I have many, many criticisms of Feser’s Augustinian proof. I explain these in one the chapters of my (unpublished-but-currently-being-discussed-with-academic-publishers) book. I explain the book at the end of my post here. Later this summer I’ll probably make a video on the Augustinian proof, too. [Update: here’s that video!]
Today, though, I want to focus on Feser’s arguments against Platonism. As Feser recognizes, the Augustinian proof crucially relies on Platonism’s falsity. But—as I hope to show in this post—Feser has failed to justify this reliance. I can find about six problems Feser (2017, pp. 97-99) raises for Platonism. I shall tackle each of these in turn in the following sections.
1 Epistemic Access
The first problem Feser raises is the epistemic access problem. Feser writes:
But what would it be for triangularity, in the abstract and all by itself—existing, not in an actual material triangle and not in any mind, but somehow as an object in its own right—to cause something? … Yet if it has no causal powers, and thus no effects on anything, then it would follow that it has no effects on us. And in that case, how could we possibly even know about it? (2017, p. 98)
The idea seems to be as follows. Prima facie, knowledge requires causal contact with the thing known. But when it comes to Platonic abstracta, causal contact is impossible. Therefore, if there are Platonic abstracta, we cannot have knowledge of them.
I do not pretend to have a knock-down solution to this problem. I will argue, though, that the Platonist’s options are not as bleak as they would need to be for Feser’s criticism to have teeth. For, plausibly, knowledge doesn’t require causal contact with the object of knowledge. There seem to be lots of legitimate ways to acquire knowledge that don’t require causal contact. Consider, for instance, inference to the best explanation in the context of Platonism. We certainly have causal contact with ordinary particulars. Moreover, we observe various facts about them—facts that call out for explanation. For instance, we observe objective resemblances or commonalities among them; we observe the success of science and the fact that certain natural classification schemes seem essential to such predictive, explanatory, and technological success; and so on. Finally, we use standard explanatory reasoning (e.g., comparing theoretical virtues of the competing explanations) to infer that abstract objects best explain (or, as the case may be, provide the only explanation for) the relevant phenomena. None of this seems to require abstracta to causally impinge on us in some way. But yet this is surely a legitimate way to acquire knowledge thereof.
Here’s my second response. Plausibly, any intuition or argument supporting the thesis that knowledge requires causal connection will equally support the claim that knowledge requires only some explanatory connection with the object(s) of knowledge. But in that case, the Platonist can easily accept that there is an explanatory connection between abstract objects and our belief therein. For instance, according to Berman (2020), abstracta (partly) explain various features of science itself; and our beliefs about science are themselves explained by such features (e.g., science’s predictive, explanatory, and technological success); by the transitivity of explanation, it follows that abstracta (partly) explain our belief therein. Similar things can be said about, for instance, objective resemblance among particulars: abstracta explain objective resemblance or commonalities among particulars; such mundane facts about particulars explain our beliefs about those facts; so, by transitivity, abstracta (partly) explain our belief therein. Thus, plausibly, our belief in abstracta is, indeed, relevantly explanatorily connected to the fact that there are such abstracta.
2 Explanatory Impotence
Feser’s articulates a second problem for Platonism:
Furthermore, if [an abstract object like triangularity] has no effect on anything, then it has no effect on individual material objects, like the triangular billiard ball rack or dinner bell. But in that case, how could it be that which explains why those things fit the particular pattern they do? (2017, p. 98)
It’s not exactly clear what the argument here is supposed to be, but perhaps it goes something like:
- If x has no effect on y, then x cannot explain facts about y.
- Platonic abstracta have no effects on material objects.
- One fact about material objects is that they fit certain patterns.
- So, Platonic abstracta cannot explain why material objects fit the patterns they do. (1-3)
But not only does premise (1) seem unmotivated (and not only does Feser provide no justification thereof), but it also seems straightforwardly false. For there are all sorts of non-causal explanations—that is, cases where x explains y (or some fact about y) without having any causal effect on y. Suppose Ratio is the father of two children, Philo and Sophia. Suppose Ratio has 23 cookies and wants to give each of his children the same whole number of cookies without any left over. To his dismay, he finds that he can’t do this. What explains this? 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 Ratio’s desire to be unsatisfied. Nevertheless, it explains it. So, premise (1) is false: the mathematical constraints here have no (causal) effect on Ratio’s desire, but they nevertheless explain facts about it.
Moreover, the explanatory arguments for Platonism themselves constitute responses to premise (1). For instance, suppose the Platonist appeals to some one thing shared in common among numerically distinct individuals in order to explain their objective resemblance with one another. It’s not difficult to see how this explains objective resemblance—things resemble because they literally share one and the same thing in common among themselves. And this is explanatorily illuminating, says the Platonist, regardless of whether that one thing shared in common has causal effects on the particulars in question.
Thus, I conclude that Feser’s second problem is not a serious challenge for the Platonist. Onward we march, then, to the third problem.
Platonic realism seems to regard a Form as something both universal—that is, instantiated in many things—and also existing as a particular, individual thing in its own right. This seems incoherent… (2017, p. 98)
Once again, it’s not exactly clear what the problem is. Consider the case of one universal, triangularity. This is a universal—it is a single entity. But that doesn’t mean it cannot be universal in the sense of multiply instantiable or exemplifiable. All this amounts to is that numerically distinct individuals can stand in an exemplification relation to it. There is nothing contradictory or incoherent here. It’s not as though the universal is both a single entity and not a single entity. There is one entity here to which multiple distinct entities stand in exemplification relations. This is not incoherent.
Second, arguably the same problem (assuming, as we shouldn’t, that it is a problem) afflicts any realist view of universals, including Feser’s theistic conceptualism. Any realist view of (e.g.) the universal triangularity says that there is such a thing as triangularity—that is, that there exists an x such that x is identical to triangularity. To deny this is simply to adopt anti-realism about triangularity—to say, in other words, that there is no such thing. But in that case, we have one thing—triangularity—which is nevertheless in some respect universal. The problem of being both ‘individual’ and universal will therefore afflict any realist view, it seems.
To make this concrete, consider Feser’s own theistic conceptualism. Under such a view, universals are ideas in the divine intellect (2017, p. 108). But in that case, Feser’s own view treats triangularity as something universal—that is, present in many things—and also existing as a particular, individual thing (namely, an idea).
This third problem, then, doesn’t seem particularly challenging to the Platonist.
4 Third Man Argument
Feser’s fourth argument against Platonism is the so-called Third Man Argument. He writes:
Take the Form of Man, for example. Individual men are men only because they ‘participate’ in this Form, says the Platonic realist. But if the Form of Man is itself an individual object, doesn’t that entail that there must be some other Form that it ‘participates’ in and by reference to which it counts as the Form of Man specifically? Don’t we have to posit a Super-Form of Man over and above the Form of Man, in which both individual men and the Form of Man itself all ‘participate’? Indeed, wouldn’t we have to posit a Super-Super-Form of Man over and above that Super-Form, in which the Form of Man, the Super-Form, and individual men ‘participate’? We seem led into infinite regress, and absurdity. (2017, p. 98)
I don’t think a Platonist should be sanguine about this argument. Something only instantiates or exemplifies [or ‘participates in’] the universal manhood if it is an individual, particular, flesh-and-bones man. But the universal manhood is not a man. For one thing, the former is non-spatiotemporal while the latter is spatiotemporal. So, there is no need (nor could there be need) for some higher order universal, manhood1, that both individual men and manhood exemplify. This would only be the case if there were some common or shared feature among individual men and the universal manhood. But there isn’t much in common between them—the former are concrete, flesh-and-bones, bipedal, hairy, rational animals, whereas the latter is an abstract, non-causal, non-spatiotemporal universal or property. If they have something in common, it surely won’t be anything relevant to some high order manhood, manhood1.
Moreover, an exactly parallel problem (assuming, as we shouldn’t, that it is a problem) faces Feser’s theistic conceptualism. Consider the following parody. Take the divine idea of manhood. Individual men are men only because they share (or ‘participate in’ or exemplify) this same manhood, says the theistic conceptualist. But if manhood is itself an individual divine idea, doesn’t that entail that there must be some other divine idea that it ‘participates’ in and by reference to which it counts as the idea of manhood specifically? Don’t we have to posit a super-divine-idea of manhood over and above the divine idea of manhood, in which both individual men and the divine idea of manhood itself all ‘participate’? Indeed, wouldn’t we have to posit a super-super-divine-idea of manhood over and above that super-divine-idea, in which the divine idea of manhood, the super-divine-idea, and individual men ‘participate’? We seem led into infinite regress, and absurdity.
Thus, Feser’s fourth argument is not a problem for the Platonist, and even if it were, it would equally pose a problem for Feser’s own theistic conceptualism.
5 Individuality and Incoherence
Here’s Feser’s fifth argument against Platonism:
Consider a universal like animality. Every individual animal is either rational (as human beings are) or nonrational (as all other animals are). But what about animality itself? Precisely because it is universal, it has to apply to both rational and non-rational animals. But it can’t itself include both rationality and non-rationality, for these are contradictory. So, we have to say that inherently it entails neither rationality nor non-rationality. But no individual thing can be neither rational nor non-rational; any existing thing has to be one or the other. Hence, the universal animality cannot be said to exist as an individual thing in its own right; that is to say, it cannot properly be thought of as a Platonic Form. (2017, pp. 98-99)
But this argument fails. In one sense, the universal animality entails only non-rationality. In another sense, the universal animality entails neither rationality nor non-rationality. There is nothing absurd here.
Here is the perfectly respectable sense in which animality entails only non-rationality: the universal itself is not rational (i.e., capable of rational deliberation, thought, and activity), and hence the universal itself exemplifies non-rationality.
Here is the perfectly respectable sense in which animality entails neither rationality nor non-rationality: that which exemplifies animality does not automatically (else: thereby) exemplify rationality (in particular) or non-rationality (in particular). In other words, x’s exemplification of the property of animality entails neither x’s exemplification of the property rationality nor x’s exemplification of the property non-rationality.
There is nothing contradictory or absurd here, and a fortiori there is nothing absurd here that results from treating animality as a single entity.
Finally—and once again—an exactly parallel problem (assuming, as we shouldn’t, that it is a problem) faces Feser’s theistic conceptualism. Consider the following parody. As Feser himself argues, we have to say that animality inherently entails neither rationality nor non-rationality. But no divine idea can be neither rational nor non-rational; whatever exists has to be one or the other. Hence, animality cannot be said to exist as an individual entity (which would have to be the case if it were a divine idea). That is to say, it cannot properly be thought of as an idea in the divine intellect.
Feser’s fifth problem, then, shouldn’t worry Platonists.
6 Immanence Problem
Here’s Feser’s sixth criticism of Platonism:
Platonic realism implies that the natures of the things of our experience are not in the things themselves, but are instead non-spatiotemporal. The nature of a tree, for example, is not to be looked for in the tree itself, but in the Form of Tree; the nature of a human being is not to be looked for in any human being but rather in the Form of Humanity; and so on. Now, if treeness is not to be found in a tree, nor humanity in a human being, then it is hard to see how what we call a tree really exists as a tree or what we call a human being really exists as a human being. (2017, p. 99)
Again, it’s not clear what the argument is here, but I’ll consider the claims made in turn. First, Feser says that natures are ‘not to be looked for’ in the things whose natures they are. Perhaps this is simply a statement of the Platonic commitment to the non-spatiotemporality of abstracta—one will not find the universal spatiotemporally located where its instances are. This is correct, but it seems uninteresting—it’s just a restatement of the Platonic view. Feser does go on to report that it is ‘hard for him to see’ how this commitment allows us to say that ordinary particulars exist as those particulars. But I must admit that it is hard for me to see how this is at all problematic for the Platonist. Platonists don’t deny the reality of the ordinary particulars around us, and nor do they deny that such ordinary particulars have various intrinsic and extrinsic properties. Finally, they also don’t (or needn’t) deny that many such intrinsic properties are essential to those particulars. None of this is compromised by the mere fact that properties are non-spatiotemporal. I confess, then, that I simply don’t see any force behind Feser’s sixth criticism.
Feser claims that his objections to Platonism are “insuperable” (2017, p. 109). While many adjectives describe his objections, I hope to have shown that ‘insuperable’ is not one of them.
 As I use it, Platonism is the view that there are mind-independent, necessarily existent, non-spatiotemporal, non-causal abstract objects like properties/universals, propositions, and mathematical objects.
 It is also called the Benacerraf problem (cf. Benacerraf (1973)).
 For an exploration of debates surrounding the Benacerraf problem, see Panza and Sereni (2013).
 See Berman (2020) for an argument from the success, progress, and practice of science to the truth of Platonism.
 Plausibly, those who mount the Benacerraf problem have to admit the legitimacy of this kind of transitivity, since obviously not all our knowledge of ordinary particulars is directly caused by them.
 Obviously, there are different accounts of exemplification. For present purposes, a generic and simplified notion will suffice.
 Pun intended.
 Universals are typically posited (in part) to explain commonalities among individuals. It is therefore only because things have the relevant universal, F-ness, that they are all F. Hence, if universals are divine ideas, it follows that individual men are men only because they share (or ‘have’ or ‘participate in’ or ‘exemplify’ or whatever) the divine idea of manhood.
 Or, if we want a sparser ontology, it fails to exemplify rationality.
 Feser offers what might be interpreted as a seventh criticism when he writes: “Indeed, the trees and human beings we see are said by Plato merely imperfectly to “resemble” something else—namely, the Forms. So, what we call a tree seems at the end of the day to be no more genuinely treelike than a statue or mirror image of a tree is; what we call a human being seems no more genuinely human than a statue or mirror image of a human being is; and so forth. But this is absurd” (2017, p. 99). But this criticism does not attack the kind of Platonism with which I am concerned in this post. The particulars of our experience, according to this contemporary analytic version of Platonism, do not ‘resemble’ abstract objects. They are not imperfect or ‘less real’ images of them. So this criticism is neither here nor there for present purposes.