Feser on Schmid on Existential Inertia: Vicious Circularity and the Metaphysics of EIT | Part 5

Feser has recently responded to my IJPR article. I will respond to his post in a series of blog posts. Check out Part 1, Part 2, Part 3, and Part 4. This post is Part 5, which deals with everything Feser says on the metaphysics of existential inertia. There are only two more installments of the series left: Part 6, which deals with everything Feser says in his section “Theoretical vices”, and Part 7, which addresses everything Feser says in his section ‘An argument against EIT’.

Feser writes: “Schmid next considers two possible ways of spelling out EIT.  The first account goes like this: Consider the water in our earlier example.  Its existence at some time t is sufficiently explained by (a) the state and existence of the water at an immediately preceding time t – 1 together with (b) the absence of anything acting to destroy the water.”

This first account of EIT is another aspect of the paper I would modify. When I wrote the paper, I wanted to leave this first account open between (what I have more recently termed) ‘transtemporal accounts’ and ‘no-change accounts’ (cf. those links for more detailed characterization). Both of these accounts cite at least one fact about past things to explain the present existence of the water. Naturally, then, I thought they could both fall under my first account in the IJPR paper, since they both — at least under one way to articulate them — appeal to the moment immediately before the present moment as part of the explanation of the present existence of an object. But I have now come to see that this was a mistake, since transtemporal and no-change accounts of EIT — while unified by such an appeal — are nevertheless fundamentally different kinds of explanations. The result is that my IJPR’s first account can be read in either transtemporal-account terms or no-change-account terms, and this makes things difficult (as all ambiguities do). So, let me tease out these two different ways of understanding the first IJPR account.

On a transtemporal understanding of the first IJPR account, some kind of transtemporal relation(s) to explain persistence. In the paper, I leave open the precise nature of the relevant transtemporal explanatory relation, allowing the relation to be either causal or non-causal. According to this understanding of the first IJPR account, then, temporal concrete object S’s persistence is explained by (i) the absence of sufficiently causally destructive factors operative on S, plus (ii) transtemporal explanatory relations (causal or otherwise) obtaining between the temporally successive states of S’s life (so to speak).

Now, for the sake of concreteness, let’s stipulate that the relation at play is causation. Moreover — again for the sake of concreteness — let’s stipulate that time is discrete and so composed of smallest units. Let’s call these smallest units of time moments. Thus, a transtemporal understanding of the first IJPR account says that S’s existence at moment m is explained by (i) the absence of sufficiently causally destructive factors operative on S from (m-1) to m [where (m-1) is the moment immediately prior to m], plus (ii) S’s state and/or existence at (m-1) causally producing S’s existence at m (and perhaps state at m). An explanation for S’s persistence simpliciter through a series of moments will come by way of the application of this explanatory schema to each non-first moment m of S’s existence in the series of moments.

That, then, is the transtemporal understanding of the first IJPR account that we’ll work with. Now let’s consider the no-change understanding of the first IJPR account.

In general, no-change accounts of EIT view persistence as an absence of change and take this fact to be central to their inertialist-friendly explanation of persistence. The no-change account we’ll consider here is one I’ve mentioned both in this series of responses to Feser and in Section 3 of my lengthier response to Feser. Here’s how it goes.

For S to fail to exist at m despite existing at (m-1) is for some change to occur.[Fn] 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 at (m-1), 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-1) to m — it simply follows that S exists at m. [Cf. Section 4.1 in the lengthier blog post for those interested in pursuing this line of thought even further.]

Here, we seem to have a perfectly respectable, perfectly legitimate explanation of S’s existence at m — and the explanation 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 [to wit, at (m-1)], (ii) nothing causally induced S’s cessation at (m-1) or m [i.e., nothing destroyed S from the immediately prior moment, (m-1), through m], and (iii) a change occurs only if some factor causally induces said change]. And so we do, indeed, have sufficient explanation for S’s existence at m. For me at least, the explanation certainly seems to remove mystery as to why/how S exists at m. The present explanation does, indeed, tell us how S exists at m.

————

[Fn] I addressed an objection to this argument [or, rather, explanatory schema] based on the claim that cessation isn’t a change in Section 4.1.2 of my lengthier blog post. Use the command F function to find it quickly — just search “4.1.2”.

————

To be sure, the disambiguation above — that between transtemporal and no-change understandings of the first IJPR account — is not found in the IJPR article. That’s why I said, at the beginning, that this is an aspect of the IJPR paper I would modify. Of course, Feser is not responsible (and nor will I take him to be, as we proceed in this post) for ignoring this disambiguation. But given that this disambiguation represents my current, actual views on the IJPR paper, I will proceed in my analysis of what Feser says with the disambiguation in mind. (It would be unreasonable, of course, for one to demand that I defend the non-disambiguated-version from the IJPR paper if I disagree with its being non-disambiguated in the manner it was!)

Let’s pick back up, then, with what Feser says: “Now, an objection that might be raised against existential inertia thus understood (and one I have raised in my exchanges with Graham Oppy and in my previous reply to Schmid) is that it is viciously circular.  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.”

I have responded to this circularity objection in Section 3 of my lengthier blog post, but I will also address it here.

First, neither my transtemporal understanding nor my no-change understanding of the first IJPR account entail that inertial persistence is a property or power. Both of the accounts could, for instance, be perfectly acceptable to philosophers of an anti-realist bent who think there are no such things as properties or powers. There is simply nothing in the articulations above that entail EI’s being a property or power.

Second, in order for Feser’s circularity argument to work, the two articulations must cite explanatory facts that presuppose the (explanatorily or ontologically) prior obtaining the relevant explanandum. But that is simply untrue, as I will now show.

In the case of the transtemporal understanding of the first IJPR account, we have:

Explanans 1: (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 (m-1) at which they exist causally produce their existence at the next moment, m, provided that no sufficiently causally destructive factors are operative.

Explanandum: S’s existence at m

In the case of the no-change understanding of the first IJPR account, we have:

Explanans 2: (i) S existed immediately before m — that is, 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) some kind of 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.

Explanandum: S’s existence at m

Now, I think it’s quite clear that neither Explanans 1 nor Explanans 2 presuppose the prior reality or obtaining of Explanandum. In other words, none of the explanatory facts adduced in Explanans 1 or Explanans 2 are dependent upon the fact cited in the Explanandum. And in that case, Feser’s allegation of viciously circular (explanatory) dependence has no teeth against the transtemporal and no-change understandings of the first IJPR account. It is simply false, of both of these understandings, that there is some property or power that both explains and is explained by some fact.

[Keep in mind, moreover, that there are many more inertialist-friendly explanations besides, and that these, too, do not fall prey to charges of vicious circularity. Cf. the various documents I linked in Section 2.2 of my lengthier blog post.]

So far I’ve offered two responses to the vicious circularity charge. First, neither of the understandings of the first IJPR account treat (or entail that) inertial persistence is a property or power. Second, the explanations proffered in neither of the understandings are viciously circular. Here’s a third response to the vicious circularity charge. Suppose — contrary to what I believe — that existential inertia were a property. This would only be problematic if we accepted the controversial thesis that properties ground character — that is, 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/etc. 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 [it’s not], Feser’s argument still won’t work.

One of my friends [Luiz] suggested to me a fourth response to Feser’s argument. Luiz argues that it is false that all properties of a substance inhere in the substance and have an asymmetric dependence relation to the substance. And in that case, one of Feser’s key claims — that properties depend for their reality on the substances that possess them — is false. (Or, at least, I assume Feser is saying that this applies to all properties. For if it only applies to some properties, then his argument would be invalid [this is invalid: (1) EI is a property; (2) some properties depend on substances [i.e., the substances that have them]; (3) Therefore, EI depends on substances.)

Luiz gives examples of properties of substances that don’t depend thereon. (Or, at least, examples that would be kosher by the lights of our Aristotelian-Thomistic friends.) These include what Aristotle calls secondary substances (species, genera) and specific differences (essential properties) as well as transcendental properties (being, unity, goodness, truth) which cannot depend upon the substance that possesses them in the same asymmetric way an accident is.

In any case, I’m still reflecting on Luiz’s argument as a fourth response. I offer it here because I see value in it and because you might too.

Feser: “Schmid considers something like this “circularity” objection (though his exposition of it seems to me to be quite murky, so it is possible that he has something else in mind).  In response to it, he says that if the objection had any force, it would have force against any account of the persistence of the water, including an account that attributes its persistence to God.  For if we suppose that God causes the water to persist from t – 1 to t, then we will be presupposing that it is possible for it to persist from t – 1 to t, and thus won’t be giving a non-circular explanation of how it is possible for it to do so.  And if the theist replies that God gives the water the ability to persist, then this will only push the problem back a stage insofar as it will presuppose that God has the ability to do so.”

I find this to be a very odd response, and I confess that I’m not sure I even understand what Schmid is going on about.  The circularity objection has nothing do with presupposing that it is possible for something to persist, or with presupposing that things have abilities, or anything like whatever Schmid is talking about.  Rather, it has to do with the fact that properties and powers are ontologically dependent on substances, so that substances cannot without circularity be said to be ontologically dependent on properties or powers.

Again, perhaps that is not the objection Schmid is talking about.  But if it isn’t, then I’m not sure what he is talking about.  Certainly he doesn’t seem to be talking about (a) an objection that any critic of EIT has actually given, or (b) an objection that is interesting.”

Oddly, Feser says “Certainly [Schmid] doesn’t seem to be talking about (a) an objection that any critic of EIT has actually given, or (b) an objection that is interesting.”

First, I never claimed any critic of EIT has given this in the literature. But I did receive precisely this objection from those critical of the account in question. (I wrote the paper years ago, and so I don’t remember who precisely raised the objection to me. In all probability, it was an anonymous referee.) And Feser’s finding it uninteresting is itself uninteresting and irrelevant, since it was an objection I received and needed to deal with.

Moreover, I don’t find the objection I was considering in the paper murky at all, and nor did the reviewers of the paper or other philosophers [e.g. Josh Rasmussen] I’ve discussed it with. But I guess I can try to explain it in a clearer way here for Feser and those who find it murky.

For starters, the objection I was considering is not the circularity objection Feser raises in the blog post and to which I responded above, and so it is irrelevant for Feser to point out that the objection I was considering ‘has nothing do’ with his objection. The correct response is ‘And? I wasn’t addressing your objection here.’

To get a better sense of the objection, let’s have a look at the paper:

The objection is that the explanans [the immediately temporally prior state and existence of O plus nothing intervening between that prior moment and the present moment] cannot explain the present existence of O (pace the account) — that is, it cannot explain the persistence of O from that immediately prior moment to the present moment — since it merely presupposes that O has the ability to persist from that prior state to the present state. But surely it was precisely this ability to persist that was trying to be explained!

My response was that it is not (pace the objection) the ability of O to persist that was trying to be explained. Rather, it is O’s actually persisting that is trying to be explained. So my response was that the objection under consideration mischaracterizes the explanandum. I then went on to circumvent an objection that was next leveled to me [after making the above response]: but doesn’t O’s actually persisting itself presuppose O’s ability to persist, i.e., the possibility that O persists? To circumvent this, I pointed out that any explanation of persistence will, of course, presuppose the possibility that O persists — after all, if it’s impossible that O persist, then clearly there won’t be any true explanation of O’s persistence. So yes, my account does presuppose the possibility that O persists, but this is not a problem for the account. What matters is whether or not the account presupposes the actual persistence of O. And I proceeded to point out that it doesn’t; instead, it provides a means by which such actual persistence is secured.

I don’t see anything murky here. It’s a perfectly natural dialectic, with a clear objection and rejoinder, responding to a natural-but-ultimately-misguided objection I had received.

Feser: “Anyway, Schmid goes on to discuss a further possible objection to this first way of spelling out EIT, one grounded in a presentist theory of time.  The objection would be that what happens at t – 1 cannot explain what happens at the present moment t, because (according to presentism) past moments like t – 1 no longer exist, and what does not exist cannot be the explanation of anything.  Schmid responds to this possible objection by setting out several arguments in defense of the claim that past events can play a role in explaining present ones.

Schmid does not attribute this objection to anyone, and as he rightly notes, presentists in fact do not in general claim in the first place that past events play no role in explaining the present.  So what is the point of devoting several pages to an argument no presentist has given or is likely to give?  I’m not sure, and I don’t myself have anything to add to what Schmid says in response to it.”

It’s not hard to think of an obvious purpose that justifies its inclusion: maybe I received the objection from someone critical of the account. And that is precisely what happened. (Once more, I don’t remember who it was, as this all happened years ago. But once more, in all probability, it was an anonymous referee.)

Feser: “Certainly the fact that the past is relevant to explaining the present gives (contrary to what Schmid seems to think) no support to EIT.”

This is odd. To my knowledge, nowhere do I claim in the article that the explanatory efficacy of past things supports EIT. My goal in sketching the account, there, is to point out that one metaphysical account of EIT says that past things are sufficient to explain present things. I wasn’t trying to support [i.e., positively justify] the account in talking about the explanatory efficacy of past things. Rather, I was first and foremost responding to an objection I had received that denied this, and secondly I was using it to set up my undercutting defeater for premise (7) of the Aristotelian proof.

Here’s how the undercutting defeater goes:

‘All I need to do is point out that nothing in premise (7) or that which is said on its behalf gives those who do think past things suffice to explain present things sufficient reason to abandon their position. I do not need to positively justify why past things do suffice to explain present things. I need only point out that nothing said in premise (7) or on its behalf gives those who accept the explanatory sufficiency of past things adequate reason to abandon their position. And here’s something that seems to bolster this undercutting defeater: we know that past things can and do legitimately explain, at least in part, present things. Given this, why should someone who thinks this suffices be moved to postulate something in addition?’

What I write about this in the article is as follows:

In fairness to those who think <Joe claims that the arguments Joe gives positively justify or support the thesis that past things suffice to explain present things>, in the passage above I did not distinguish between ‘sufficiently explain’ and ‘explain at least in part’, and I certainly should have. For I am intending here to point out that past things can and do legitimately explain, at least in part, present things. For purposes of clarity, I should have included this. But it is still true that I reference, in the sentence, ‘the above arguments’. And so my conclusion here — a conclusion of those arguments — is supposed to be drawn precisely from those arguments. And as we inspect the arguments, it is clear that I mean that past things can and do explain, at least in (large) part, present things. E.g., I say therein that “although our experience of an object temporally lags behind the object itself, the object itself nevertheless explains our experience. And this, in turn, entails that earlier states explain later states.” But here I clearly don’t mean that the object sufficiently explains our experience. A sufficient explanation of my experience also clearly needs to cite a functioning visual and neurophysiological apparatus, and so on.

Now let’s continue with what Feser says.

Feser: “For what is at issue in the debate over EIT and EET is not whether what happens at t – 1 is part of the explanation of what is true of the water at t, but rather whether it is by itself sufficient to explain what is true of it at t.”

I have two responses.

First:

It is true that this is part of what’s at issue in the debate over EIT and EET, but we also have to be clear about the burden of proof in the present dialectical context. In particular, we must remember that in the dialectical context of the IJPR paper [namely, the Aristotelian proof and whether one can “sufficiently undercut the argument” (p. 203 of my paper, emphasis added) by showing that its rejection of EIT “is not adequately justified” (ibid.)], 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, in spelling out my first account of EIT and in leveling my overarching EIT-based undercutting defeater of premise (7), I do 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. [Yes, according to the first account, it is sufficient to explain it. But I don’t positively claim, in the article, that the first account is true. I sketched it to facilitate the overarching EIT-based undercutting defeater of the Aristotelian proof.] Rather, my aim is to point 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’ — that is, those who do accept the first IJPR account — sufficient reason to abandon their view. The onus is thus 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 not succeeded in showing this. And so we must keep in mind that it is no mark against my paper or my case that I don’t positively justify the first IJPR account, i.e., that I don’t positively justify that past things do suffice to explain the present existence of something.

——–

[Fn] To be sure, I do go on to mount some reasons favoring EIT in the theoretical virtues section of the paper. But this is simply meant to bolster and strengthen my EIT-based undercutting defeater. I had already explained the undercutting nature of my overall thesis in the stage-setting bit of my paper [e.g., I was seeking to “sufficiently undercut the argument” (p. 20e, emphasis added) by arguing that its rejection of EIT “is not adequately justified” (ibid.).

——–

In short, I want us all to keep in mind the dialectical context: 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.

Second:

I have argued in this blog post, Section 3 of my lengthier blog post, and Part 2 that at least with the no-change understanding of the first IJPR account, the explanatory facts adduced are, indeed, sufficient. Indeed, 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 — the no-change understanding certainly removes mystery [for me, at least] as to why and how S exists at m. The explanandum was simply derived from the explanatory facts cited, and to me at least, they illuminate precisely why S exists at m. And I also confess that the same is true [again, by my lights] with the transtemporal understanding of the first IJPR account. By examining Explanans 1 from earlier, once more S’s existence at m is simply derived from the explanatory facts adduced. And those facts certainly do seem, to me, to remove mystery as to why S exists at m.

To be sure, there might be the further question about 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 is so constituted so as 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).

Anyway, let’s move on.

Feser: “(I have to say that I wonder what kind of rhetorical effect this kind of stuff has on Schmid’s readers, some of whom – judging from my combox – seem very impressed by it.  Schmid’s discussion of this first interpretation of EIT occupies almost five pages of analysis, with the standard bells and whistles that we analytic philosophers pick up in grad school and from reading academic journal articles – semi-formal formulations, the entertaining of various hypotheticals, and so on.  Other things Schmid has written, such as the article addressed in my previous post on Schmid, have a similar character.  Untutored readers, especially those whose knowledge of philosophy is largely drawn from blog posts, Reddit discussions, and the like, are bound to think: “Wow, this is so technical and rigorous!”  Yet in fact the analysis is sometimes not terribly clear, and in this case it is devoted to criticizing claims that no critic of EIT has actually made or is likely to make in the first place!  So it seems to me that some of the rigor is specious.)”

I have already addressed the allegation that the analysis is not terribly clear. To be sure, I can’t stop Feser from finding it murky. But I can at least report that neither I myself, nor the referees of the paper, nor philosophers and other colleagues I’ve conversed with about the matter, found it murky. I have also already addressed Feser’s claim that “no critic of EIT has actually made” these criticisms “or is likely to make [them] in the first place”. Apparently addressing actual criticisms I received [in all probability, from reviewers] hasn’t crossed Feser’s mind as a plausible explanation of why it’s included. And so I take his allegations here to be specious.

Feser: “Schmid considers a second possible account of EIT, according to which existential inertia is simply a basic or primitive feature of reality.  He suggests that one way of reading this claim, in turn, is that it is a necessary feature of reality that things have existential inertia. But there are two obvious problems with this.  The first is that there is no reason to believe it.  (I’ll come back to that.)”

But there are many obvious problems with Feser’s response here. First, the purpose of spelling out the metaphysical accounts of EIT is not to provide positive reasons for accepting them. Rather, the purpose is to flesh out the inertial thesis, to pinpoint that in virtue of which it is true if it is true at all.

Second, in the present dialectical context, I do not need to give positive reasons to believe this account of EIT. Once again — and this is a point I made repeatedly in Section 3 of my lengthier blog post — in the present dialectical context of Feser’s Aristotelian proof, the onus is not on detractors to positively justify this account of EIT. They need only point out that (i) nothing in premise (7) or that which Feser says on its behalf adequately justify denying the account, and that (ii) justifying such a denial is precisely what would need to be done for the proof to work. In other words, the account need only be proffered as an undercutting defeater, and one need only point out that nothing that Feser says in his chapter gives those who do accept the account sufficient reason to abandon their position.

Third, it’s false that ‘there is no reason to believe it’. You may not agree with the following line of reasoning, but it’s obvious that it at least represents a reason to believe this account of EIT: Suppose we have reason to think presentism is true. Suppose we have reason to think that if presentism is true, then the facts or truths [about temporal realities, like ‘dinosaurs exist’] change. Suppose we have reason to think that if the facts or truths themselves change, then any omniscient being’s knowledge likewise changes [since knowledge is factive, and hence if the facts or truths themselves change, the knowledge couldn’t remain unchanged]. Suppose we think that if a being’s knowledge changes, it’s not timeless. Then we have reason to think that any omniscient being isn’t timeless. Suppose, then, that we have reason to think that if there is a timeless sustaining cause of temporal things, then this timeless sustaining cause would be an omniscient God. Since we have reason to think that no omniscient being is timeless, we thereby have reason to think that there is no timeless sustaining cause of temporal things. We thereby have reason to think that at least some temporal things [debarring infinitely descending chains of per se causal dependence] persist in the absence of sustenance from without. We thereby have reason to think (a version of) EIT is true. And as I argued in the stage-setting portion of my IJPR paper, EET or EIT would be necessarily true if true at all. So we have reason to think EIT is necessarily true. This alone could justifiably raise one’s credence in the present account of EIT and thereby constitute some reason to accept it.

To be sure, I’m not here claiming that the above argument succeeds in demonstrating EIT or the second IJPR account under consideration. I’m simply pointing out that it’s just false — pace Feser — to say that ‘there is no reason’ to accept the account in question. Even if you don’t think the reason is adequate, or don’t think some step in the line of argument works, to say there is no reason here at all just strikes me as obviously wrong.

For these three reasons, Feser’s first “obvious” problem for the second IJPR account ‘obviously’ fails.

Before turning to what Feser says next, I once more want to flag that the second IJPR account is another element of the IJPR paper I would slightly modify if I could. In particular, I think it needs some clarification about what, precisely — if anything — is doing the explanatory heavy-lifting here.

In subsequent work, I’ve come to call this account a ‘propositional necessity account’ of EIT, and I wish to add greater clarity to the account from the IJPR paper. Henceforth, I’ll call this the propositional necessity understanding of the second IJPR account. To be sure, not everything I’m about to say is found in the IJPR article. That’s why I said that this is an aspect of the IJPR paper I would modify. Of course, Feser is not responsible (and nor will I take him to be, as we proceed in this post) for ignoring these clarifications. But given that such clarifications represent my current, actual views on the IJPR paper, I will proceed in my analysis of what Feser says with the clarification in mind.

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.

According to this propositional necessity understanding of the second IJPR account, inertial persistence is “a basic, primitive, foundational feature of reality”, by which the paper 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 the IJPR paper means that EIT itself—that is, the existential inertia thesis—is simply a metaphysically necessary truth.[1] 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, arguably 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.[2] In that case, though, it is surely no mark against the propositional necessity understanding of the second IJPR account that it, too, ends in something primitive, since every explanation of p—if the aforementioned reasoning holds—must likewise end in something primitive.[3]

But perhaps someone will object that this 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 metaphysical necessity, which in turn is taken to be a brute or primitive necessity. But there is a viable, readily-available, illuminating explanation of persistence in terms of the sustaining activity of something timeless. Yes, 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.

Or so the objection goes. What to make of it?

I have two responses. First, even if the alternative account adducing timeless sustenance has an explanatory advantage over the 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 the propositional necessity 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)?[4] 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 an observation made in Oderberg (2014): 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).[5] (I note that this problem is compounded even further for classical theism, since absolutely nothing about God himself varies across worlds in which he timelessly sustains O from t* to t (t* < t) versus worlds in which God withdraws his timeless sustenance at any point between t* and t — in which case, we cannot point to any fact about God that varies or differs across such worlds to account for why, in one world, O persists from t* to t as opposed to being annihilated at some point therebetween. (Is ‘therebetween’ a word? IT IS NOW!!!!) And I add, moreover, that there are many, many more worlds of the latter kind than the former, given how many moments there are in the typical object’s life.]

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 my propositional necessity account.


[1] The paper describes the 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.

[2] 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 we 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 my 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.

[3] We 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 my propositional necessity account that its explanation of p terminates in a primitive necessity.

[4] 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). Much, much more can be said here on both sides of the issue, but this suffices for a footnote.

[5] 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.)


Onward we march to Feser’s second “obvious” problem.

Feser: “The second is that there is positive reason to disbelieve it.  Again, with lions, Tyrannosauruses, water, etc., there is simply nothing about their natures or essences that entails that they exist at all.  So how could it be just a basic and necessary feature of a world comprised of such things that they persist in existence?”

There isn’t much I can evaluate here, since Feser just asks a question and, in doing so, perhaps supposes that his audience shares his intuition that the following conditional claim is true:

CONDITIONAL: If O’s existing at all is not a basic and necessary fact, then O’s persisting (once in existence) is not a basic and necessary fact.

I don’t find CONDITIONAL implausible, but I also don’t quite find it plausible. Someone who accepts the propositional necessity account will presumably just say: ‘CONDITIONAL amounts to a mere denial of my view. Why, then, should I accept it?’.

Feser next says: “Schmid also suggests that the thesis that it is a necessary feature of reality that lions, water, etc. have existential inertia is no less plausible a terminus of explanation than the thesis that God, qua pure actuality, exists of necessity.  Both theses, he claims, posit something “primitive,” but EIT is more parsimonious.”

I do suggest this in the paper, but I would modify the second IJPR account to include the clarifications above about what, precisely, is doing the explanatory heavy-lifting and how, precisely, the account compares with a rival timeless sustenance account. Next Feser says:

“But this is quite absurd. As I argue in Five Proofs and in my article on existential inertia (both of which Schmid purports to be responding to in the present article), the reason contingent things are contingent is that they are composed of parts, and in particular that they have potentialities as well as actualities.”

But it is quite absurd to say that the reason contingent things are contingent is that they are composed of parts, and in particular that they have potentialities and actualities. Consider the number two. The number two has various properties, such as the property of being even. But anything with various properties is a composite thing, by the lights of those who accept a broadly classical theistic understanding of parthood. (This is one reason they deny that God has a multiplicity of properties — that would, by their lights, entail that God has parts, whereas God is simple.) So the number two is a composite thing. But the number two is not a contingent thing. It’s not like the number two just happens to exist in some worlds and not others. So surely the reason why something is contingent is not that it is composed of parts. Or consider a view of God on which God is timeless, immutable, impassible, necessarily existent, but nevertheless has some potential for cross-world variance [say, God has potential to have timelessly performed an act of creation that is numerically distinct from his actual act of creation, or God has potential to have different intrinsic knowledge states (timelessly) across worlds, or whatever). It is obvious that nothing about God (so construed) having some potential for cross-world variance compromises God’s being metaphysically necessarily existent. And so surely the reason why something is contingent is not that it has potentialities as well as actualities.

Again, whether or not you think the number two exists, or whether or not you accept this model of God, that’s irrelevant, since we are concerned with in principle counter-examples to the claim that the reason something is contingent is because it is composite, or that it has potentialities as well as actualities.

I don’t think, then, that we should be sanguine about what Feser says in the quoted passage.

Feser: “So, when we say that God is absolutely simple rather than composite and that he is pure actuality devoid of potentiality, we have given an explanation of his lacking contingency – that is, of his existing of necessity.”

But that isn’t the point I was getting at. The point I was getting at is that even if you have an explanation of the necessity, we can equally well ask, of your explanation, why is that the case? What explains God’s being purely actual and non-composite? [Side note to some: I am not here asking about what justifies believing that God is purely actual or simple. I’m asking about what explains why this is so.] And as I argued above in connection with the propositional necessity understanding of the second IJPR account, it will ultimately be a primitive or basic fact — a necessary one, to be sure — that there is this simple, purely actual thing. One cannot appeal to its simplicity or its pure actuality to explain its simplicity or pure actuality, for that is circular. And one cannot appeal to its necessity to explain its simplicity or its pure actuality, since Feser appealed to its simplicity and/or pure actuality precisely to explain its necessity. And so both accounts end in some primitive fact. Feser might argue that his account is more illuminating, or that my account stops the explanatory buck too early. But then I will simply raise the same responses I leveled earlier in comparing the propositional necessity account to the timeless sustenance account.

Feser: “By contrast, Schmid’s proposal is that the world is made up of things that are contingent, composite, and have potentialities as well as actualities – and yet for all that it is still somehow just a necessary fact about the world that these things have existential inertia!”

Feser isn’t to blame for this, but — importantly — nothing Feser says here does anything to target the earlier comparison of the timeless sustenance account with the propositional necessity understanding of the second IJPR account. I argued therein that not only is the latter far simpler [and, hence, even if I granted that Feser has an explanatory advantage over this specific account, it’s an open question whether this sufficiently outweighs the cost in complexity], but also that the timeless sustenance account raises more questions than it answers and more mysteries than it resolves, and hence it is far from obvious whether it is explanatorily superior to the propositional necessity account in the first place.

Feser: “This is not a case of being presented with a choice between two alternative possible ultimate explanations, the Thomist’s and Schmid’s.  Rather, it is a case of being presented with a choice between an explanation and an unexplained and indeed counterintuitive brute fact.”

I have already addressed the reasons Feser proffered on behalf of his conclusion here, and so I have nothing to add except for the fact that Feser has not at all successfully done away with the propositional necessity understanding of the second IJPR account.

In Part 6, I’ll examine Feser’s section ‘Theoretical vices’.

———-

Bonus Stuff

For those who want to explore different metaphysical accounts of EIT even further, I advise y’all to check out the following:

(a)  A tendency or disposition to persist in existence (à la tendency-disposition accounts, which can be construed in metaphysically heavyweight or lightweight ways);

(b)  Transtemporal explanatory relations obtaining among the successive stages of objects’ lives or among their temporal parts (à la transtemporal accounts);

(c)   Laws of nature that govern or otherwise explain the evolution of systems and/or objects over time (à la law-based accounts);

(d)  The primitive metaphysical necessity of the inertial thesis (à la propositional necessity accounts);

(e)   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

(f)   Persistence being the absence of change and so adequately explained by the absence of sufficiently destructive change-inducing factors (à la no-change accounts — cf. also Section 4.1 of the lengthier blog post wherein I argue that the Aristotelian proof entails EIT).

To be sure, there are more besides. I’m simply giving you a flavor of the explanations on offer that make no appeal to conserving or sustaining causes. (Want to pursue them further? Click on those hyperlinks. I don’t claim Feser needs to do this or should do this if he wants to respond to me further. They’re included for those who want to dig deeper into inertialist-friendly explanations of persistence.]

———-
Author: Joe

9 Comments

  1. Gunther Laird

    For S to fail to exist at m despite existing at (m-1) is for some change to occur.[Fn] 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 at (m-1), 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-1) to m — it simply follows that S exists at m. [Cf. Section 4.1 in the lengthier blog post for those interested in pursuing this line of thought even further.]

    Yup! This is pretty much the point I was reaching at in my previous comment 🙂 I wonder how the Thomists would respond to it.

    That said, forgive me if I missed something, but is your friend Luiz an academic as well? I find his Aristotelian argument to also be interesting, if there’s anyplace he laid it out in more depth that would be great.

  2. Walter Van den Acker

    The Thomist would probably deny that non-persistence is a change. After all, they also claim that creation is not a change.

    • Gunther Laird

      I think they say that creation isn’t change from God’s perspective, but do they say the same for the creation (and by extension destruction) of ordinary things of our experience? I’m not sure they do, but I might be wrong.

      • Walter Van den Acker

        Creation isn’t a change in that, for the Thomist, It doesn’t entail an actualization of a potential.
        Now, if going out of existence is not the actualization of a potential, it is not a change either, so, according to the Thomist, it does not need a cause. Existence is something “added” to an essence, and if God stops adding existence to some essence, the idea is that this essence becomes a non-existing essence.
        How an immutable God is supposed to start or stop doing anything is beyond me, but I guess that’s another issue.

      • Gunther Laird

        Hi Walter,

        How an immutable God is supposed to start or stop doing anything is beyond me, but I guess that’s another issue.

        Hah, exactly! Forgive me for self-promotion, and I hope Joe won’t mind, but I actually touch on this subject a bit in chapter 6 of my book, the ebook of which is pretty cheap on Kindle. I describe how God’s incomprehensibility, in the Thomist view, actually undermines a lot of their ethical and metaphysical proscriptions:

        https://www.amazon.com/Unnecessary-Science-Critical-Analysis-Natural-ebook/dp/B08M68LP8F

        Aside from that,

        Creation isn’t a change in that, for the Thomist, It doesn’t entail an actualization of a potential.

        If you’re saying creation isn’t change for the Thomists *at all* (not just from God’s perspective but from everything else as well), I think in that case we can just dismiss the Thomist/Aristotelian definition of change as the actualization of potential as flatly useless and incoherent. How anyone can deny creation or destruction are changes is beyond me.

        Now, maybe you could deny they’re changes *from the perspective of the thing being destroyed,* but I don’t see how one could deny they’re changes *in any sense.* Take a chair sitting in a room that just pops out of existence. Even if we say the passing out of existence isn’t a change in the chair, it is a change in the room, as the room’s potentiality to be empty has been actualized. So we need a cause or explanation for that, which entails that a chair or any other object in the room has a sort of existential inertia, otherwise we wouldn’t need an explanation for the room suddenly becoming empty.

  3. Pingback:Feser on Schmid on Existential Inertia: Theoretical Virtues and Vices | Part 6 | Majesty of Reason

  4. Pingback:Feser on Schmid on Existential Inertia: An Argument Against EIT | Part 7 | Majesty of Reason

  5. Pingback:Feser on Schmid on Existential Inertia | A Comprehensive Response | Majesty of Reason

  6. Pingback:So you think you understand Existential Inertia? | Majesty of Reason

Leave a Reply to Gunther Laird Cancel

Your email address will not be published.