against EIT.

To drive this point home, consider again Explanandum, Transtemporal Explanans, and No-change Explanans from earlier:

Explanandum: S’s existence at m.

Transtemporal Explanans: (i) There is an absence of sufficiently causally destructive factors operative on S from m-1 to m (where m-1 is the moment immediately prior to m), and (ii) the state and/or existence of temporal concrete objects (or, at least, those within EIT’s quantificational domain) at a given moment at which they exist causally produces their existence at the next moment provided that no sufficiently causally destructive factors are operative.[11]

No-change Explanans: (i) S existed immediately before m (i.e., at m-1); (ii) if S existed immediately before m but fails to exist at m, then S’s cessation is (or involves, or entails) a change; (iii) nothing causally induces S’s cessation at m-1 or m—that is, nothing destroyed S from the immediately prior moment, m-1, through m; and (iv) a change occurs only if some factor causally induces said change.

For Feser’s circularity objection to work, the explanatory facts in each explanans must presuppose the (explanatorily or ontologically) prior obtaining of Explanandum. But that is simply untrue. I think it is clear from inspection that neither Transtemporal Explanans nor No-change Explanans presuppose the prior reality or obtaining of Explanandum. In other words, none of the explanatory facts are dependent upon S’s existence at m. And in that case, Feser’s allegation of viciously circular (explanatory) dependence has no teeth against such explanans. It is simply false, of both explanans, that there is some property or power that both explains and is explained by some fact. And the other metaphysical accounts likewise do not fall prey to charges of vicious circularity.

Here’s another response to Feser’s vicious circularity charge. Suppose—contrary to what I believe—that existential inertia is a property. This would only be problematic if we accepted the controversial thesis that properties ground character—i.e., that it is in virtue of possessing (exemplifying, instantiating) (say) the property redness that something is red. But suppose we reject this thesis and adopt its opposite: it is rather in virtue of being red that something possesses the property redness. Under this anti-character-grounding view, it is simply false—pace Feser—that existential inertia’s being a property entails that the water exists at m (or persists from m-1 to m) because it has the property of existential inertia. Rather, the substance has the property of existential inertia because it exists at m (or persists from m-1 to m) in an inertial fashion. So even if existential inertia were a property, Feser’s argument still wouldn’t work.

In the next sub-sections, we will consider arguments against EIT deriving from Feser’s various causal ‘proofs’ of God.[12]

7.8 Aristotelian proof causal principle

Premise (7) of Feser’s Aristotelian proof states that the “existence of S at any given moment itself presupposes the concurrent actualization of S’s potential for existence” (2017, p. 35). This premise says that S is concurrently causally sustained at each moment of its existence. On behalf of premise seven, Feser (2021) writes:

The basic idea is this. Consider a collection of particles of type P which constitute water at time t.  Though they actually constitute water at t, there is nothing in the particles qua particles of type P that suffices to make them water rather than one of the other alternatives mentioned.  Again, qua particles of type P they have the potential to constitute water, or separate quantities of hydrogen and oxygen, or some other substance or aggregate of substances.  So, there must at t be something distinct from the collection which actualizes its potential to be water, specifically.

What to make of this justification?

Here are two replies.

First, the inference to the claim that “there must at t be something distinct from the collection which actualizes its potential to be water, specifically” is a non-sequitur. Merely from the facts that the collection as such (qua collection) does not suffice for the collection’s actually constituting water at t, the only thing that follows is that there must be some other sufficient condition(s) for the collection’s constituting water at t. What doesn’t follow is that this other sufficient condition(s) is an external sustaining, efficient, actualizing cause. The other sufficient condition(s) need only be some explanation of why the collection constitutes water at t. But there are whole swathes of explanations of why the collection constitutes water at t that don’t adduce some outside sustaining or conserving efficient cause. I have already explored these in detail in Section 5. Here’s a reminder and quick-and-dirty summary:

  • A tendency or disposition to persist in existence (à la tendency-disposition accounts, which can be construed in metaphysically heavyweight or lightweight ways);
  • Transtemporal explanatory relations (e.g., causal relations) obtaining among the successive phases of objects’ lives or among their temporal parts (à la transtemporal accounts);
  • Laws of nature that govern or otherwise explain the evolution of systems and/or objects over time (à la law-based accounts);
  • The primitive metaphysical necessity of the existential inertial thesis (à la propositional necessity accounts);
  • The metaphysically necessary existence of some foundational temporal concrete object(s), such as the neo-classical theistic God or various naturalist-friendly proposals (à la objectual necessity accounts); and
  • Persistence being the absence of change and so adequately explained by the absence of sufficiently destructive change-inducing factors (à la no-change accounts).

To be sure, there are more besides. I’m simply giving a flavor of the explanations on offer that make no appeal to conserving or sustaining causes.

Second, even if successful (it’s not), the reasoning here would only justify the claim that contingent objects require a sustaining cause. For the reasoning hinges on the idea that the particles could compose something else (i.e., that they need not constitute S) at t, and hence that there must be some cause that explains why they constitute S at t. But this is simply inapplicable to necessary beings, since their parts (if they have them) don’t have the potential to compose something else. And recall that EIT only quantifies over temporal concrete objects or some sub-set thereof, and hence it is compatible with every contingent thing require a sustaining cause (so long as the necessary concrete object(s) upon which they depend inertially persist).

But perhaps someone will argue that, per the Principle of Sufficient Reason (PSR), the present existence of the water (or any substance for that matter) requires an explanation. And absent a concurrently operative existential sustaining cause, there is no explanation for the present existence of the water. Call this response the ‘PSR Response’

The PSR Response, however, won’t work. For, plausibly, past things can and do legitimately explain (at least in large part) the existence of present things.[13] Indeed, it seems difficult to square our ordinary, common sense explanatory practices (and, for that matter, our scientific practices) with a position on which past things lack explanatory force. Delving into arguments in favor of the legitimacy of past things’ explaining present things, though, would take us too far afield given present purposes. Two notes suffice for now.

First, a dialectical point: Feser himself agrees that past things can causally explain present things.[14] Second (and more importantly), a point concerning defeaters: even if one holds that past things cannot legitimately explain present things, what matters for present purposes is that neither Feser nor the (hypothetical) proponent of the PSR Response have—in the dialectical context at hand—given those who think the present existence of an act-potency composite object could be explained by the some past thing (say, the immediately temporally prior[15] state and existence of the object in question in conjunction with the absence of sufficiently destructive factors operative) any reason to abandon their view. And in that case, the PSR Response and premise seven face an undercutting defeater, since they (i) require it to be false that the present existence of such an object could be so explained, but (ii) do not provide adequate reason to think it is false. So, I do not think the second response (the PSR Response) on behalf of Feser succeeds.

Now, Feser is happy to grant that what happened prior to t is part of the explanation of the water’s existence at t(Feser 2021). But what is in question is whether what happened prior to t is by itself sufficient to explain the water’s existence at t. And yet nothing we said above justifies why this is by itself sufficient. How might we respond to this charge?

In response, I note that the charge misunderstands the dialectical context. Yes, what is in question is whether what happened prior to t is by itself sufficient to explain the water’s existence at t. But—and this is crucial—Feser is the one offering a positive argument one of whose premises requires that what happened prior to t is not sufficient to explain the water’s existence at t. By contrast, I did not take a stance on whether or not what happened prior to t is by itself sufficient to explain the water’s existence at t. Rather, I simply pointed out that nothing in premise (7) or what Feser says on its behalf gives those who do think <what happened prior to t is sufficient to explain the water’s existence at t> sufficient reason to abandon their view. The onus is not on me to give positive reasons for thinking what happened prior to t is sufficient to explain the water’s existence at t. Rather, all I need to do is point out that (i) in order for Feser’s proof to succeed, he needs to positively show that what happened prior to t is not sufficient for the water to exist at t, and that (ii) he has simply not showed this. It is thus irrelevant to point out that I say nothing to justify the claim that what happens before t is explanatorily sufficient.[16]

Now, in Feser (2017, ch. 1), all we’re offered by way of justifying the claim that what happened prior to t is insufficient to explain why the water exists at t, as far as we can tell, is the following passage:

[I]t is that matter’s potential to exist as water that is being actualized right now. Why? It is no good to answer that such-and-such a process occurred at some time in the past so as to combine the hydrogen and oxygen in just the right way. That tells us how the water got here, but that is not what we are asking about. It is also no good to point out that nothing has yet come along to separate out the hydrogen and oxygen. That tells us how the water might someday go out of existence, but that isn’t what we’re asking about either. What we’re asking about, again, is what keeps the water in existence at any instant at which it does in fact exist. (2017, p. 26)

But this amounts to simply asserting that the appeal to past things is insufficient to explain the present existence of the water; it is hardly a justification for it. The philosopher who thinks that the existence of S at moment m is adequately explained by the conjunction of (i) S existed immediately before m and (ii) nothing destroyed S from then through m will simply say: “Au contraire! On my view, these do suffice to explain it. Nothing you say in the quoted passage gives me any reason to think my proffered explanation is inadequate. You can retort that it’s ‘no good’, but you need to show why it’s no good. My view is precisely one according to which the conjunction of (i) and (ii) tells us not merely how the water got here, and not merely how the water might go out of existence, but also why the water exists at m. All you have done is simply asserted that the conjunction of (i) and (ii) doesn’t tell us why/how the water exists at m. But that’s precisely my view. And merely asserting a denial of my view is not grounds for rejection of said view. And note, moreover, that the onus in the present dialectical context is not on me to positively demonstrate why (i) and (ii) suffice to explain S’s existence at m; rather, you are the one giving a positive argument here, and hence you are the one who needs to give me sufficient reason to think my view is false.”

Moreover—and note that we do not need to do this for our criticism to succeed—here is a plausible positive justification of why something like (i) and (ii) do suffice to explain S’s existence at m. For S to fail to exist at m despite existing from [m*, m), where m* < m, is for some change to occur. But a change occurs only if some factor causally induces said change. Hence, if no factor causally induces a change, then the change won’t occur. Thus, if no factor causally induces S to fail to exist at m despite existing from [m*, m), then S exists at m. Once we add that nothing came along to causally induce this—that is, once we add that nothing came along to destroy S from m* to m—it simply follows that S exists at m.

Here, we seem to have a perfectly respectable, perfectly legitimate explanation of S’s existence at m. And this does, indeed, tell us how S exists at m. That was a straightforward deduction of the explanatory facts cited (namely, (i) S existed immediately before m, and so S’s ceasing at m would constitute a change; (ii) nothing causally induced S’s cessation at m (i.e., nothing destroyed S from the immediately prior moment(s) through m); and (iii) a change occurs only if some factor causally induces said change). And so we do, indeed, have a sufficient explanation for S’s existence at m, one that doesn’t adduce some extrinsic sustaining efficient cause. For me at least, the explanation certainly seems to remove mystery as to why/how S exists at m. And unlike what allegedly afflicts the explanatory facts adduced in the quoted passage above from Feser (2017, p. 26), the present explanation does, indeed, tell us how S exists at m. (I discuss and defend other metaphysical accounts of EIT in Section 5.)

7.9 Neo-Platonic proof causal principle

Premise (3) of the Neo-Platonic proof embodies what I will call the Neo-Platonic Causal Principle as applied to composite objects (i.e., concrete objects with parts):

Neo-Platonic Causal Principle (NPCP): Any composite object requires a sustaining efficient cause of its existence—that is, an outside or external cause that concurrently combines the parts together to constitute the object.

Why should we believe NPCP?

The central motivation seems to be a kind of explicability requirement—that is, a requirement that the combination of parts (at any given moment of a composite object’s existence) be explained. Feser proceeds through candidate explanations (the whole itself, say, or one or more of its parts) and finds them all wanting. Rejecting appeal to inexplicability or brute facts, the only remaining explanation (according to Feser) would be some external cause or principle that accounts for the unity of the parts.

The first thing to note is that Feser’s survey of explanations of the combination of parts entirely overlooks every single metaphysical account of EIT sketched earlier. None of those require the whole composite object at t to explain the combination of its parts at t, and none of them require one or more parts of the composite object at t to explain the combination of its parts at t. And so NCPC and what Feser says on its behalf are utterly impotent to refute or undermine EIT. For as we saw in Section 5, there are whole swathes of legitimate, inertialist-friendly explanations of persistence—and, a fortiori, explanations of why a temporal objects parts are unified or combined at a given non-first moment of its existence.

Thus, the NPCP is dead on arrival when it comes to arguing against EIT. Nevertheless, let’s continue with our appraisal of it, since we can glean insights therefrom.

In order to evaluate the central motivation for NPCP from explicability, let’s proceed through Feser’s reasons for ruling out various alternative explanations of the unity (at any given moment) of a composite object. For ease of exposition, I’ve broken down the reasoning (in Feser, 2017, ch. 2) into a series of four steps. Let’s consider these steps in turn.

Step one is that all composites depend on their parts. Feser writes that “a composite is less fundamental than its parts in the sense that its existence presupposes that its parts exist and are put together in the right way” (2017, p. 70). Moreover, argues Feser, “a composite depends on its parts not merely (and indeed not necessarily always) in a temporal sense, but more fundamentally (and always) in an atemporal sense. At any particular moment, a composite thing’s existence will presuppose that its parts exist and are put together in the right way at that moment” (p. 70).

One worry thus far is that the mere fact that a given composite object presupposes the existence and combination of its parts does not entail that it depends (in some causal or metaphysical sense) on those parts. It seems only to entail that the existence and combination of the parts is simply a necessary condition for the existence of the whole. But this only implies a kind of logical or counterfactual dependence, not an explanatory dependence. (X logically or counterfactually depends on Y just in case had Y not existed, X would not have existed. Explanatory dependence has much more metaphysical ‘oomph’ (as it were) and conveys a causal or non-causal (i.e., grounding) relation between X and Y. In such cases, X’s existence and/or character is explained or accounted for in terms of Y’s existence, character, or activity.)

One way to see that logical dependence neither means nor entails explanatory dependence is to consider that logical dependence can be symmetric, whereas explanatory dependence is necessarily asymmetric. On certain plausible views in the philosophy of mathematics, in any world in which Socrates exists, Socrates’ singleton set exists. Socrates therefore cannot exist without his singleton set existing. But it is likewise the case (again, under plausible views in philosophy of mathematics) that singleton sets exist only in worlds wherein their members exist. It would follow that Socrates’ singleton set cannot exist without Socrates existing (cf. Baddorf (2016, p. 410)). Hence, there’s a logical dependence in both directions. But it’s also clear that the set is in some sense derivative of Socrates’ more fundamental existence.

Indeed, even in the case of the radically independent God of classical theism, there are a whole host of necessary conditions or presuppositions of his existence. For instance, God’s existing and being the thing he is presuppose the laws of identity and non-contradiction. And despite plausibly presupposing these as necessary conditions, God is still not dependent on another for his existence.

It should be clear, then, that logical dependence can be symmetric. But could explanatory dependence be symmetric? I don’t think so. A (vicious) circle of causal, ontological, or explanatory dependence for existence seems metaphysically impossible. For if x causes or explains the existence of y, and y causes or explains the existence of x, then x is both prior to y (on account of x’s causing/explaining y’s existence) and posterior to y (on account of y’s causing/explaining x’s existence), which is a contradiction.

But perhaps Feser’s thought is one about fundamentality or grounding: wholes are dependent on their parts insofar as parts are more fundamental than and/or ground the existence of their wholes. But while this might be true in some cases, it’s by no means clear that all cases of part-whole relationships fit this schema (especially given the wide classical theistic understanding of parts). For it seems eminently plausible that some wholes ground the existence of their parts. As Baddorf writes:

[I]t is far from obvious that the only kind of thing that could satisfactorily explain compresence is an outside sufficient cause. … [The neo-classical theistic] God’s tropes are dependent upon God. This suggests another explanation for their compresence: they are compresent because they are each grounded in God. This is not a causal explanation, but it is plausible to think that it is an explanation nonetheless. … This conclusion can also be supported by more general argument. It is plausible that tropes are individuated by their bearers and so cannot exist without them. Or, similarly, it is plausible to think that tropes cannot exist without their bearers since they are merely ways their bearers are. (2016, pp. 408-409)

While Baddorf focuses solely on the distinction between God and his property instances, his central point is simply that it seems eminently plausible for some wholes (e.g., the neo-classical God) to ground their ‘parts’. And this isn’t an ad hoc move, either; on one prominent view, tropes are explanatorily posterior to (in virtue of being individuated by) their substances.

Moreover, whole-to-part explanation or grounding is a very (broadly) Aristotelian notion. For Aristotle, a substance’s form “makes [its] parts what they are and organizes them into a unified whole” (Cohoe 2017, p. 756). More generally, Aristotelianism conceives parts of substances as in some sense less fundamental than the substances they compose, since their identities are intelligible only in light of the substances to which they belong. (For instance, something’s being your heart seems to presuppose your existence as a whole, integrated, functionally unified substance.)

Naturally, debates concerning divine aseity overlap quite heavily with the considerations adduced above. According to divine aseity, God is “an absolutely independent being—a being that does not depend on anything else for its existence” (Fowler 2015, p. 115). One of the principal motivations for DDS derives from aseity. If there were some item intrinsic to but distinct from God—that is, if God had some part—then God would be dependent on something that is distinct from God. As Aquinas writes, “every composite is posterior to its component parts, and is dependent on them; but God is the first being” (Summa Theologiae I, q3 a7). One influential response to this line of reasoning derives from what Fowler calls the Doctrine of Divine Priority (DDP): “For all x, if x is a proper part of God or x is a property of God, then x depends on God for its existence” (2015, p. 122). And since dependence is asymmetric, God does not depend on God’s parts. Like Baddorf, then, Fowler offers whole-to-part grounding as a viable model for understanding divine aseity and the independence of at least one whole.

What matters for present purposes, though, is that nothing in Feser’s Neo-Platonic proof—and, moreover, nothing Feser says on behalf of its premises—gives any reason for thinking that such views of divine aseity fail.

In any case, the live possibility (and, I would argue, plausibility) of whole-to-part grounding renders Feser’s step one far from demonstrated. It is not at all clear that any whole whatsoever depends on its parts, and there are plausible views in mereology and metaphysics according to which this step comes out false. Nothing in Feser’s step justifies a denial of these mereological and metaphysical positions. I conclude, then, that step one fails.

Let’s turn, then, to step two: wholes cannot cause the combination of their parts. Feser writes the following in connection to a chair (the whole) and its parts:

How do the parts of a composite come together to form the whole? It can’t be the composite itself that causes this to happen. … [A]t any particular moment, the existence of the whole depends on the existence and proper arrangement of the parts. And the chair as a whole can’t be the cause of those parts existing, and being assembled in just the right way, at that moment. We would in that case have an explanatory vicious circle, insofar as the existence of the whole would depend on the existence and arrangement of the parts, and the existence and arrangement of the parts would depend on the existence of the whole. (2017, pp. 70-71)

Here’s the problem with this step: even granting that wholes cannot efficiently cause the existence of their parts, this by no means entails that wholes cannot explain their parts. We’ve already found defensible views according to which wholes can non-causally explain (i.e., ground) their parts. And because vicious explanatory circularity is impossible, it follows that in cases of whole-to-part grounding, the parts do not explain the existence of the whole.

Step three of Feser’s case for NPCP is that it’s clear that there are extrinsic causal factors that sustain composite objects in existence:

[T]he existence and arrangement of the chair’s parts at any moment does not depend on the chair itself, but on myriad other factors. … The legs and screws themselves exist at that moment because their respective molecules exist and are combined in certain specific ways… Then there are other factors, such as the temperature in the room in which the chair sits being within the right range. … At any moment at which they exist, their parts exist and are arranged in just the right way, and that is the case only because various other factors exist and are combined in just the right way at that moment. Composite things have causes… (2017, p. 71)

There are numerous problems with this step. First, even if it’s true that macroscopic physical objects clearly have such causes (in virtue of their situatedness within a physical context wherein a host of external conditions (like temperature, pressure, etc.) are in place; though, I’ve already shown why this is wrong—cf. Sections 6.1, 7.1, 7.3.1, inter alia), it by no means follows from this that any composite object whatsoever requires an efficient sustaining cause.

Consider, for instance, the God of neo-classical theism. According to the classical theistic understanding of parts, this neo-classical God is a composite object. But it is extremely unclear why there must be a sustaining efficient cause of his existence (or even how there could be). He is not situated within a physical context; he does not require a host of conditions to be in place in order to exist (like temperature, pressure, etc.); he is a necessary being; and he causes the existence of every concrete object apart from himself. Under such an understanding, it is simply metaphysically impossible for there to be a pre-existing context of concrete objects that causally sustain the neo-classical God’s existence, since there being such a context already presupposes the neo-classical God’s creative activity. Nothing Feser says in any of the steps justifies a demand for a sustaining cause of the neo-classical God’s existence.

Or consider a naturalistic view according to which there exist ultimate, fundamental constituents of physical reality (whether they be particles (like quarks), one or more foundational physical fields, or what have you). No reason is given as to why these composite objects are likewise situated within a physical context wherein a host of external sustaining causes have to operate in harmony in order for such objects to exist. It’s not at all clear why or how temperature, pressure, and the other factors cited are even relevant to the ground layer of physical reality. After all, such a ground layer is precisely that which ultimately accounts for how such physical contexts could even arise in the first place—and so it would be absurd for there to be a more fundamental, pre-existent physical context that sustains them in being.

Now, one might at this juncture raise worries for this last example. Here are two worries one might level. First, one might question whether there even is such a fundamental, ultimate ground layer of physical reality. That’s a good question, but it misunderstands the dialectical context at hand. Feser is aiming to give a positive demonstration of NPCP. One of the approaches he takes to justifying this is step three (as we’ve outlined it), which appeals to (purportedly) clear cases of composite objects’ being sustained in being by a pre-existent physical context. But as we argued in the previous paragraph, one clear way to circumvent this justification is to hold that there is such a fundamental, ground layer to physical reality that provides a context for less fundamental physical realities but does not itself have some more fundamental physical context. Given that this circumvents step three, it follows that the success of step three as a justification for NPCP presupposes that there couldn’t be such a fundamental, ground layer as we’ve described it. The onus of justification is therefore on Feser to justify why there couldn’t be such a layer. Without such justification, step three rests on an unjustified presupposition.

A second worry one might raise is that while there may not be a physical sustaining cause for this ground layer of physical reality, there must nevertheless be a non-physical cause sustaining it in being. But once again, this misunderstands the dialectical context at hand. For even if it’s true that a non-physical sustaining cause is required, nothing in step three justifies this. Step three simply adduces a broader physical context in which non-fundamental physical objects are situated. But no reason (thus far) has been given as to why there must exist a non-physical sustaining cause of this fundamental layer of physical reality. This bridges nicely into step four, one wherein Feser aims to do precisely that.

In the fourth and final step, Feser extends his arguments concerning physical composition to metaphysical composition:

The point is just that what has been said here about ordinary physical parts like chair legs and screws would be true also of metaphysical parts like form and matter, if they exist. … For on the Aristotelian analysis, the form of something like copper or a tree is, all by itself and apart from matter, a mere abstraction rather than a concrete object. … But matter all by itself and apart from any form is, for the Aristotelian, nothing but the potential to be something. It is only actually some thing if it has the form of some particular kind of thing. So, though form and matter are different, there is a sense in which form depends on matter and matter depends on form. We would thus have an explanatory vicious circle if there were not something outside them which accounted for their combination. (2017, p. 73)

Feser then goes on to apply the same reasoning to other (purported) forms of metaphysical composition, like essence-existence composition. There are, however, several problems with this fourth step. First, it’s quite dialectically limited insofar as it rests on contentious commitments to various kinds of metaphysical parts.

Second, even granting such commitments, vicious explanatory circles are metaphysically impossible regardless of whether there’s an extrinsic cause accounting for the viciously intertwined things. As we saw earlier, vicious explanatory circularity entails that x causes or grounds y while y causes or grounds x. This is impossible regardless of whether there is some z that accounts for both x and y, since the very fact of the vicious circularity entails that x is both prior and posterior to y, which is a contradiction. If hylemorphism entails that two distinct things are viciously intertwined—such that the existence of each explains the existence of the other—then that would simply be a reductio of hylemorphism. Far from demonstrating the need of an extrinsic sustaining cause for any composite object, Feser’s step four imputes a metaphysical impossibility to composite objects from the get-go. And if Feser holds instead that x only causes or explains y in some respect distinct from the respect in which y causes or explains x (in order to avoid contradiction), then his argument no longer has teeth. For then there is no vicious circularity, and hence the very means by which he motivated the need for an extrinsic source of both x and y is undercut.

Third, even ignoring the above point, there seem to be perfectly legitimate explanations of the combination of parts that avoid vicious circularity but that don’t adduce extrinsic sustaining efficient causes. We’ve already seen a panoply of such explanations in Section 5. But also consider:

For instance, there is the whole-to-part grounding that we covered earlier. Another seemingly legitimate candidate explanation for the unity of an object’s parts is that it is metaphysically necessary that they be so combined. If we take the ultimate reality (be it the neo-classical God or some ground layer of physical reality) to be a necessary being, the unity of its parts could easily be explained in terms of the metaphysical necessity of said unity. Now, we’re not claiming that metaphysical necessity categorically precludes any further explanation. Rather, we’re simply noting that metaphysical necessity may plausibly itself constitute a kind of explanation for something’s obtaining. Even if the reader disagrees with this, the point is that nothing in Feser’s Neo-Platonic proof gives those who accept (or are even neutral on) the explanatory legitimacy (in principle) of metaphysical necessity any reason to abandon their position.[17]

A third seemingly legitimate candidate explanation is that the kind of thing in question simply requires the obtaining of the explanandum. For instance, perhaps the neo-classical theistic God’s essence simply requires existence, or perhaps unlimited perfection requires existence. And the legitimacy of this form of explanation seems eminently plausible. Consider the really distinct properties[18] ‘being triangular’ and ‘being trilateral’. Why are these co-instantiated? Because they are simply the kinds of properties that require co-instantiation. This seems to be a perfectly legitimate explanation of their compresence or unity in something. It does seem somewhat odd to demand some sort of concurrent sustaining cause keeping the two properties together. And this plausibly generalizes to a being whose parts are all the kinds of things that require their unity and co-instantiation. (This is compounded even further if there is a kind of intrinsic intelligibility to the parts’ unity together: it is no coincidence, for instance, that omnipotence, omniscience, and moral perfection are all compresent and unified together in the neo-classical theistic God.)

Here’s yet another seemingly legitimate candidate explanation. It seems that the neo-classical theist (if they’re unwilling to adopt any of the above explanations) could be well within their rights in holding that the unity of God’s parts flows from one simple component of God. Perhaps pure perfection is the root, the core, the fundamental aspect of God. From pure perfection flows the purely positive great-making properties like omniscience, omnipotence, essential moral goodness, and so on. Pure perfection is like a spring from which flow the individual perfections or great-making properties. In this case, we have one ‘divine part’ (in the broad, classical theistic sense) that explains the unity of God’s parts: all the distinct properties and powers flow from or are explained by the simple property of sheer perfection. Once again, we have an explanation of the unity of the parts here, but nothing outside of or external to God is explaining or causing the unity.[19]

At the very least, Feser’s arguments for NPCP do nothing to address such alternative explanations. The central takeaway of this section, then, is as follows. The detractor of classical theism—whether it be a neo-classical theist or a non-theist—need not appeal to inexplicability or brute facts to account for unity. There are perfectly sensible explanations that don’t adduce sustaining efficient causes. I conclude, then, that Feser’s third premise (i.e., NPCP) is not only unjustified but also dialectically ill-situated (insofar as detractors of classical theism have perfectly legitimate alternative explanations of unity).

I also want to cover something Feser fails adequately to rule out (and which would need to be ruled out for his argument to succeed): why couldn’t the unity of something’s parts have an internal cause for their combination or ‘holding together’ as opposed to an external sustaining cause?

Feser does consider the proposal that something might be held together internally (i.e., without the need for an external sustaining cause to unify the thing’s parts). He considers two ways in which this might happen. First, a thing with two parts, A and B, might have some additional part, C, which unifies them. Second, it might be an unexplained or irreducible brute fact about the thing that it is made up of parts A and B. Neither suggestion, according to Feser, will do. We can set aside the second proposal, as I’m keen to avoid inexplicability. But on the first proposal, Feser demurs:

[T]he argument assumes that for a composite thing to exist, its parts have to be unified by some external cause. But why assume this? Why not suppose instead that it is precisely some part of a composite thing that unifies its parts, rather than something external? Or why not suppose that the fact that a composite thing’s parts are unified is just an irreducible fact about it?

But as Vallicella has argued, neither of these suggestions really makes any sense. Start with the suggestion that the parts of a thing are unified by some further part. For instance, consider a thing composed of parts A and B. What makes it the case that parts A and B are united in such a way that the composite thing in question exists? The suggestion at hand is that there is some further part, C, which accounts for A and B being united. But the problem is that this just pushes the problem back a stage, since we now need to ask what unites C together with A and B. If we posit yet another further part, D, in order to account for the unity of A, B, and C, then we will merely have pushed the problem back yet another stage. And of course the problem will just keep recurring for each further part we posit. (2017, pp. 83-84)

The problem with Feser’s reasoning here is that the two suggestions Vallicella proffers—the only two suggestions Feser considers, mind you—are not exhaustive. Recall that Vallicella and Feser are here considering the possibility of an internal explanation for the unity of a thing. But their suggestions for an internal explanation of A’s being joined to B do not exhaust the possibilities as to why A is joined to B. For if the cause is internal to the composite object in question, it may be either A itself, or B itself, or some further part C.[20] Vallicella and Feser criticize only the third possibility. That is, both Feser and Vallicella criticize an infinitely regressive postulation of additional parts (C, D, E, F, and so on). But this still leaves us with the possibility that either A itself or B itself is the cause of the conjoined-ness or unity of A and B. Feser here explores neither of these possibilities.[21]

Consider, at this juncture, salt. That is, consider a sodium chloride crystal, composed of repeated formula units of sodium and chloride ions bonded together in a lattice or crystal structure. Why are the sodium ions right now joined to or unified with the chloride ions? By our lights, it seems like a perfectly adequate explanation to say that the sodium ions themselves bring about, right now, their conjoined-ness or unity with the chloride ions, while the chloride ions in turn bring about their conjoined-ness or unity with the sodium ions.[22]

And nor is this explanation viciously circular. It would be circular to hold that the sodium ions cause the existence of the chloride ions, and that the chloride ions in turn cause the existence of the sodium ions. But this is decidedly not what this explanation says. Instead, this explanation says that each causes the other to be attracted to itself. This mutual attraction is perfectly intelligible and non-circular, since A’s being attracted to B does not presuppose B’s being attracted to A (and vice versa). These are distinct states of affairs, neither of which presupposes the prior reality or obtaining of the other. (I should add, moreover, that an explanation of the unity of A and B in these terms is similar to one of the explanations of the unity or combination of form and matter—in terms of conditional potencies—adumbrated earlier.)

Furthermore, note that the explanation here does not adduce some further, additional component C that unifies or combines A and B. The question Feser poses, then, as to what unites or combines C with A and B is a non-starter.

NPCP and Trinitarianism

Consider again NPCP’s primary motivation: explicability. More specifically, the compresence of distinct elements together into a unified being demands some explanation—and because the explanation cannot be in terms of the being itself or one of its parts, an extrinsic cause (so the argument goes) is required.

But the exact same motivation for demanding an extrinsic cause of the unity of composite objects seems equally to apply to the multiplicity of divine persons unified in a single Godhead. In other words, there seems to be no justification for NPCP that doesn’t also justify demanding a cause of any Trinitarian being. There seems to be no non-question-begging, non-arbitrary, principled way to delineate the demand for an extrinsic cause of unity in the case of things that are composites of (say) distinct attributes, or of essence and accident, or what have you (on the one hand) and things within which there is a multiplicity of distinct persons and processions (on the other hand).

Indeed, the fact of distinct x’s within a being—i.e., a multiplicity or differentiation among numerically distinct ontological items—is precisely what requires an extrinsic cause in Plotinus’ view. For Plotinus, the One’s simplicity is utterly unqualified (Gavrilyuk, 2019, p. 442). At “Enn. 6.9.4 Plotinus observes that… the One transcends all differentiations characteristic of being… Plotinian simplicity excludes any multiplicity” (Gavrilyuk, 2019, pp. 447-448). Anything apart from the One requires a cause for the unity of its multiplicity, distinction, differentiation, and qualification. Hence, for Plotinus, a differentiation or multiplicity of distinct divine persons would demand an extrinsic sustaining cause.

Clearly the Trinitarian God cannot have an extrinsic sustaining cause of its being. But what, then, could explain the unity of the distinct persons in God? If the answer is that there’s no explanation, then that seems to be a perfectly legitimate move for the detractor of the Neo-Platonic proof in explaining the most fundamental composite thing(s). If the answer is that it’s simply metaphysically necessary that there be three distinct persons unified in one God, then that also seems to be a perfectly legitimate move for the detractor of the Neo-Platonic proof.

Perhaps the explanation is in terms of one of those very divine persons (the Father, say)? But once again this will undercut the Neo-Platonic proof. For this explanation amounts to the proposal that one of the x1, x2, … xn explains the unity of those very x’s. But once this kind of explanation is granted, then it seems that the unity of the fundamental composite object(s) could be explained in virtue of one (or more) of its parts. If one of those very x’s can explain the unity of them all, then it seems this should equally be the case when the ontological items in question (the x’s) are parts (again, in the very broad classical theistic understanding of parts). And hence it would not necessarily be the case that any composite object must be explained in terms of some extrinsic sustaining cause.

Perhaps there is some fourth thing, some fundamental aspect of God, that explains or accounts for their unity? But (i) this is plausibly incompatible with DDS (since this thing would be some positive ontological item within but distinct from God himself), and (ii) this reduces to the answer in the previous paragraph, since there’s the further question of what explains the unity or connectedness or compresence of {Father, Son, Holy Spirit, this fourth fundamental aspect of God}. And any explanation in terms of the final element in the set is just to cite one of the very x’s among x1, x2, … xn to account for their unity.

Fundamentally, then, my challenge for the Trinitarian proponent of the Neo-Platonic proof is as follows. When it comes to explaining to unity (togetherness, conjoined-ness, compresence) of the distinct divine persons within one being (the Godhead or G), the explanation is either (i) internal to G, (ii) external to G, (iii) G itself, or else (iv) there is no explanation. But if the explanation is internal to G, then something within a unified plurality can explain the unity of said plurality—in which case, the same could apply to the objects within NPCP’s domain of quantification, thereby undermining NPCP’s demand for an outside explanation of such objects. If the explanation is external to G, then there is something outside of or external to God which explains something about God, which is obviously incompatible with classical theism. If the explanation is G itself, then an object within which there is a unified plurality can explain the unity of said plurality without recourse to any external thing—in which case, the same could apply to the objects within NPCP’s domain of quantification, thereby undermining NPCP’s demand for an outside explanation of such objects. Finally, if there is no explanation, then once again the same could apply to the domain of NPCP, thereby undermining it. Thus, either classical theism is false, or else NPCP (and, hence, the Neo-Platonic proof itself) is undermined.

In general, then, it seems that any proposal for explaining the unity of the distinct Trinitarian persons will undercut NPCP and a fortiori the Neo-Platonic proof.

7.10 Thomistic proof causal principle

Maybe Feser’s Thomistic proof can count as an argument against EIT?

One of its premises is the following:

Thomistic Proof Causal Principle (TPCP): Since its essence and existence remain really distinct at every moment at which it exists, including here and now, its existence must be imparted to it by some cause distinct from it at every moment at which it exists, including here and now.

TPCP denies EIT which, you will recall, states (roughly) that temporal concrete objects (or some subset thereof) persist in existence in the absence of sustenance and destruction. Because temporal objects are (we can grant arguendo) things in which essence and existence are distinct, the truth (or rational defensibility) of EIT would (or does) undermine the Thomistic proof. And as we’ve seen, the arguments thus far articulated against EIT in the literature can be met, and there are also many motivations for holding the inertialist view.

But let’s consider the distinctive reasons cited on behalf of TPCP.

The argument Feser (2017, pp. 124-125) provides runs as follows. Existence cannot follow or flow from the essence of an essence-existence composite; nor can it impart existence to itself (for that would be self-causation). There must, then, exist some extrinsic cause that imparts existence to an essence-existence composite.

As Feser recognizes, however, this reasoning is incomplete. For even if successful, it does not entail the need for concurrent, existential causal sustenance of essence-existence composites—for all the reasoning shows, the object may very well need to be brought into existence (thereby effecting composition between essence and existence, if we must speak this way). Thereafter, however, the union of the two may very well be inertial.

To solve this, Feser argues that the line of reasoning “applies to a thing not only before it comes into being and as it comes into being but always, even after it has come into being” (2017, p. 126). He continues:

Fido’s existence is distinct from Fido’s essence, does not follow from Fido’s essence, and cannot be imparted by Fido to his essence. All of these things are true not only before Fido exists and at the time he is conceived, but also after he comes into being, and indeed at every moment he is alive. Fido’s existence here and now is distinct from his essence and does not follow from his essence. So, here and now there must be some cause that adds or imparts existence to that essence. Otherwise Fido would not exist here and now any more than he did before he was conceived. He would ‘blink out’ of existence or be annihilated. (2017, p. 126)

The principal problem with this argument is that it simply does not justify the requirement of a concurrent sustaining cause. For even if S’s existence at t is distinct from S’s essence at t, such that neither S at t nor S’s essence at t could account for or explain S’s existence at t, it simply doesn’t follow that the only remaining option for explaining S’s existence at t is a concurrent, sustaining cause which imparts existence to S at t. But this is precisely what Feser needs for his argument to succeed.

Moreover, we have seen that there exist alternative explanations of S’s existence at t that decidedly reject efficient sustaining causes. For instance, S’s existence at t could easily be explained (for all Feser’s argument shows) by the state and existence of S immediately prior to t in conjunction with there being no sufficiently destructive factors operative (à la transtemporal accounts), or by the very nature of persistent existence as the maintenance or non-disruption of a state of unchanging actuality (à la no-change accounts), or by the metaphysical necessity of persistence in the absence of sufficient destructive factors (à la propositional necessity accounts), and so on. Feser has given no reason to rule out any of these explanations. I conclude, then, that EIT is not at all threatened by TPCP and what Feser says on its behalf.

I think an analogy with mechanical or spatial inertia is helpful at this point. Consider: spatial location is almost always (perhaps always) numerically distinct from the essence of the thing occupying said location. A cup, for instance, is distinct from a particular spatial location, since the cup can exist without being located at such a position. (This shows, moreover, that such a spatial location is not even essential to the cup.) But the mere fact that the cup’s particular spatial location is distinct from its essence (and not even included in its essence as one of its essential properties) doesn’t entail that the cup’s being in this particular spatial location (at any moment at which it is in such a location) requires some kind of continuously concurrent external causal ‘keeper’ or ‘sustainer’ of the cup’s being in said location. Indeed, the opposite seems to be the case: we have good reason to think that no such continuously concurrent sustenance is required in order for the cup to simply retain its spatial position. And if this is true of spatial location—which, again, is distinct from the essence of objects (and isn’t even one among their essential properties)—why couldn’t it equally be true of existence?

In any case, nothing in the Thomistic proof threatens EIT. There’s a whole host of inertialist-friendly explanations of the existence of S at t. One needn’t adduce an extrinsic sustaining cause, contra TPCP.

7.11 Rationalist proof causal principle

Maybe Feser’s Rationalist proof can count as an argument against EIT?

One of its premises is the following:

Rationalist Proof Causal Principle (RPCP): Furthermore, that an individual contingent thing persists in existence at any moment requires an explanation; and since it is contingent, that explanation must be in some simultaneous cause distinct from it.

In light of the various metaphysical accounts of EIT from Section 5, RPCP is straightforwardly false. For there are other ways to explain the persistence of temporal objects that don’t invoke sustaining efficient causes of their existence. Nothing in RPCP or what Feser says on its behalf addresses any such accounts. There is thus no argument against EIT here.

Moreover, RPCP is actually compatible with EIT. We’ve already seen why: EIT quantifies over temporal concreta or some subset thereof. Thus, even if contingent things require sustenance from without, some necessary temporal foundation may very well inertially persist. So even ignoring the first point above, this is not an argument against EIT.

Now, perhaps you think a necessary concrete object must be purely actual and hence timeless (and hence not within EIT’s quantificational domain). But this is straightforwardly false: there could easily be a non-purely actual, necessarily existent being. Its potencies must simply be potencies merely for accidental (rather than substantial) change (or else mere cross-world variance in properties).

Feser tries to argue as much:

[F]rom the fact that it is necessary, it follows that it exists in a purely actual way, rather than by virtue of having potentialities that need to be actualized. For if it had such potentialities, then its existence would be contingent upon the existence of something which actualizes those potentialities—in which case it wouldn’t really exist in a necessary way after all. (2017, p. 159)

But this is confused. First, the passage equivocates between ‘contingent’ in the sense of ‘possibly absent from reality’ (on the one hand) and ‘dependent upon another’ (on the other).[23] Second, and more importantly, when one accepts the necessary (and independent) existence of a non-purely actual thing (like the God of neo-classical theism), one does not say that it ‘exists by virtue of having potentialities that need to be actualized’. One is only saying that the being in question has various potentialities to accidentally change, none of which require actualization in order for the being in question to exist. Feser needs to argue that the mere fact of having potentialities for accidental change entails that the being couldn’t exist necessarily. But instead he has targeted the wildly different proposal that the necessary being has potencies in need of actualization in order to exist.

7.12 Nemes’ argument

Steven Nemes offers an argument against EIT. I explain and respond to Nemes’ argument in this blog post.

7.13 De Ente argument

One might proffer the De Ente argument as an argument against EIT.

I discuss the De Ente argument—including its relation to EIT—from 2:11:52 to 2:52:20 in this video.

Perhaps a more fundamental problem for any argument against EIT based on the De Ente argument, though, is that the De Ente argument requires pluralism about being. But many philosophers have pushed quite forceful criticisms of pluralism about being. See, for instance, Merricks (2019), “The Only Way To Be”, for a pretty forceful attack on pluralism. [Fun fact: I had Dr. Trenton Merricks on my channel to discuss this article! You can watch it here.] [Note: I’ll be articulating, in broad outline, Merricks’ argument below. Note that Builes (2019), in his Analysis paper “Pluralism and the problem of purity” criticizes Merricks’ argument, but in doing so, he develops a new problem for pluralism that—by Builes’ lights—is even more forceful than Merricks’.]

Before articulating the problem in more detail, let’s get a clear understanding of pluralism and its rival, monism. Trenton Merricks pithily summarizes them as follows:

Monism about being (monism for short) says that everything enjoys the same way of being. So monism implies, for example, that if there are pure sets and if there are mountains, then pure sets exist in just the way that mountains do. Monism can be contrasted with pluralism about being (pluralism for short). Pluralism says that some entities enjoy one way of being but others enjoy another way, or other ways, of being. (2019, p. 593)

The thrust of the objection I’ll be leveling here, then, is that the De Ente argument requires pluralism about being—the view that there are multiple ways to exist. But we should reject pluralism.

Why does it require pluralism about being?

Because the argument requires that God have an esse or act of existence different from creaturely acts of existence. Under such a view, God and creatures exist in different ways. For suppose that God’s existence was the same existence as creatures’ existence. Since God is identical to his existence, it would follow that creatures’ existence just is God. But this is absurd for several reasons. First, with ordinary objects like “stones, trees, dogs, human beings, and the like… there is in each of them a real distinction between its essence and its existence” (Feser 2017, p. 122). The distinction is in objects, i.e., it is within them or intrinsic to them. (At least, that’s how I read the passage.) If I read the passage right, then essence and existence are component parts of ordinary objects. But God is obviously not a component part of things. Second, “there is no such thing in mind-independent reality as a thing’s essence existing apart from its existence (whatever that would mean) or a thing’s existence existing apart from its essence (whatever that would mean)” (2017, p. 123). But if a thing’s existence is identical to God, then clearly a thing’s existence can exist apart from its essence, since God can exist apart from anything else (had he chosen to refrain from creating). Third, the argument says (or requires, or entails, or whatever) that the existence of creatures must be caused by something outside it—something which adds existence to its essence, as it were. Thus, if God were identical to ordinary objects’ existence, then God would be caused. But God is not caused. For these three reasons, it follows that creatures’ existence is not identical to God (and hence God’s existence). And as we’ve seen, this entails pluralism about being. Hence, the De Ente argument requires pluralism about being: God and creatures exist in different ways. God’s existence is not the same existence as creaturely existence.

Why should we reject pluralism?

The argument against pluralism in what follows is applied to act-potency pluralism, but it applies mutatis mutandis to the De Ente argument—simply replace ‘a-existence’ and ‘p-existence’ with ‘d-existence’ and ‘c-existence’ (representing ‘divine existence’ and ‘creaturely existence’, respectively).[24]

The argument as applied to act-potency pluralism

With generic monism and pluralism articulated, we can now consider a specific form of pluralism. This brand of pluralism—what we might call act-potency pluralism—recognizes two different ways of being: being-in-act and being-in-potency. Thus:

Act-potency pluralism: Everything either exists-in-act (hereafter, a-exists) or exists-in-potency (hereafter, p-exists), and a-existing is not the same thing as p-existing.[25]

We will let ∃a (‘there a-exists’) range over all and only those entities that a-exist and ∃p (‘there p-exists’) range over all and only those entities that p-exist. Because ∀xFx iff ~(∃x~Fx), act-potency pluralists endorse parallel biconditionals: ∀axFx iff ~(∃ax~Fx) and ∀pxFx iff ~(∃px~Fx). ‘∀axFx’ is read as ‘everything that a-exists is F’ while ‘∀pxFx’ is read as ‘everything that p-exists is F’. Neither ∀a nor ∀p are fully general, in which case act-potency pluralists—at least those that don’t accept generic existence (which will be defined below)—don’t have a universal quantifier.

We can now consider the following thesis, (26):

(26) We should not accept act-potency pluralism.

In its defense, I will draw on Merricks’ (2019) recent argument for monism as well as the insights in Loux (2012). Here is the argument for (26):

18. Act-potency pluralists should not accept generic existence.

19. If act-potency pluralists do not accept generic existence, then they cannot—by their own lights—state that everything is thus and so.

20. If act-potency pluralists cannot state that everything is thus and so, then they cannot state their view.

21. If one cannot state a view, one should not accept that view.

22. So, if act-potency pluralists do not accept generic existence, then they should not accept act-potency pluralism. (19-21)

23. If act-potency pluralists should not accept generic existence, and if their not accepting generic existence implies that they should not accept act-potency pluralism, then act-potency pluralists should not accept act-potency pluralism.

24. So, act-potency pluralists should not accept act-potency pluralism. (18, 22, 23)

25. If (24) is true, then we should not accept act-potency pluralism.

26. So, we should not accept act-potency pluralism. (24, 25)

I will justify each premise—(18), (19), (20), (21), (23), and (25)—in the following paragraphs. The premise under consideration will be placed in italics at the beginning of the relevant portion of my discussion.[26]

         Premise (18). Generic existence is the ordinary, humdrum existence enjoyed by everything. (Many pluralists reject generic existence, and—as Merricks argues—all pluralists should reject it. For now, I am simply characterizing it.) ‘Generically exists’ is not shorthand for some disjunction like: a-exists or p-exists (or exists1 or exists2 …). “To generically exist”, writes Merricks, is “to exist, or to have being, or to be something” (2019, p. 599). Merricks gives at least three reasons why pluralists should not embrace generic existence. I adapt his reasons and apply them to the present context of act-potency pluralism. Here, then, are three reasons supporting premise (18).

         First, act-potency pluralism is motivated, at least in part, by the conviction (or insight, or intuition) that potential things and actual things don’t exist in the same way. On this point, for instance, Feser writes: “For if potentiality and actuality had being or reality in exactly the same sense, then what could that mean if not that potentiality is really a kind of actuality or that actuality is really a kind of potentiality?” (2017, p. 182). This question seems to voice the conviction that there is not a way of being that potency and act alike enjoy. But this conviction, of course, contradicts the thesis that act and potency alike generically exist. Those who share the conviction should therefore reject generic existence. “But”, writes Merricks concerning abstract/concrete ways of being (which applies equally to actual/potential ways of being), “this conviction is the best motivation for pluralism. So this new version of pluralism [i.e., one that embraces generic existence] contradicts the best motivation for pluralism” (2019, p. 602). This is the first reason why act-potency pluralists should not accept generic existence.

         Second, pluralisms that accept generic existence render themselves vulnerable to the most natural objection to pluralism. This objection—voiced, for instance, in van Inwagen (2014, p. 23)—alleges that radically different properties had by things suffices to explain the vast differences between them. No appeal to different kinds or ways of being is required. But once one accepts that everything generally exists, “it really does seem like a mistake to add that there is another way of being that is correlated with being concrete [or being actual], and another still that is correlated with being abstract [or being potential]” (Merricks 2019, p. 602). Why not simply hold that these correspond to different properties rather than additional ways of being? “And why pick ways of being that are correlated with those particular differences among generically existing entities, as opposed to others?” (ibid, emphasis added). In short, embracing generic existence renders pluralism “particularly vulnerable to the objection that pluralists posit a difference in being where there is instead but a difference in kind among entities that exist in the same way” (ibid, pp. 602-603). This is the second reason why act-potency pluralists should not accept generic existence.

         Third, if everything enjoys one-and-the-same generic existence, then ‘being’ is (or can be) predicated univocally of act and potency. But Feser (2017, pp. 176-184, esp. 180) is explicit that—according to the version of act-potency pluralism presently under consideration—being is not predicated univocally of act and potency but instead only analogically. Thus, act-potency pluralists should not accept generic existence. Merricks extends this third reason further:

This new version [of pluralism] implies that all entities—properties, numbers, mountains, God, creatures, everything—generically exist. This implication is clearly in tension with the sorts of views that virtually all pluralists have tried to articulate and defend. This tension is illustrated by the fact that… historically influential motivations for pluralism are inconsistent with the claim that all entities generically exist. And this tension is not surprising. That is, it is not surprising that pluralism about being is in tension with the idea that there is a single way of being that everything enjoys. (2019, pp. 603-604)

We thus have three reasons to think that act-potency pluralists should not accept generic existence. I take it, then, that premise (18) is on solid footing.

         Premise (19). If act-potency pluralists do not accept generic existence, then they cannot—by their own lights—state that everything is thus and so. This is the parallel of Merricks’ (2019) central contention that “pluralists can, by their own lights, make claims to the effect that everything is thus and so if and only if they take such claims to invoke… generic existence.” (p. 606). And in that case, “they cannot, by their own lights, state claims to the effect that everything is thus and so” (ibid, p. 608).

Without generic existence, act-potency pluralists are left only with a-existence and p-existence and their corresponding existential-like quantifiers ∃a and ∃p.[27] Given interdefinability—that is, given that ∀axFx iff ~(∃ax~Fx) and ∀pxFx iff ~(∃px~Fx)—act-potency pluralists are left only with ∀a and ∀p in terms of their universal-like quantifiers; they don’t have a fully general, universal quantifier.[28] But in that case, act-potency pluralists cannot state that everything—full stop, with perfect generality—is thus and so. Thus, premise (19) is true.

Now, pluralists are aware of this problem—that is, the problem of stating pluralism without employing the universal quantifier. Perhaps the act-potency pluralist can use (mutatis mutandis) Turner’s (2010, p. 33) way of stating (what Merricks calls) two-ways-of-being pluralism.[29] That is, perhaps the act-potency can state their view as follows:

26. ∀ax(∃ay(y=x) or ∃py(y=x)) and ∀px(∃ay(y=x) or ∃py(y=x))

For act-potency pluralists think that everything either a-exists or p-exists. Thus, by the pluralist’s lights, making a claim about all the entities in the domains of each of ∀a and ∀p leaves nothing out.

         The problem with (26), though, is that it is simply a logical truth. This is because (26) simply says that (i) everything that a-exists either a-exists or p-exists and (ii) everything that p-exists either a-exists or p-exists. This is a mere logical truth. But as Merricks points out, “[e]veryone should endorse logical truths. So everyone should endorse [(26)]. So monists should endorse [(26)]. But monists should not endorse [act-potency] pluralism. So [(26)] fails to state [act-potency] pluralism” (2019, p. 595).

         Merricks proceeds through different ways of modifying, supplementing, or replacing (26) in order to state pluralism. But none of them, argues Merricks, succeed. For instance, one might add to (26) the claim that something a-exists and something p-exists. But this modification still fails to state act-potency pluralism, since a pluralist who accepts a-exisence, p-existence, and another particular form of existence—someone we might call an act-potency-plus pluralist—also accepts the modified thesis. Hence, someone who denies act-potency pluralism would accept the modified thesis, in which case it fails to state pluralism. Alternatively, one might add to (26) that there are exactly two ways of being. But stating that there are exactly two ways of being ultimately requires a truly universal quantification: there is this way of being, that other way of being, and these are all the ways of being there are. Stating the latter conjunct requires a universal quantifier. The underlying problem for any such attempt to fix or replace (26)—indeed, the underlying problem for any attempt to state act-potency pluralism—is precisely the failure to state truly universal claims.

Let’s summarize the argument. Act-potency pluralists cannot state that everything is thus and so, and hence cannot state their view (since their view requires stating that everything either a-exists or p-exists). They cannot state that everything is thus and so because they don’t have a truly universal quantifier. For having a truly universal quantifier would entail having the simple, standard existential quantifier (given the interdefinability of the universal and existential quantifiers), which in turn expresses or captures generic existence (‘there exists’ rather than ‘there a-exists’ or ‘there p-exists’ or ‘either there a-exists or there p-exists’). But pluralists should reject generic existence, as we’ve seen.[30]

Premise (20). Recall the premise: If act-potency pluralists cannot state that everything is thus and so, then they cannot state their view. This is clearly true, since act-potency pluralism, recall, states that everything either a-exists or p-exists. This is a statement to the effect that everything is thus and so. Hence, if act-potency pluralists cannot state that everything is thus and so, then they cannot state their view.

Premise (21). This premise says that if one cannot state a view, one should not accept that view. Like (20), I take this to be clearly true.

Premise (23). This premise says that if act-potency pluralists should not accept generic existence, and if their not accepting generic existence implies that they should not accept act-potency pluralism, then act-potency pluralists should not accept act-potency pluralism. Like (20) and (21), I take this to be clearly true. It seems eminently plausible (if not self-evident) that if S should F, and if <S does F> implies <S should Y>, then S should Y. If—given my evidence—I should believe the Earth is round, and my believing the Earth is round implies that I should act as if the Earth as round, then surely it follows that I should act as if the Earth is round. I think, then, that premise (23) is on good footing.

Premise (25). According to this premise, if act-potency pluralists should not accept act-potency pluralism, then we should not accept act-potency pluralism. Once more, I think this premise is clearly true. If, by adopting a view, one thereby ought not adopt it, surely no one should adopt that view!

Having justified each of the respective premises, I conclude that (26) is true.


[1] There may be a sense in which focusing on the redness of the chair as opposed to its existence avoids the problems we raise concerning parts and wholes (and parts’ not being efficient causes of the very being or existence of their wholes). For on some theories of color (e.g., some versions of color realism), color is not composed or constituted by the microchemical properties but is instead efficiently caused by (or realized by, or emergent from, or whatever) them. In response, we note that it’s true that on some theories of color, these will be efficient rather than material causes. But, first, Feser is fundamentally concerned not with the properties of things but instead with the existence of things like the chair. And, indeed, this is precisely the regress he sends us off on in his chapter: parts of coffee ‘actualizing’ the coffee’s existence. Second, if we take this line of response, then we have brought in a deeply controversial account of the nature of color, one present in neither Feser (2017) nor McNabb and DeVito (2020). If a defense against Oppy’s point needs something like this highly controversial account of color, then the Aristotelian proof loses much of its force. (And, moreover, the detractor can simply reject the account and thus reject the purported counterexample to Oppy). Thanks to Tyler McNabb for bringing this to my attention.

[2] As Ingthorsson (2021, ch. 6) points out, it’s well-nigh standard/universal in contemporary philosophy of causation and mereology to hold that the constitution relation is not an efficient causal relation. Note, also, that I gave some Aristotelian-inspired reasons for this position earlier.

[3] It is clear from context, it seems, that this is what Feser means. Consider, for instance, that Feser seems to infer that <no material substance has a tendency of itself to persist in existence once it exists> from the facts that <there are no other internal principles from which such a substance might derive such a tendency> and <form and matter are internal principles from neither of which the substance can derive a tendency to persist>. This inference only seems to work if ‘of itself’ expresses ‘intrinsically or internally’.

[4] See Section 6.2 for an objection to this claim and our responses.

[5] Two notes. First: For simplicity, we here assume that time is composed of smallest units termed moments. Nothing much hangs on this, though; we could explicate each explanans in a continuous-time-friendly manner mutatis mutandis. Our more general point remains unaffected. Second: These are not the only explanantia under no-change and transtemporal accounts. We choose only one explanans from each of these two accounts for ease of exposition.

[6] Alternatively, if one wants to say—at least with respect to the Transtemporal Explanans—that the explanans is intrinsic to S in the sense that the ‘principal explanatory mechanism’—transtemporal causation relating the successive phases of S’s life—is entirely contained within S’s life, that’s fine. Then premise (7) (“there are no other internal principles from which such a substance might derive such a tendency”) would be false, since there is something else intrinsic to S’s life that accounts for S’s inertial persistence—to wit, transtemporal causal relations among the successive phases or states of S’s life.

[7] I would argue, for instance, that Feser’s (2019, pp. 20-27) arguments for hylemorphism—if successful—only establish that there is such a thing as prime matter and substantial form. Feser argues, for instance, that the determinability, changeability, particularity, diversity, and imperfection exhibited by material substances is (or must be) explained by prime matter, whereas the determinacy, unchangeability or permanence, universality, unity, and perfection exhibited by material substances is (or must be) explained by substantial form. But this is perfectly compatible with there being other facts about material substances that something else (i.e., something other than prime matter and substantial form) explains. The hylemorphically-inclined existential inertialist may very well say that one such fact is precisely inertial persistence.

[8] By ‘feature’, I don’t mean to commit to the view that an inertial tendency is some kind of positive ontological item existing in or exemplified by the object in question. By ‘manifest a further feature’, then, we just mean ‘behave in a way that neither of its parts on their own behave’.

[9] How we spell out the metaphysics here may again come down to the metaphysical accounts of EIT articulated earlier.

[10] Once more, I’ve discussed this point at length elsewhere.

[11] Two notes. First: For simplicity, I here assume that time is composed of smallest units termed moments. Nothing much hangs on this, though; I could explicate each explanans in a continuous-time-friendly manner mutatis mutandis. My more general point remains unaffected. Second: These are not the only explanantia under no-change and transtemporal accounts. I choose only one explanans from each of these two accounts for ease of exposition.

[12] There are criticisms of something like existential inertia in Kvanvig and McCann (1988). I have not considered them in this post for several reasons, foremost among them being that I’m not convinced that they’re actually criticisms of existential inertia as such. Kvanvig and McCann are explicit that they only criticize the thesis that there is some positive ontological item within a temporal substance—a self-sustaining capacity—whose operation ensures or explains or otherwise maintains the persistent existence of the substance. But few if any of the metaphysical accounts of Section 5 require positing a self-sustaining capacity of this sort, and it is certainly not inherent to or required by EIT as such. (Beaudoin 2007 also addresses Kvanvig and McCan’s points.) I have also not addressed Feser’s (2011) claim that (at least some of) the five ways themselves represent arguments against EIT. Here are some reasons for this. First: I have already addressed the first way elsewhere, and in the next sub-section I’ll be considering the Aristotelian proof’s demand for a sustaining cause. I will also address what Feser says on behalf of the demand for sustaining causes in his essence-existence or De Ente-esque argument (i.e., his Thomistic proof) as well as what Feser says on behalf of the demand for sustaining causes based on contingency (i.e., his Rationalist proof). What I say therein applies mutatis mutandis to Feser’s claim that the first three ways constitute arguments against EIT, since Feser gives the second and third ways essence-existence and contingency interpretations (respectively). And I don’t find the fourth or fifth ways plausible. Second: I find Feser’s claim deeply implausible, given that the renditions in question of the five ways have premises that—by my lights—simply assume, rather than justify, the falsity of EIT.

[13] What’s more, we need not even appeal to past states to explain present ones. I gave a variety of metaphysical accounts of EIT in Section 5 that do not require this.

[14] For instance: that something “did exist simpliciter at one time… is enough for it to bear a causal relation to me” (2019). To our minds, it also seems deeply plausible upon reflection. For instance: plausibly, the immediately temporally prior state, action, and/or obtaining of something can explain the immediately temporally posterior obtaining of something else.

[15] If time is continuous, we can let the temporal state immediately prior to t refer to some suitably small finite interval of time—perhaps infinitesimally small—with t as its later-than bound.

[16] In short, detractors of the Aristotelian proof don’t need to positively justify or establish that what happened prior to t is sufficient to explain the water’s existence at t. They only need to point out that the Aristotelian proof fails to justify why they aren’t sufficient.

[17] Here’s a way to appreciate why some philosophers think necessity can explain. Suppose the law of non-contradiction is true. Why? What explains that? Perhaps the explanation is that reality is consistent. But why is reality consistent? One of the only plausible answers that comes to mind is that ‘it just must be that way!’. By my lights, playing the ‘why?’ game—that is, repeatedly asking why in a chain of explanations—will ultimately bottom out in ‘well, things just have to be that way—end of story.’ It should be noted, though, that my rejoinder to the Neo-Platonic proof does not rely on showing or justifying why necessity can explain. That would be a rebutting defeater. But I am offering an undercutting defeater. In other words, I am arguing that (i) nothing in Feser’s Neo-Platonic proof rules out the legitimacy of this kind of explanation (in principle), and that (ii) ruling it out is required in order for the Neo-Platonic proof to succeed.

[18] Which are, again, parts according to the classical theistic understanding of parthood.

[19] And this kind of internal explanation of unity, moreover, is not restricted to neo-classical theists. Ingthorsson (2021, ch. 6) develops an interactive view of causality informed by contemporary science and the metaphysics of causation. Ingthorsson’s explanation of the unity and existence of composite objects is of particular relevance to the Neo-Platonic proof. He provides a number of reasons for thinking that the unity of composite material objects is explained internally in terms of the causal, glue-like interactions among parts. In contradistinction to NPCP’s requirement of an extrinsic source or principle that causally sustains the unity, Ingthorsson cites internal factors to do the explanatory work.

[20] Or perhaps a combination of these. We need not consider this here, since it is not pertinent to our criticism in the main text.

[21] Note that these suggestions are not (necessarily) suggestions that a proper part of something is the efficient (sustaining) cause of the whole of which it is a component. Rather, I’m talking about one part of something explaining why it is conjoined or unified with the other parts of the thing. At least prima facie, this seems like it could provide an explanation of a composite thing’s existence without adducing an (external) efficient cause of the existence of the composite thing. Prima facie, we could simply appeal to the explanation of the unity or conjoined-ness of the parts—in conjunction with the principle that a whole exists if its parts are unified or conjoined in the relevant way—in order to explain the existence of the composite object. This does not adduce an outside/external sustaining efficient cause of the very being of the whole composite object. (And this is precisely the kind of cause that Feser tries to infer in his argument, since the classical theistic God would clearly be an external/outside sustaining efficient cause of composite objects.)

[22] Ingthorsson (2021; Forthcoming) gives a causal interaction account of both constitution and persistence, arguing that the continued unity (and, hence, existence) of composite temporal concrete objects is explained internally by the continuous and dynamic glue-like interactions of their parts. Thus, Ingthorsson writes:

I think there are very obvious causal aspects to the constitution and continued existence of compound entities… [C]onsider the explanations given in secondary education physics and chemistry of the physical bonds that hold compound objects together, say, how elementary particles constitute an atom through continuous interaction. Physics, chemistry, and biology simply do not depict the entities they study as things that passively continue to exist as long as nothing destroys them; they continuously and actively preserve themselves through the interaction of their constituents. (2021, p. 99)

Ingthorsson’s account of the persistence of composite material objects is inertialist insofar as their persistence is not explained by some non-part concrete object that sustains them in being. Importantly, nothing Feser or Vallicella say constitute an objection to Ingthorsson’s causal interactionist explanation of both the synchronic and diachronic unity of composite material objects.

[23] For all Feser has shown, the necessary foundation of, source of, or explanation for contingent beings might be necessarily existent (in the sense of not possibly absent from reality) but nevertheless dependent. (But wouldn’t this implicate us in a regress of dependent necessary beings, which must ultimately terminate in a first, independent necessary being? No—at least, not if the ‘dependence’ in question is mere linear or per accidens dependence. For Feser nowhere argues that regresses of linear or per accidens dependence must have a first member. Rather, he only argues that hierarchical or per se chains of dependence must have a primary or first member. But the necessary dependent thing in question might—for all Feser has shown, that is—be dependent merely in a per accidens chain.) Note, though, that this first response I’ve given to Feser’s line of reasoning isn’t my main response. My main response is my second response, which is in the main text.

[24] If you don’t think there is a single creaturely way of being, the argument could equally apply to however-many-creaturely-ways-of-being-there-are plus the divine way of being. The argument for pluralism goes through either way, since the problem of truly universal quantification will arise either way.

[25] This formulation seems to encapsulate the first of the famous 24 Thomistic theses: “Potency and Act so divide being that whatsoever exists either is a Pure Act, or is necessarily composed of Potency and Act, as to its primordial and intrinsic principles” (Lumbreras 2016, p. 6).

[26] I follow Merricks in assuming the following: “monists take the existential quantifier, ∃, to capture (what they say is) the one and only way of being… That is, I shall assume that monists take the existential quantifier to range over all and only those entities that enjoy (what they say is) the one and only way of being. And I shall assume that pluralists take various existential-like quantifiers—∃1, ∃2, etc.—to capture (what they say are) the various ways of being” (2019, p. 593). Furthermore, as Merricks points out on the same page, the anti-pluralist argument does not depend on such quantifiers; for the arguments and debates at hand are not symbolic but fundamentally metaphysical.

[27] More precisely, they might think there are other ways of being (that are not a-existence or p-existence). They might, for instance, think there is a divine way of being (d-existence, say) and creaturely way of being (c-existence, say). But the sentence in the main text is fine as is, since the main point is that act-potency pluralists are, without generic existence, left simply with various specific existential-like quantifiers (∃a and ∃p and, as the case may be, ∃d, ∃c, and so on).

[28] Cf. Merricks (2019, pp. 596-598) for more on this commitment of pluralism.

[29] As Merricks (2019, p. 594) points out, several authors seem to agree that Turner’s approach (or something exactly like it in relevant respects) is the pluralist’s best hope for stating their view. See, e.g., Loux (2012) and van Inwagen (2014).

[30] To avoid expressing, by the (simple, standard) existentially quantified sentence ‘∃x(x=x)’, that something generically exists, act-potency pluralists might be tempted to read it as ‘something a-exists or p-exists’. But this won’t work, for the reasons articulated by Merricks (2019, pp. 600-601). In short, if ‘∃’ expresses the disjunction ‘∃a or ∃p’, then ‘~∃x(~Fx)’ expresses ‘~(∃ax(~F) or ∃px(~F))’ , which is equivalent to ‘~∃ax(~F) and ~∃px(~F)’, which in turn—given interdefinability—is equivalent to ‘∀ax(Fx) and ∀px(Fx)’. And since ‘~∃x(~Fx)’ is equivalent to ‘∀x(Fx)’, it would follow that ‘∀’ would express ‘∀a and ∀p’. But we have already seen that this renders the universal quantifier not truly universal, and hence we would land right back in the problem for act-potency pluralism.


7.14 Nemes and Kerr on the De Ente Argument

The final argument against EIT I’ll consider in this post derives from Aquinas’s De Ente Argument for God’s existence. I’ll focus in particular on Nemes’ and Kerr’s recent formulation and defense. Nemes spells it out as follows:

Pat Metheny… is an existing human being. He therefore is a substance that possesses an accidental esse [i.e., existence] of its own in composition with the essence of a human, viz. rational humanity. According to the ‘causal principle’ of Kerr and Aquinas, he must possess this proper actus essendi [i.e., act of existence] either (a) in virtue of the principles of his nature qua human being, i.e. his form and matter, or else (b) in virtue of receiving it from something outside of him. But his possession of being cannot be caused by the principles of his nature, e.g. his forma substantialis [i.e., substantial form] (a). The possession of esse is the absolutely prior condition of the possibility of the causal and explanatory efficacy of these principles of nature. The nature must first exist if it is to do or cause anything at all. Esse is prior to the substance, which means that it is prior to the composite of forma substantialis and materia prima [i.e., prime matter]. But the concern here is to understand how it is that Pat Metheny possesses existence in the first place. It follows therefore that he must receive his existence from something else (b) in which there is no distinction between essence and existence. This would be “something” which simply is, entirely in virtue of itself: esse tantum or pure being. (Nemes Forthcoming, p. 7)

Nemes is here expressing a causal principle found in Aquinas, which Kerr (2015, p. 93) translates as follows:

Whatever belongs to a thing is either caused by the principles of its nature (as the capacity for laughter in man) or comes to it from an extrinsic principle (as light in the air from the influence of the sun). (De Ente, Cap. 4, p. 377: 127-3)

There is much to say in response. First, the argument as Nemes articulates it requires whole swathes of highly contentious metaphysical commitments the existential inertialist need not accept. The argument, for instance, requires a constituent ontology on which features of a thing compose it (and, in particular, on which the metaphysical principle of esse (existence) is a constituent part of the substance). It also requires (i) realism about essences, (ii) realism about esse or existence, and (iii) a view of existence as a first-order property had primarily by substances (rather than (say) a second-order property (i.e., a property of properties) or not a property at all). But the inertialist isn’t (and needn’t be) beholden to these metaphysical commitments.

Second, Nemes’ argument (and, by extension, the De Ente argument) requires pluralism about being. For suppose—as monism would have it—that there aren’t different ways to exist or modes of being. In that case, God enjoys the same generic existence enjoyed by everything else. But—per DDS—God is identical to everything he has, and hence God is identical to his existence. But in that case, God is identical to generic existence—the existence that you and everything else shares (under monism). But your existence, as Nemes explicitly says, is a constituent part of you. God would therefore be a constituent part of you, which contradicts classical theism. So, the De Ente argument (conjoined with DDS, which is plausibly an entailment of the De Ente argument’s conclusion) requires pluralism about being. But we saw in earlier that we should reject pluralism, and so we should reject the De Ente argument.[1]

Third, consider again the causal principle underlying the argument: if x is F, then either (a) x’s being F is caused by the principles of x’s intrinsic nature, or (b) x’s being F is caused from without by y—that is, x’s being F is caused by some y such that y is extrinsic to (i.e., outside of, external to, disjoint from) x.[2] For starters, this principle—as currently stated—is compatible with x’s inertially persisting provided that x was caused to begin to exist. For if x was caused to begin to exist, then x’s existing is caused from without. The cause, though, is merely an originating (rather than continuously sustaining) cause. Second, I simply reject this principle. I think, instead, that if x is F, then x’s being F is explained—either by the intrinsic nature of x, or by an extrinsic cause, or by some non-causal explanation.[3] Not all explanations adduce causes, and my explanatory version of the principle is equally well-supported by both empirical and a priori considerations as the causal version. Indeed, we saw in Section 5 that there’s a panoply of explanations of why an inertially persistent object (x) exists at a given non-first moment of its life (F) that do not adduce extrinsic efficient sustaining causes.

The first of these worries is not an issue, though, for the principle can be modified as follows: if x is F at time t, then either (a) at t, x’s being F is caused by the principles of x’s intrinsic nature, or (b) at t, x’s being F is (concurrently) caused from without by y. But as with the unmodified principle, this modification succumbs to the second worry articulated above. There need not be an extrinsic efficient sustaining cause of x’s being F at t (when F isn’t caused by x’s nature); there need only be an explanation of x’s being F at t.[4] And as we saw in Section 5, explanations abound for the inertialist.

Kerr (2015, pp. 100-105) considers whether (i) there could be any uncaused non-intrinsic properties of a thing, and, more specifically, (ii) whether existence could be uncaused in one or more essence-existence composites. In order to complete my third reply to Nemes’ argument, then, an evaluation of Kerr’s reasoning is in order. Kerr writes:

To begin with (i), that there are no uncaused non-intrinsic properties, it must be borne in mind that Aquinas applies his causal principle so as to account for a fact requiring explanation: that is to say, if one can ask and answer the question of why a thing possesses a certain property, then one can offer a causal explanation for the possession of that property. (Ibid, p. 100)

I have one note and two responses. The note is that in appraising Kerr’s case, I will grant the legitimacy of constituent ontology, realism about essences, realism about existence, and so on. Now let’s consider my three responses.

First, if Aquinas applies his principle in order to account for facts needing explanation, then if we have principled reasons to think that there is some fact that doesn’t need an explanation, Aquinas’s principle, it seems, will be inapplicable. But why can’t the fact that a specific essence-existence composite object exists be one such fact not needing an explanation?

In fact, this is precisely what non-classical theists hold. Under both classical and non-classical models of God, the unique foundation of reality is a metaphysically necessary, unlimited, axiologically supreme, perfect being upon which everything else depends. But so long as the non-classical theist has principled reason to think that a perfect being exists, they thereby have a principled reason to think that this perfect being’s existence is uniquely unexplained. And this is true regardless of whether this perfect being is such that its essence is numerically distinct from its existence, since God’s status as uniquely unexplained simply falls out of his perfect and unlimited nature. It is intuitively obvious that perfection—axiological supremacy and unlimited value—precludes dependence on something else. In that case, there is independent reason to think that a perfect being—even if (under non-classical theism) the being’s essence and existence are distinct—is uniquely unexplained (and, indeed, unexplainable).

According to these non-classical theistic views, only finite, limited, imperfect, contingent things cry out for further explanation. But this cry is entirely silenced when we consider an infinite, unlimited, perfect, necessary being. There is no mystery as to why such a being is uniquely unexplainable, and there are principled, independent reasons for thinking as much—reasons entirely separate from considerations about the essence-existence distinction. Aquinas’s principle, then, is impotent to establish that an infinite, unlimited, perfect, necessary being requires a cause by dint of its being an essence-existence composite. For, again, it is clear that such a being couldn’t have a further explanation. And as Kerr points out, Aquinas applies his principle precisely to account for those facts that do require further explanation.[5]

Second, the principle <if one can ask and answer the question of why a thing possesses a property, then one causally explain the possession of that property> seems straightforwardly false. Suppose Stephen is hanging out with two friends, Cameron and Joe. Suppose Stephen has 23 cookies and wants to give each of his two friends the same whole number of cookies without any left over. To his dismay, he finds that he can’t do this. What explains the fact that Stephen’s desire has the property being frustrated? The explanation is in terms of mathematical constraints: 23 cannot be divided by 2 to yield a whole number. Importantly, though, this is not a causal explanation. The relevant non-divisibility of 23 isn’t causing Stephen’s desire to be frustrated. Nevertheless, it explains it. And, indeed, this is the only explanation of why the desire is frustrated. (We are not asking why Stephen exists, or why his desire exists, or whatever; we are asking why the relevant desire is frustrated given that he exists, has the relevant desire, and so on.) The principle Kerr adduces, then, is false. For we can ask and answer the question of why Stephen’s desire possesses a property, and yet we can only non-causally explain the possession of that property.

Kerr does consider the prospect of uncaused properties. Here’s what he writes:

But if such properties are uncaused, there is no explanation for why they exist in the thing, that is, there is no explanation of why they are there. Granted that it is a fact they are, such a fact is without explanation. . . . If the objector grants that there is an intelligible framework within which beings can be analysed and accounted for, then such uncaused non-intrinsic properties must be primitive. (Ibid, p. 101)

There are two problems with this. First, these are mere assertions. Kerr asserts but does not justify the claim that uncaused properties are unexplained. That alone suffices as a response. (Remember, the onus of justification in this dialectical context is on the proponent of the De Ente argument against EIT. For they are the ones leveling an argument aiming to disprove EIT, and hence the onus is not on the detractor of the argument to positively show why or how some uncaused property is nevertheless explained.)

Second, Kerr’s claims here are simply false. It is clear that the mathematical constraints that explain the frustration of Stephen’s desire do not cause Stephen’s desire to be frustrated. Since the mathematical constraints alone explain why the desire is frustrated (given, of course, Stephen exists, that his desire exists, etc.), this seems like a case of an uncaused-but-explained property possession. More fundamentally, it is simply false that <if S’s existence at a non-first moment of S’s life has no (concurrent sustaining) cause, then S’s existence at a non-first moment of S’s life has no explanation>. We have already seen in Section 5 that there are whole swathes of inertialist-friendly explanations of persistence on which this conditional has a true antecedent but false consequent.

All of this, moreover, is highly germane to the De Ente argument against EIT. As Kerr rightly points out,

The upshot of the objection that there could be non-intrinsic, uncaused properties is that if correct, it would entail that one cannot move from the distinction of essence and esse to the caused character of esse in essence-esse composites, and because setting up a causal series in the line of esse is essential to the argumentation that Aquinas is making, the possibility that esse could be an uncaused yet distinct (non-intrinsic) property of a thing would undermine Aquinas’s argument from the outset. (Ibid)

If my above responses work—and if the metaphysical accounts of EIT from Section 5 are defensible—then the De Ente objection to EIT is undermined. In fact, as should be clear by now, I actually don’t need—in the present dialectical context—to positively show that the metaphysical accounts of EIT are defensible. I only need the weaker claim that neither the De Ente argument nor what its proponents (e.g., Nemes, Kerr) say on its behalf gives those who accept (or are neutral on) one or more of the metaphysical accounts of EIT sufficient reason to abandon their position and instead affirm that all such accounts fail. And this weaker claim is, I think, clearly true. Thus, the De Ente objection to EIT is undermined.

Recall that Kerr is interested in whether (i) there could be any uncaused non-intrinsic properties of a thing, and, more specifically, (ii) whether existence could be uncaused in one or more essence-existence composites. Having argued that his case against (i) fails, I will now turn to his case against (ii). He writes:

Esse stands to essence as act does to potency, such that esse is what actualises essence and makes it exist. No essence would thus exist without esse. Given that essence would not exist without esse, why does some existing essence have esse in the first place? Even though esse is primitive and there is nothing more fundamental than esse, its being composed with some essence is not primitive; and since the essence with which it is composed does not possess such esse essentially, it possesses it from without. (Ibid, p. 102)

I have two replies. First, in the case of neo-classical theism, the very question at issue is whether there is some essence such that its composition with esse is primitive (i.e., not dependent on anything else). The very question at issue, in other words, is whether there is (or can be) an essence-existence composite (viz. the God of neo-classical theism) whose existence is not derived from something else. It is question-begging, then, to simply assert that esse’s being composed with some essence is not primitive.[6] Second, even if the composition of esse and essence is not primitive (i.e., there is some further explanation for why they are composed—including an explanation for why they’re composed at any given moment of an object’s existence), EIT’s falsity doesn’t follow. For—as we’ve seen time and again—there are whole swathes of inertialist-friendly explanations of the esse of essence-existence composites (at non-first moments of their lives).

         Kerr continues:

So, why does it thus possess esse? What is the cause of that essence’s esse? These are not unreasonable requests, and so the burden of proof is on the objector to show how that in which essence and esse are distinct could have esse and thus actually exist without its esse being caused. (Ibid)

But, first, this response illicitly shifts the onus of justification. It is Kerr who is offering a positive argument in the present dialectical context. The onus is therefore not on the detractor to positively show how there could be an uncaused essence-existence composite; the onus, instead, is on Kerr to show how there couldn’t be an uncaused essence-existence composite. Without showing this, a claim key to the De Ente argument is simply unjustified. Second, even ignoring the illicit burden-shifting maneuver, I’ve developed a variety of metaphysical accounts of EIT, and such accounts—granting Kerr constituent ontology, realism about essences and esse, and so on—represent workable accounts on which at least some essence-existence composites continue to exist without continuously concurrent sustenance from without. They therefore directly answer Kerr’s challenge to spell out accounts on which an essence-existence composite can exist at a non-first moment of its life without a (concurrent sustaining) cause.

         Finally, Kerr writes:

Given the above, I submit that (ii) [the claim that existence could be uncaused in an essence-existence composite] can be rejected because insofar as esse is related to essence as act to potency, yet no essence-esse composite need ever exist, there is a cause for the esse that the essence enjoys. (Ibid)

I have two responses. First, Kerr simply asserts but does not justify that no essence-existence composite need ever exist. Neo-classical theists, panentheists, and non-theists who accept a necessarily existent foundation of reality will all simply reject that no essence-existence composite need ever exist. More generally, Kerr’s claim is flatly rejected by those who accept (e.g.) the objectual necessity account of EIT. Nothing Kerr says here gives any reason to rule out such views.[7] Second, even if a given essence-existence composite need not exist (or need not exist at non-first moments of its life), all that follows from this (by my lights) is that there needs to be an explanation of its existence (or existence at non-first moments of its life). (At the very least, Kerr has failed to justify why a cause is needed rather than a mere explanation.) But as we’ve seen, there are lots of explanations that make no reference to (sustaining) causes.

I conclude, then, that my third reply to Nemes remains forceful. My fourth and final reply is a Moorean one. The De Ente argument, if successful, entails the Big Four: DDS, timelessness, immutability, and impassibility. But even if one cannot pinpoint precisely where an argument goes wrong, one is well within one’s rights in denying the conclusion and inferring the disjunction of the negations of the premises so long as one has sufficiently strong independent reasons to think the conclusion is false. But for many non-classical theists and non-theists, this condition is met with respect to the Big Four. I myself have developed and defended elsewhere develop a variety of formidable challenges to classical theism from God’s changing knowledge, various considerations pertaining to modal collapse, abstracta, and so on.[8]


[1] See earlier in this post for more on the relationship between pluralism and Feser’s own rendition of a broadly De Ente-esque argument he terms the ‘Thomistic proof.’

[2] This is essentially how Kerr (2015, p. 95) puts it. He writes: “if something, x, possesses some property, F, then x possesses F either as a result of the principles of its own intrinsic nature, its x-ness, or as a result of some extrinsic principle(s), y.” Kerr goes on to explain that the ‘as a result of’ signifies causality.

[3] A complication arises when we consider that, plausibly, all global explanatory hypotheses—including that purportedly delivered by the De Ente argument—have one or more primitive (basic, fundamental, foundational, not-further-explained) entities and (perhaps) properties (ese: predicates). For instance, God (x) is good (F), but God’s being good, under classical theism, is caused neither by God’s intrinsic nature nor an extrinsic principle. We can set this complication aside, though, since (i) I am focused, in the main text, on non-fundamental things and (ii) the principle can be modified to say something like ‘if x is F, then (i) if we have sufficient, principled reason to think x’s being F is a fundamental fact, then x’s being F is unexplained, and (ii) if we don’t have sufficient, principled reason to think x’s being F is a fundamental fact, then x’s being F is explained.’ Alternatively, we could restrict the explanatory principle to (say) contingent cases of x’s being F, or cases where x is a limited being (i.e., not a perfect or axiologically supreme being), or whatever.

[4] At the very least, nothing Nemes says gives the inertialist any reason to abandon this explanatory principle, which is what he would need to do for his argument against EIT to have teeth.

[5] The same reasoning applies to non-theists who have principled reasons for thinking that some layer of (say) natural reality (say, a necessarily existent foundational quantum field, or the universal wavefunction, or whatever) couldn’t be explained in principle. And such independent reasons aren’t too difficult to come by: suppose one is convinced by Pruss and Rasmussen (2018) that there is at least one foundational necessarily existent concrete object, and suppose further that one has strong reasons to think theism is false (deriving from, say, the problem of evil, or divine hiddenness, or religious diversity and confusion, or paradoxes of omniscience or perfect rationality, or whatever). Then, one has independent reason to think that there is some non-theistic necessary being that is unexplainable in principle. And as I point out in the main text, this will debar the need for a cause regardless of whether such a being is an essence-existence composite.

[6] I assume—quite innocuously, by my lights—that by ‘some’ Kerr means ‘any’. At least, this is what he would need to mean for his case to go through (applying, as he wants it to, to all essence-existence composites.)

[7] Note that the proposal here is not that the necessary foundational essence-existence composite is such that its existence is caused by the principles of its essence. The proposal, instead, is that the essence necessarily exists and that the conjoinedness of its essence and existence is a primitive metaphysical necessity. And as we saw earlier, there are (or can be) principled reasons for affirming this primitive element.

[8] Cf. Schmid and Mullins (Forthcoming) and Schmid (Forthcoming-a (my IJPR article on modal collapse); Forthcoming-b (my EJPR article on Neo-Platonic proof); this video and the forthcoming Parts 2 and 3).

8 Resources

Below you’ll find some very helpful resources on EIT from both sides of the debate.

8.1 Articles

Some works that don’t explicitly discuss EIT but nevertheless implicitly discuss EIT or deeply related issues

Rundle (2004). Why there is something rather than nothing. Oxford University Press.

Millor (PhD thesis)

Kvanvig and McCann (1988 and 1991)

  • Kvanvig, J.L. & McCann, H.J. (1988). Divine Conservation and the Persistence of the World. In T.V. Morris (Ed.) Divine and Human Action: Essays in the Metaphysics of Theism. Cornell University Press.
  • McCann, Hugh J. & Kvanvig, Jonathan L. (1991) “The Occasionalist Proselytizer: A Modified Catechism,” in Philosophy of Religion, Philosophical Perspectives, vol. 5, Philosophy of Religion, ed. James E. Tomberlin (Atascadero, CA: Ridgeview Pub.), 587-615.

Kerr (2015) and various papers published in response

  • Kerr, G. 2015. Aquinas’s Way to God: The Proof in the De Ente et Essentia. New York: Oxford University Press.
  • Critical and constructive commentary and engagement of Kerr (2015) can be found in an entire issue—freely available in PDF format—in the journal Roczniki Filozoficzne here: https://ojs.tnkul.pl/index.php/rf/issue/view/175

Mackie (1974). The Cement of the Universe: A Study of Causation. Oxford: Clarendon Press.

Works that explicitly discuss EIT

Beaudoin (2007)

  • Beaudoin, J. (2007). The world’s continuance: Divine conservation or existential inertia? International Journal for Philosophy of Religion, 61(2), 83-98.

Feser (2011)

  • Feser, E. (2011). Existential Inertia and the Five Ways. American Catholic Philosophical Quarterly, 85(2), 237-267.

Oderberg (2014)

  • Oderberg, D. (2014). Being and Goodness. American Philosophical Quarterly, 51(4), 345-356.

Feser (2017)

  • Feser, E. (2017). Five Proofs of the Existence of God. San Francisco: Ignatius Press.

Benocci (2018)

Audi (2019)

  • Audi, P. (2019). Existential Inertia. Philosophic Exchange, 48(1), 1-26.

McNabb and DeVito (2020)

  • McNabb, T. & DeVito, M. (2020). Has Oppy Done Away with the Aristotelian Proof? The Heythrop Journal, 61(5), 723-731.

Schmid (2021)

  • Schmid, J.C. (2021). Existential inertia and the Aristotelian proof. International Journal for Philosophy of Religion, 89: 201-220.

Ingthorsson (2021, ch. 6)

  • Ingthorsson, R. (2021). A Powerful Particulars View of Causation. Routledge.

Oppy (2021)

  • Oppy, G. (2021). On Stage One of Feser’s ‘Aristotelian Proof’. Religious Studies, 57(3): 491–502.

Feser (2021)

  • Feser, E. (2021). Oppy on Thomistic cosmological arguments. Religious Studies, 57: 503–522.

Hsiao and Sanders (Forthcoming)

  • Hsiao, T. & Sanders, G. (Forthcoming). The Contingency Argument in Plain Language. The Heythrop Journal.

Schmid (Forthcoming)

  • Schmid, J.C. (Forthcoming). Stage One of the Aristotelian Proof: A Critical Appraisal. Sophia.

Schmid (Forthcoming)

  • Schmid, J.C. (Forthcoming). Simply Unsuccessful: The Neo-Platonic Proof of God’s Existence. European Journal for Philosophy of Religion.

8.2 YouTube videos

I discuss existential inertia in at least five videos: (a) A User’s Guide, (b) Response videos to Intellectual Conservatism (Part 1 and Part 2), (c) the end of my discussion with Oppy on the nature and purpose of arguments, (d) Arguments for Classical Theism | Part 1/2.

8.3 Blog posts

(1) I respond to Feser regarding existential inertia and the Aristotelian proof here.

(2) I respond to Feser regarding existential inertia here.

(3) I have written on existential inertia (including some metaphysical accounts thereof) in my response to Hsiao and Sanders here.

(4) I respond to Nemes regarding existential inertia here.

(5) I discuss existential inertia in my response to RM and HoH here.

(6) I respond to Thomistic Disputations on existential inertia here.

(7) I cover existential inertia in sections of this post, too.

Author: Joe

7 Comments

  1. I’m keen on reading this (for a forthcoming paper on an Avicennian argument against EIT that I’m developing and defending). The gist of it is here on my blog: https://kimiyagard.wordpress.com/2020/07/06/a-sharqi-argument-against-existential-inertia-ei/

    Let me know what you think. pax.

    • Wonderful to see you here!! Thanks for your lovely comment!

      I actually address the argument you develop in your post in my post. I address it in numerous places, but I think you be especially interested in Section 5 of this post [wherein I develop a host of explanations of contingent objects’ existence at non-first moments of their existence that don’t adduce a sustaining cause] as well as Section 7 [wherein I address many criticisms of EIT, many of which make essentially same point you made about things’ *remaining* contingent even at non-first moments of their existence. Feser, for instance, makes this point in his Rationalist proof. I address it in Section 7.6, Section 7.11, and other sections too.

      Hope this helps! 🙂

  2. Pingback:Comments on Feser on Oppy on Thomistic Cosmological Arguments | Majesty of Reason

  3. Really interesting post! As a Classical Theist, I’ll make a few comments in response to the so-called theoretical virtues of EIT against CTST.

    1. You say that EIT is superior to CTST viz-a-viz explaining the fact that we do not observe existential sustaining causes in the world of experience. However, this is not correct. On Classical Theism, the power to create or to conserve something in existence is such that no creature can possibly possess or exercise it (a point emphasised by Aquinas.). Naturally, then, observation of sustaining causes is metaphysically impossible *even on* CTST (since the only candidate, God, is non-observable). Given this, I think it is false that EIT has any explanatory superiority to CTST w.r.t the dearth of observational evidence of sustaining causes.

    2. In a similar fashion, you say that our observation of objects ceasing to exist always and only due to a positive cause is superiorly explained by EIT than by CTST. Now the problem with this is that, strictly speaking, we have never observed (nor could we observe) the occurrence of any *positive* causal destruction. The reason is that the very notion is incoherent. Positive causal destruction implies that the non-existence of a concrete object O is the effect of — brought about by — some cause. However, ‘non-existence’ cannot be an effect. Effects must exist in order to stand in the causal relations they do. It follows that objects cannot cease to exist or corrupt through being caused to do so. But then, what does EIT predict over CTST?

    3. It seems to me that your Bayesian argument proves too much. On *any* brand of Theism (e.g. Neo-Theism, Panentheism, even Deism, ect.), we could speak of the probability that God might annihilate the world provided He has sufficient power to exercise or withdraw. The implication of your argument is that this is distinctively a problem for *classical* theism (namely, CTST) in particular. But, if it is a problem, it is one for all theistic views and indeed all Abrahamic religions. For on any of those views, it is possible for God to annihilate the universe and hence open to the Bayesian argument you propose here.

    4. As evidenced by Oderberg’s position, the Classical Theist can take a view according to which [contingent] things have a natural tendency toward existing, which is nevertheless ultimately dependent upon God. On this view, it seems like the Bayesian argument does not go through. For while God could in principle annihilate some objects, that would count as a miracle (a suspension of the natural order) which, by definition, is irregular. Hence, like EIT, it would be far more likely for us to observe things persisting in existence.

    Overall, then, I do not think that EIT is theoretically any more virtuous than CTST.

    • Thanks for the comment! 🙂 <3

      You say:

      “1. You say that EIT is superior to CTST viz-a-viz explaining the fact that we do not observe existential sustaining causes in the world of experience. However, this is not correct. On Classical Theism, the power to create or to conserve something in existence is such that no creature can possibly possess or exercise it (a point emphasised by Aquinas.). Naturally, then, observation of sustaining causes is metaphysically impossible *even on* CTST (since the only candidate, God, is non-observable). Given this, I think it is false that EIT has any explanatory superiority to CTST w.r.t the dearth of observational evidence of sustaining causes.”

      I actually have a footnote in my book manuscript under review addressing precisely this objection. [I call it the ‘incommunicability thesis’.] Here’s what I say:

      “One might object that CTST likewise predicts a dearth of observational evidence for sustaining efficient causes, since—so the objection goes—the power to sustain something in existence could only be had and exercised by God. Such a power is an incommunicable characteristic of God (so the objection goes). In response, I note that this is not entailed by CTST as such. Instead, it is an auxiliary hypothesis that must be added to CTST. Thus, while one might add it to CTST to predict the data at hand, doing so lowers CTST’s probability. (Even if one disagrees and thinks that the incommunicability thesis here is entailed by CTST, we still have a dearth of observational evidence of instrumental sustaining efficient causes in chains of efficient sustaining causes. And surely creaturely instrumental sustenance is not debarred by CTST. My broader point, then, stands.)”

      And this is true — there’s nothing in CTST, as I’ve outlined it, that alone entails that ‘being a sustaining efficient cause’ is necessarily unique to God. Sure, you might add some auxiliary metaphysical theses to try and derive this. But then you’ve predicted the data only at the cost of lowering the probability of the hypothesis.

      It also seems to me — prima facie — that Aquinas allows that there may be *instrumental* causes in the chain of sustaining causes of existence. Now, I’m not trained in Aquinas interpretation, so we must keep this in mind. But prima facie, this seems to be what Aquinas is getting at when he says, in his second way, that “In the world of sense we find there is an order of efficient causes”. And the order of causes he talks about in his ways are subordinated per se. [Given the passage, it would be strange if Aquinas were talking about per accidens series here. He would have to be read as abruptly shifting gears to a per se order despite all appearances of continuing to speak about the same order he said was manifest to the senses.] And so here we seem to have Aquinas saying here that we *can* sense per se chains of efficient causes, albeit ones wherein the creaturely causes are instrumental. And this is all I need for my point to go through.

      You say:

      “2. In a similar fashion, you say that our observation of objects ceasing to exist always and only due to a positive cause is superiorly explained by EIT than by CTST. Now the problem with this is that, strictly speaking, we have never observed (nor could we observe) the occurrence of any *positive* causal destruction. The reason is that the very notion is incoherent. Positive causal destruction implies that the non-existence of a concrete object O is the effect of — brought about by — some cause. However, ‘non-existence’ cannot be an effect. Effects must exist in order to stand in the causal relations they do. It follows that objects cannot cease to exist or corrupt through being caused to do so. But then, what does EIT predict over CTST?”

      The effect isn’t non-existence. It is the *cessation* — the *going out of existence* — of the object in question. So, we have the event of (say) Jones being hit by a bus [cause], and the event of Jones dying [effect]. This doesn’t require non-existence to stand in a relation.

      You say:

      “3. It seems to me that your Bayesian argument proves too much. On *any* brand of Theism (e.g. Neo-Theism, Panentheism, even Deism, ect.), we could speak of the probability that God might annihilate the world provided He has sufficient power to exercise or withdraw.”

      This isn’t proving too much; it’s the correct result. This is true: all forms of theism face some evidential confirmation here. [And this just follows from Bayes theorem, of course.] So it’s not proving too much; it’s just showing that persistence evidence for a necessitarian view of persistence [i.e., it is necessary that something persists in the absence of destruction and sustenance] vis-a-vis non-necessitarian ones.

      You continue:

      “The implication of your argument is that this is distinctively a problem for *classical* theism (namely, CTST) in particular.”

      I do not recall saying or implying that this is distinctively a problem for *classical* theism. Yes, I’m focusing on CTST. But that doesn’t mean I think it’s *distinctively* a problem for classical theism; it just means I think it’s a problem for classical theism. Showing that there’s Bayesian disconfirmation of H1 vis-a-vis H2 doesn’t mean or imply that H2 is *uniquely* disfavored by the data in question.

      You continue:

      “But, if it is a problem, it is one for all theistic views and indeed all Abrahamic religions. For on any of those views, it is possible for God to annihilate the universe and hence open to the Bayesian argument you propose here.”

      Correct — it’s a problem for all such views.

      But — perhaps more fundamentally — classical theism *is* uniquely disadvantaged vis-a-vis non-classical views. For under classical theism, *nothing* about God varies across *any* possible world. And hence there is *nothing* about the cause/explanation across worlds that we can cite to explain *differences* in the effect [ceasings to exist]. And hence classical theism doesn’t seem to give us any probabilistic resources to expect one world *as opposed to* any other. This contrasts with non-classical views, since these views allow that there is something about the cause/explanation that *does* vary across worlds, and hence they *can* cite facts about the cause to explain *differences* in the effect [ceasings to exist] across worlds. And hence we have probabilistic resources to ground differential expectations as to why one world obtains rather than others.

      You say:

      “4. As evidenced by Oderberg’s position, the Classical Theist can take a view according to which [contingent] things have a natural tendency toward existing, which is nevertheless ultimately dependent upon God. On this view, it seems like the Bayesian argument does not go through.”

      On the contrary; for such a tendency is not entailed by CT, and hence it represents an auxiliary thesis that lowers the probability of the hypothesis [and hence you gain greater expectability of the data only at the expense of comparably lowering your prior].

      You continue:

      “For while God could in principle annihilate some objects, that would count as a miracle (a suspension of the natural order) which, by definition, is irregular. Hence, like EIT, it would be far more likely for us to observe things persisting in existence.”

      This is a fascinating response — thanks for it. Two notes. First, this accrues the probability cost mentioned above. Second, it still wouldn’t completely make up for the gap in expectability of the data, and hence n-EIT would still garner *some* evidential confirmation from the data. For by your own lights, the probability of cessation at any given moment — even in the absence of destruction — is still non-zero, and this is all we need to get at least *some* evidential confirmation for n-EIT vis-a-vis CTST. I will add, moreover, that even if the relevant interruption is miraculous, God can still freely interrupt this tendency whenever he wishes to brings things out of existence, and he can do this at any of the infinitely many points in something’s life [under continuous time] or the bajillions of points [under discrete time]. And there is no difference in God across any such world wherein God indeterministiclly causes them to cease at different points, and hence there seems to be nothing to ground differential probability assignments as to what we would expect to occur.

      You then say:

      “Overall, then, I do not think that EIT is theoretically any more virtuous than CTST.”

      This is a non-sequitur even granting everything else you said. You did not address the points about the massive simplicity advantages gained by EIT vis-a-vis CTST. And your four points only aimed to gain an explanatory/predictive parity between EIT and CTST, from what I can tell. And so even if all of your points succeeded [contrary to what I’ve argued], we have explanatory parity but parsimony superiority, which means EIT is theoretically more virtuous than CTST *even granting* everything you said.

    • I will also note that it’s the first day of classes for Purdue tomorrow, so I might not be able to continue this conversation further. You’re certainly welcome to reply, it’s just I’m going to be very, very low on time starting tomorrow. <3

    • Btw, I enjoyed your fourth objection so much that I actually added it to my book manuscript! Here’s what I added:

      Now, one might object that the proponent of CTST—following Oderberg—can adopt a view on which temporal objects have a natural tendency to persist, which is nevertheless ultimately dependent upon God. On this view, the objection continues, the Bayesian argument fails. For while God could in principle annihilate any given temporal object at any given time, that would count as a miracle (i.e., a suspension of the natural order)—which, by definition, is improbable. Hence, like n-EIT, it would be far more likely for us to observe things persisting in existence in the absence of destruction.

      I have four replies. First, adopting this view seems to do significant damage to Persistence Arguments for classical theism. For this view grants the central contention of (some versions of) tendency-disposition accounts of EIT—namely, that things by nature tend to persist. If we grant this central contention, then—plausibly—we have a readily available inertialist-friendly explanation of persistence. Why posit anything else to explain persistence if we already have—as this objection seems to grant—a workable tendency-disposition account? The objection, then, seems to undermine Persistence Arguments for classical theism.[Fn]

      Second, such a tendency is not entailed by CTST as such, and hence it represents an auxiliary thesis that lowers the probability of the hypothesis. Thus, greater expectation of the data is gained only at the expense of lowering the hypothesis’s probability.

      Third, even granting the objection’s central point about the miraculous (and hence improbable) nature of God withdrawing his sustenance, this won’t completely close the gap between the P(F|CTST) and P(F|n-EIT), and hence n-EIT still garners some evidential confirmation vis-à-vis CTST from F. For by the objection’s own lights, the probability of O’s cessation at any given moment—even in the absence of destruction—is still non-zero, and this is all we need to get at least some evidential confirmation for n-EIT vis-à-vis CTST. (Things only get worse for CTST when we consider that there are boatloads of such moments in O’s life.)

      Finally, the objection’s inference from “temporal objects have a natural tendency to persist” to “it would therefore be miraculous (and hence improbable) for God to remove his sustenance” is a non-sequitur. After all, many classical theists think things have a natural tendency to expire (in the sense that things are disposed to cease to exist when a certain manifestation condition (say, removal of divine sustenance) is met). And yet this clearly doesn’t imply that objects’ persistence is a miracle. Moreover, (i) many features of objects are “non-natural” in the sense that the object in question naturally has different, incompatible features, and yet (ii) this does not (typically) involve any miracle. Consider, for instance, that there is a sense in which humans are naturally two-legged.[Fn] But nearly all instances of zero- or one-leggedness in humans are not miracles. This point generalizes to any physical, genetic, or psychological defect or illness anywhere in the biological world. The objection, then, simply rests on a non-sequitur.

      [Fn] Two notes. First, this is not to say it undermines all or even most arguments for classical theism. Second, I don’t claim that a tendency to persist is incompatible with classical theism; I have only claimed that it plausibly undermines Persistence Arguments. For if we grant such a tendency, then there seems to be a workable tendency-disposition account of persistence that illuminates why things persist without appeal to external sustenance. (To be sure, classical theists may hold that appeal to external sustenance is needed for reasons other than Persistence Arguments. My point here, though, is restricted to Persistence Arguments.)

      [Fn] I’m here using “natural” in the sense that many classical theists use it (as in Feser (2014, pp. 258-261)). In this sense, “S is naturally F” means (roughly) that F-ness is proper to or characteristic of the kind of thing S is. It does not mean whatever belongs to that kind must—as a matter of metaphysical necessity—be F in order to fall within the kind.

Leave a Reply to Majesty Of Reason Cancel

Your email address will not be published. Required fields are marked *