Thirteen Years, Two Critics, Same Page

On September 21, 2026, Thomist philosopher Edward Feser posted the second installment of a takedown of William Lane Craig’s newly published Systematic Philosophical Theology, Volume IIa — specifically, Craig’s roughly 70-page treatment of divine simplicity (Feser, “Craig on Thomism and divine simplicity, Part II”). Page 151 of Craig’s book contains the line at the center of the fight: “We make-believe that there are such things as properties in talking, for example, about God’s perfections.” Properties, for Craig, are not real. They are a convenient fiction — a nominalism he has held since at least his 2004 essay “Creatio Ex Nihilo and Abstract Objects,” where he wrote that “the chief theological failing of Platonism… is that Platonism is incompatible with the doctrine of creatio ex nihilo and so fundamentally compromises divine aseity… An orthodox Christian theist, then, cannot be a Platonist.” If properties, propositions and numbers exist as abstract objects the way Plato thought, Craig argues, then God is no longer the sole uncreated, self-existent reality — so they must not exist that way at all. The dispute worth watching isn’t theism versus atheism. It’s whether Craig’s nominalism can survive contact with the very job he built it to do, and two entirely independent critics — thirteen years apart, arguing from opposite ends of philosophy — say it can’t, on the same page, for related reasons.

Feser’s version: Craig needs some fact about the world to explain why fire reliably burns, acorns reliably grow into oaks, and the like — some ground for those regularities holding rather than an infinite number of others. Either there’s an objective, necessary fact about how properties relate (which is realism, restored through the back door), or there isn’t, and creation is arbitrary all the way down — a voluntarism Craig explicitly rejects because it violates the principle of sufficient reason he’s committed to elsewhere. Worse, Craig’s own stated fallback — “on an anti-realist perspective on properties, God is, indeed, not composed of metaphysical parts and is in that sense simple” (p. 152) — quietly grants Aquinas’s conclusion while insisting he’s refuting it.

The older version of the objection, from R. Scott Smith of Biola University, doesn’t touch theology at all. In a 2013 Philosophia Christi paper (PDF), Smith argues that Craig’s nominalism can’t even secure the meaning of Craig’s own sentences. A mental state — a belief, an experience — has to be determinately about something (its “intentionality”) for it to be that state and not some other; your experience of a coffee’s taste couldn’t have been, instead, an experience of a jackhammer. If that aboutness is a real essence, you’re back to the abstract objects Craig is trying to expel. If it isn’t real, “there is no fact of the matter what a given mental state really is about” — which, applied to Craig’s own kalam cosmological argument and his own theory of reference, makes his position self-undermining: an argument that, on its own premises, isn’t determinately about anything. Smith reran the same charge in a 2022 sequel, “Propositions: Who Needs Them? Craig’s Nominalism Revisited.” That’s three publications — 2013, 2022, 2026 — thirteen years apart, from a Thomist and an analytic epistemologist who otherwise share almost nothing, converging on the identical structural weak point: nominalism about properties, pressed hard enough, either quietly readmits realism or can’t state its own claims. Craig has a standing reply pattern for adjacent worries — distinguishing “alethic” realism (his sentences can be true) from “ontological” realism (the things they’re about needn’t exist) — but neither Feser’s 2026 post nor a search of Craig’s public record turns up a reply to Smith’s specific self-refutation charge.

What this changes: nominalism keeps getting recruited, in and out of philosophy of religion, to do a specific job — deflate an ontology while keeping everything that ontology was paying for (necessity, meaning, structure) — and it keeps failing that job the same way, whether the pressure comes from “why does nature hold together” (Feser’s route) or “how can any claim be about anything” (Smith’s route). The SEP entry on divine simplicity frames aseity as requiring that God “be rather than have” his attributes; Craig’s fix was to make nothing, anywhere, really have attributes. My bet: the next substantive move in this fight won’t be a new argument for nominalism at all — it’ll be Craig or an ally trying to change the terrain back to Platonism’s causal-access problem, because nobody defending the anti-realist side has yet answered either fork, the necessity fork or the meaning fork, on its own ground.