A Multi-Layered Accessibility in Modal Metaphysics — Epoche B2
True and Immutable Natures: Why Descartes's Ontological Argument Is Not the Modal One The proposal to be tested It is common to present Descartes's ontological argument from the Fifth Meditation as an early version of the modal ontological argument [1] , and to say that possible-world semantics supplies the formal machinery he lacked. On that view, casting the argument in the modal system S5 makes explicit what Descartes meant. This essay argues that the recasting is not a formalisation but a substitution. Descartes's necessity is grounded in the clear and distinct perception of a true and immutable nature [2] . The S5 version discards that ground and replaces it with a bare possibility premise, which Descartes never states and never defends, and which is where the entire modern dispute takes place. The two arguments have different premises, different burdens and different vulnerabilities. A second proposal can be dismissed on the way. It is sometimes suggested that Descartes's kind of necessity requires a special or "multi-layered" accessibility relation, distinguishing ordinary from ontological necessity. Nothing supports this. An accessibility relation is a device for saying which worlds are relevant to evaluating a modal claim; a necessity that holds everywhere just is one evaluated against a universal relation, and S5 already provides that. The instinct that S5 does not sit comfortably on Descartes is right, but the discomfort is not in the accessibility relation. What Descartes actually argues In the Fifth Meditation, Descartes has just established that whatever he clearly and distinctly perceives to belong to the nature of a thing does belong to it. He applies this to geometry: it belongs to the nature of a triangle that its angles equal two right angles, and this is discovered, not invented. He then applies it to God. The idea of God is the idea of a supremely perfect being. Existence is a perfection. Therefore existence belongs to God's nature, and separating existence from God is like separating a mountain from a valley, or three-sidedness from a triangle. Two features of this deserve emphasis, because the modal recasting loses both. First, the argument runs on a theory of ideas, not on a theory of possibility. Its warrant is the reliability of clear and distinct perception, established earlier in the Meditations . Descartes is not surveying scenarios; he is inspecting a nature. Second, Descartes is aware that the inference proves too much unless restricted, and he supplies the restriction. In the First Objections, Caterus complains that from the concept of an existing lion one cannot infer that any lion exists. Descartes replies by distinguishing true and immutable natures from natures we assemble. "Existing lion" is a composite we put together, and what we put together we may take apart. A triangle's nature is not our construction: its properties can be discovered by someone who never thought of them. His claim is that God's nature is of the second sort. Whether it is, is the argument's real load-bearing premise. The modal recasting The modern version, given its best-known form by Alvin Plantinga in The Nature of Necessity (1974), proceeds differently. Define maximal greatness so that a being is maximally great only if it exists and is maximally excellent in every possible world. Then assert that maximal greatness is possibly exemplified. In a modal logic with an accessibility relation that is an equivalence relation — S5 — the following holds: $$\Diamond\Box p \to \Box p$$ Read: if it is possible that $p$ is necessary, then $p$ is necessary. Because being maximally great already builds in existence at every world, the possibility of a maximally great being yields a necessary one, and hence an actual one. The argument is valid. Its whole content therefore sits in the possibility premise. Plantinga is candid about this: he concludes not that the argument proves God exists but that accepting its premise is rationally permissible. That is a modest claim, and it is modest for a good reason. The premise has a mirror image. If it is possible that maximal greatness is not exemplified, then in S5 it is necessary that it is not, and the argument runs to the opposite conclusion. Nothing in the logic decides between the two possibility claims, so nothing in the logic settles anything. Why this is a different argument Now compare. Descartes's premise is that we have a clear and distinct perception of a true and immutable nature in which existence is contained. Plantinga's premise is that a certain property is possibly exemplified. These are not the same claim in different notation. The difference shows up in what each must answer. Descartes must answer Caterus: why is God's nature not a composite of our making, like the existing lion? His answer is an epistemological one about the marks of a true and immutable nature. Plantinga must answer the parity objection: why is the possibility of maximal greatness better supported than the possibility of its absence? His answer is about the rational permissibility of intuitions. Neither answer transfers. Someone convinced by Descartes's account of true and immutable natures has been given no reason to accept the possibility premise, since he has not been shown that possibility, rather than conceivability by us, is the relevant notion. Someone convinced by the possibility premise has no need of clear and distinct perception at all. There is also a formal cost that is easy to overlook. Quantified modal logic must decide whether the domain of individuals is the same at every world. If it is — the constant-domain setting in which the Barcan formula holds — then everything that exists at any world exists at all of them, and "exists necessarily" no longer marks God out from anything else. If domains vary, the logic needs a free logic in which singular terms may fail to denote, and the treatment of existence as a predicate has to be built by hand. Descar