4. Metaphysical Rivals

With the positive argument in place, the chapter turns to rival accounts, each taken up in its most demanding form rather than its most convenient one. The question put to every rival is the same: does it actually eliminate T, S, and Φ — Temporality, Spatiality, and Physicality, borne by Time, Space, and Substance — or does it only redescribe them?

At the limit, a rival can make one of three moves. It can deny the structural conditions, holding that they do not obtain. It can ground them prior, granting that they obtain but locating them downstream of something more fundamental. Or it can grant the conditions while charging them with insufficiency: T, S, and Φ obtain, but do not exhaust what Reality contains. The chapter follows each position to where it actually leads.

What is at stake is whether the derivation survives contact with the live literature — so no deflationary program is dismissed as unserious, and no classical-theist position is treated as unavailable. The standard is the strongest version currently on offer; where a rival has a stronger form, known or yet to come, the argument is answerable to it, not exempt from it.

Four directions organize the engagement: modal and semantic foundations (§4.1), time and change (§4.2), physics and mind (§4.3), and modal variants (§4.4). The same question is put to each, and §4.5 draws together what the pattern across them reveals.

The chapter inherits the formal framework secured in Chapter 2, but it does not treat every rival as answering to one uniform reply. Some engagements test the fixed global domain of real possibility and the modal semantics that follow from it (§2.1). Some test whether a rival realizes only a thin abstract analogue of directed order, difference, or bearing, or realizes the fuller types required by genuine change. Others accept the structure and test whether it is rich enough. The verdict in each section therefore tracks the rival’s actual burden: incoherence, relocation, compatibility, explanatory cost, or an open seam.

4.1 Modal and Semantic Foundations

This first cluster gathers attempts to ground, deflate, or bypass the argument’s modal and structural claims: accounts that would explain modal talk away, that would relocate Time, Space, and Substance to a simulated or solipsistic substrate, or that would derive them from something more fundamental still, sufficient reason, Humean succession, or information. These positions differ, but they share one aim: to account for change and its structural conditions without appeal to the structure the argument derives. One entry shares no such aim: Pizzi’s impossibility of nothingness (§4.1.11), the literature’s nearest convergence — the same closure of modal space against ∅, reached by kindred machinery and stopping short of naming the necessary content. Three further entries close the cluster on the modal side proper. Dispositionalism about modality (§4.1.12) grounds possibility in the potentialities of actual objects rather than in a plurality of worlds, and is the account nearest to this one. Modal epistemology (§4.1.13) grants the vocabulary and ordinary modal knowledge alike, and challenges the reliability of the modal faculty at the distances this argument works at. Impossible worlds and relative modality (§4.1.14) ask whether the quantifier in □(T∧S∧Φ) ranges over everything, or over a domain that could be widened by admitting impossible worlds or narrowed by a contextual modal base.

The question is whether any of them can describe a reality in which Time, Space, and Substance are absent. The reply is that none can. A simulation still needs a host that really changes; a solipsistic mind still runs on succession and distinction; a deflationary theory still needs the structure it redescribes. The challenge bears most directly on §§2.1–2.6 and the Modal Identity of §2.6.6: change-possibility and the joint obtaining of T, S, and Φ are one fact stated from two sides.

A preliminary note before the named engagements. The objection that everything we take to be reality might be a simulation (that what we take to be physical succession, extension, and persistence could be virtual artifacts of a deeper computational substrate) does not escape structure. It relocates it. Any host system capable of hosting simulated succession must itself instantiate genuine succession; any host system capable of hosting simulated extension must itself instantiate relational extension: distinct positions standing in a traversable interval (the thin, structural S of §2.4.7), not necessarily a metric manifold of any particular geometry; any host system capable of hosting simulated persistence must itself persist. What the relocation secures is the structural role at each level (succession, traversable distinctness, persistence), not the contingent metric character (a particular spacetime or geometry), which is physical content established downstream (§7.3).

The same diagnosis applies to hard solipsism: a solipsistic mind whose experience exhibits succession, distinction, and exchange is itself a configuration within T, S, and Φ. The question of whether anything else exists does not affect the necessity of T, S, and Φ. Both retreats relocate structure; they do not remove it.

A second preliminary note concerns alethic and doxastic modality. The engagements throughout this chapter operate in alethic modality, what holds across possible worlds independently of what any mind conceives or believes. Oppy¹ argues that conceivability-to-possibility inferences are unreliable because conceivability tracks cognitive limitations rather than modal facts, and expresses agnosticism about whether broadly logical and metaphysical modality are genuinely distinct. Both points are well-taken against arguments that derive alethic modal claims from doxastic premises. The argument of this book does not make that move. Its modal claims are grounded in actuality: change, reading, and conceiving are actually occurring, and the actuality-to-possibility inference, every actual fact is possible (§1.2, the Actuality definition), is logically valid. Alethic modality is prior to doxastic modality because doxastic states presuppose it: a belief that P is a belief that P is true, truth conditions are alethic, and the very existence of a cognizer capable of doxastic states is itself an alethic fact. Oppy’s agnosticism about the broadly logical/metaphysical distinction is not dissolved merely by the Condition-Individuation Criterion (§2.6.11). As §2.10 now states, distinguishing the temporal, spatial, and physical subject matters does not by itself derive three separate consequence relations. The book proposes three formally differentiated disciplines or readings, but whether broadly logical and metaphysical modality are distinct remains a substantive semantic question rather than something the criterion settles by itself.

4.1.1 Rigid Designation

The Kripkean challenge is a wedge any argument for necessary existence has to clear. The move separates necessary co-extension (two terms picking out the same set of things in every possible world) from necessary identity (two terms picking out the very same thing as a matter of what it is), and presses that the argument has established only the former. Wherever change is possible, T, S, and Φ are correctly described as its preconditions; but this is a truth about how we specify what change requires, not a truth about what must exist. The replies that follow grant Kripke’s narrower point (necessary co-extension does not by itself establish identity, and many necessary-existence arguments fail by conflating the two) and contest a different one. In this case the two descriptions pick out one fact, and §2.6 discharges the burden of showing it, especially the Modal Identity at §2.6.6 and the Triconditional Identity at §2.6.10.

Necessarily, if something is water, it is H₂O. But water itself might not have existed: on another planet, or in a world with slightly different chemistry, there could have been no water at all. The necessity belongs to the identity of water with H₂O; it does not secure that water exists in every possible world. Saul A. Kripke² draws the distinction: some necessities are de dicto, holding of the description, while others are de re, holding of the thing. Hilary Putnam³ presses the same point with Twin Earth: the same words can pick out different substances in different worlds. Necessity lives in reference, not in existence.

The distinction has force because it exposes a pattern common to arguments that appear to prove necessary existence. An ontological argument that moves from conceivability to necessary existence needs more than rigid designation: it needs a justified possibility premise and a valid modal bridge. Kripke’s analysis does not itself collapse such arguments; rigid designation can instead help state their de re conclusion. The live question is whether the argument has earned the relevant possibility and existence claims. A Leibnizian cosmological argument moves from ‘something exists’ to ‘a necessary being explains it,’ but the necessity attaches to the explanation, not to the being explained. Kripke’s wedge has closed more arguments than it has opened; the question is whether it closes this one.

The challenge bears directly on §2.6. If the necessity is hollow (a consequence of how ‘change’ is defined rather than a modal fact), the argument collapses into conditional necessity and the classical theist regains room for a changeless necessary being outside structure. The claim of §2.6 is narrower: that change-possibility and T ∧ S ∧ Φ are not two correlated facts that march together across worlds but one structural fact disclosed from two directions. Three forms of the move take up the objection.

First. The sparse-world move does not require a fully empty world, only one in which T, S, and Φ are present by description without being metaphysically robust. Call it w′, stipulated here as a test-case for the conditional in the manner of Kripke’s Lecture I naming conventions, though the dichotomy that follows holds equally if w′ is surveyed as an independently real Lewisian world. In w′, the descriptions apply, but what they pick out is thin: no causally structured succession, no relational extension, no causal persistence, only the verbal placeholders. The argument’s necessity then reduces to a truth about descriptions, not about what worlds contain.

The move faces a dichotomy it cannot survive. Either (i) w′ contains some metaphysical content beyond the descriptions, or (ii) it contains none. If (ii), w′ is ∅, foreclosed by §2.7: a description without a world to describe is not a description at all but a string of words with no being. If (i), the remaining content either suffices for change or it does not. If it suffices, T, S, and Φ obtain in their operative sense after all and w′ is not sparse in the relevant way. If it does not, there is no change in w′, and the world is not a counterexample to the conditional but a witness to its antecedent failing. There is no third reading on which w′ contains metaphysical content while failing to instantiate the operative sense of the conditions, because the operative sense is what structural content consists in.

Second. Exempting the preconditions from S5 grants the ∅ result but restricts S5’s reach. The refinement concedes that T, S, and Φ obtain in every change-permitting world but denies that S5 licenses the further step to their obtaining in every world simpliciter. S5 governs inferences within the space of possibilities; it does not settle the status of that space itself. So the modal behavior of the preconditions can be governed by S5 while S5 still fails to bridge the conditional.

The refinement sharpens the problem rather than loosening it. There may be worlds without actual change that are nevertheless structured and distinct from ∅; a frozen or silent world is the obvious test case. Section 2.7 rules out absolute structural absence, not every world in which change fails to occur. The Kripkean therefore is not forced into a simple change-or-∅ dichotomy. To close the exemption move, the argument needs the separate claim that every possible world retains the capacities or role-structure required for change, and the role-to-type bridge must not be assumed in stating that claim. The local retorsion does not supply this cross-world result; it is supplied independently by the role-level convergence of §§3.2–3.3, which earns structure in every possible world at role-resolution, with the role-to-type bridge earned at §3.5.3 rather than assumed.

Third. The rejection of the Modal Identity is the form of the objection that goes deepest: it grants the prior replies but denies the underlying identity of §2.6. Necessary co-extension is not identity (Kripke himself distinguishes necessary identity, water = H₂O, from mere necessary biconditionals) and so the possibility of change and the instantiation of T, S, and Φ may be necessarily co-extensive without being one fact. Fine’s⁴ distinction between essence and necessity cuts the same way: Socrates and {Socrates} are necessarily co-present without the singleton being part of Socrates’s essence. The wedge would hold, and the structural conditions would be necessarily co-present with change-possibility without being what change-possibility essentially consists in.

Fine’s point reinforces rather than undermines the argument. The distinction between necessary biconditional and necessary identity is genuine, and Kripke is right to insist on it; but it cuts against the objector here. A necessary biconditional between two distinct facts is established by the world: the world just happens to be such that whenever one obtains, the other does. A necessary identity between two descriptions of one fact is established by examining the descriptions themselves: they are revealed — empirically in Kripke’s case, by analysis here — to pick out the same thing. Kripke’s water/H₂O case is the second kind. The Modal Identity of §2.6 is of this second kind, not the first.

This is shown at §2.6 by analyzing what change consists in: a succession (Time) of states (Substance) across a structured interval (Space). The analysis does not establish a biconditional between change-possibility and T ∧ S ∧ Φ; it shows that the right-hand side is what the left-hand side is, spelled out. By Fine’s lights, the argument’s claim is an essence claim: T, S, and Φ are what change essentially is, not merely what change necessarily entails. Kripke’s wedge works only where the description and the substance are logically independent, where ‘water’ and ‘H₂O’ could, in principle, have picked out different things. The structural conditions are not like that.

One point is still owed to the objector who wants more than assertion for why the analytic reading beats the discovered-identity reading. The water/H₂O case supports a genuine two-description gap because the Twin Earth scenario that exhibits it is set up using ordinary space, time, and chemistry, and only then asks which of two candidate substances a surface concept picks out; the scenario does not itself presuppose water or H₂O. No parallel scenario is available here. To construct a world in which ‘causally structured succession’ turns out, on investigation, to pick out something other than Time, the construction needs a succession across which the investigation unfolds, a distinction between the surface concept and the candidate referent under examination, and a persisting investigator to carry the inquiry from setup to verdict. The thought experiment that would exhibit the gap cannot be run without the very structure it is trying to show is merely co-extensive with change rather than identical to it. This is not the cheap point that any act of arguing takes time; it is the narrower one that this particular thought experiment’s own machinery is drawn from the domain the gap is supposed to open up in, which the water/H₂O case’s machinery is not.

The objection will press that this shows only an epistemic limit — that we cannot represent the counterexample — not a metaphysical fact. But imaginability was never the measure. One can imagine what is merely false: that one is not reading this sentence, that one is elsewhere entirely. Imagination reaches those because they are genuinely possible though not actual — contingent facts it is free to lift. What it cannot lift is the standing it lifts them from: the succession across which the imagining unfolds, the standpoint it is entertained at, the persistence of the one entertaining it. To deny change is already to transition from not-denying to denying (§2.2.1); to imagine Nature away is to instantiate the Conditions in the imagining (§2.8.2). w′ does not ask us to imagine a contingent fact otherwise; it asks us to imagine beneath this floor, and beneath it there is not a scenario hard to reach but no scenario — ∅ (§2.7). The inability is therefore not epistemic. The counterexample would have to be specified through the very roles it denies, and a specification that consumes its own content is not a possibility we cannot access but a description that never closes. This is Cartesian in the exact sense the book uses: not that the identity is vivid to us, but that its denial cannot be performed.

Their referring status is fixed accordingly. T, S, and Φ are not definite descriptions that might have designated something else, but the irreducible existence-axes individuated at §2.6.11 — real structural joints holding across every world (§2.6.6). And ‘Reality’ is reference-fixed at §2.8.5 as their joint complex, its definiendum rather than a predicate merely co-extensive with it. That fixed reference is what makes the identity an essence claim rather than a truth about how ‘change’ is labeled. They are not what we call the preconditions of change but what any possibility of change consists in.

A kindred deflationary move drops rigid designation entirely in favor of modal conventionalism:⁵ necessary truths are analytic consequences of our classificatory conventions rather than discoveries about reality. The same reply applies: classificatory conventions presuppose classifiers that persist across a domain and operate through time: Substance, Space, and Time under different names. Modal structure cannot be a convenience of our conceptual apparatus unless that apparatus is itself instantiated in the structural conditions the argument describes.

Adjacent modal deflations are engaged in their own sections: two-dimensionalism (§4.1.2), the Principle of Sufficient Reason (PSR) demand (§4.1.3), modal collapse (§4.4.9), and the changeless-being objection (§4.2.1).

One last consideration, stated narrowly. The retorsion that follows (a retorsion being an argument that turns an objection back on its maker, showing that the very act of stating the objection presupposes what it denies) establishes the conditions’ actuality wherever the Kripkean objection is articulated, not their necessity across worlds where no such articulation occurs. The modal ascent is the work of §§3–7 and does not rest here. What the retorsion shows is narrower: the Kripkean vocabulary is parasitic on the structure it purports to loosen.

What closes. Three forms of the Kripkean wedge, and none opens the gap. Each candidate world has either content enough to individuate it (in which case T, S, and Φ obtain in their operative sense) or it is ∅; the S5-exemption move collapses into that same dichotomy. Fine’s essence/necessity distinction, the deepest form, cuts the other way. The Modal Identity at §2.6.6 is one fact under two descriptions, not a necessary biconditional between two: Kripke’s own distinction, turned against the objector.

The narrower Kripkean point (that necessary co-extension does not entail identity) is not what the argument contests. On Kripke’s own taxonomy, □(T∧S∧Φ) is not a necessary a posteriori truth on the water=H₂O model either. The one place the derivation touches actuality is the datum that change occurs (§2.4), and even that enters not as ordinary empirical discovery but as Cartesian certainty, epistemically inescapable, surviving the strongest skeptical scenario (§2.2.1). The structural identity is constitutive rather than discovered: §2.6 recovers it by analysis, and its necessity is fixed a priori by the S5 lift of that datum (P2), not read off the world as the water=H₂O identity is. □(T∧S∧Φ) is therefore an a priori, constitutive necessity reached from an inescapable datum. The two-dimensional bookkeeping this raises is handled at §4.1.2.

Kripke’s own theory of reference is causal-historical: a name picks out its referent through a chain of transmissions from a baptismal event, preserved across time, through persistent speakers, in writing and speech stable enough to carry reference forward. Rigid designation is not a free-floating stipulation; it is analyzed as an actual causal chain. The paradigm case (‘water = H₂O’) reached philosophy through exactly such a chain: chemistry to laboratory to textbook to Naming and Necessity to here, rigidly designating the same content at every point on the way. The Kripkean apparatus presupposes what §2.6 describes: a manifold individuated by structural content, a referent preserved across positions, an asymmetric chain from origin to use. The machinery driving the Kripkean wedge is itself an instance of the Modal Identity the wedge was meant to loosen.

The paradigm case remains formulated within ordinary chemical and spatiotemporal theory: H₂O describes molecular structure and its physical realizations. That shows that the water/H₂O example does not itself provide a structure-free countermodel. It does not show, merely from the molecular formula, that every possible realization must exhibit motion or the book’s full loaded types. Every essence-specification, fully unpacked, runs back to T, S, and Φ. The wedge exposes necessary identities that hold in virtue of what their relata structurally consist in, and the relata, here as elsewhere, consist in the conditions.

4.1.2 Two-Dimensional Semantics

Two-dimensional semantics promises a wedge between what a concept picks out when a world is considered as actual and what it picks out across possible worlds. If that wedge holds at the constitutive level, the Modal Identity of §2.6.6 (□(◇C ↔︎ T∧S∧Φ) read as one fact under two descriptions, connecting change-permission with co-instantiation of the three structural roles) will not survive. The challenge presses at three points: the two-intensions machinery itself, the water/H₂O analogy that most cleanly exhibits it, and the indexical anchor on which primary intensions depend. Chalmers is right on a narrower point, conceded here without qualification: the primary/secondary distinction is a genuine feature of natural-kind terms, and arguments that ignore it routinely overstate what necessary co-extension delivers. The claim contested below is narrower and specifically targeted: that the structural role definitions are natural-kind terms, and that the two-intension machinery has purchase where the appearance/nature separation it exploits does not exist.

Chalmers⁶ makes the cleanest contemporary case that meaning and reference come apart. Every concept, on the two-dimensional semantic framework, admits two intensions. The primary intension picks out what the concept would refer to in any world considered as actual: what ‘water’ would pick out if the world turned out one way rather than another. The secondary intension picks out what the concept refers to across all possible worlds, given what it in fact refers to here. For natural kind terms the two come apart cleanly: ‘water’ picks out the clear drinkable stuff in any world considered as actual (primary), but once we discover the actual stuff is H₂O, ‘water’ rigidly designates H₂O across all possible worlds (secondary). The framework tracks a gap between description and reference that rigid designation alone cannot capture.

Frank Jackson⁷ develops the apparatus in parallel, pressing that conceptual analysis tracks primary intensions and that every metaphysical necessity has an a priori scrutable counterpart at the primary level (one derivable from a compact base by ideal a priori reasoning alone, without further empirical input). David J. Chalmers and Frank Jackson⁸ extend the framework into a general theory of reductive explanation: wherever a target fact is not a priori entailed by a non-target base, a two-dimensional gap opens. David J. Chalmers⁹ generalizes the move still further in Constructing the World: all truths are a priori scrutable from a compact base, and wherever primary intensions fail to track secondary ones, a genuine modal gap has been uncovered. The most aggressive deployment for modal epistemology is David J. Chalmers:¹⁰ ideal a priori conceivability, properly regimented by 2D, entails primary possibility, and primary possibility is the kind of possibility that matters for disputes over necessary existence. The framework is not a regional tool for natural-kind terms; it is a general theory of where modal weight sits and where it does not.

Applied to the present account, the framework presses a precise threat. The argument of §2.6 holds that ‘causally structured succession’ and Time are not merely co-extensional descriptions but the same structural fact under two modes of presentation: that the role definitions are constitutive rather than descriptive. The two-dimensionalist replies that even for terms the argument labels constitutive, primary and secondary intension can diverge: what ‘causally structured succession’ picks out when a given world is considered as actual may differ from what it picks out across possible worlds. The Modal Identity of §2.6.6 would then fail, and with it the step from □◇C to □(T∧S∧Φ): from the necessary possibility of change to the necessary co-instantiation of the three structural conditions. A wedge between description and reference at the constitutive level would leave the structural roles necessarily co-extensive with change without being identical to it: the exact space the classical theist needs to reassert an uninstantiated ground.

Three forms of the move take up the objection; the narrower point about natural-kind terms stands untouched.

First. The two-intensions move holds that every concept admits both intensions and that the argument illicitly collapses them at the point of constitutive terms. The pressure is sharpest in its epistemic-possibility form. The primary intension captures what the world could turn out to be for all we know a priori, and the Modal Identity of §2.6.6 is precisely the kind of a priori necessity claim Chalmers¹¹ has shown can fail even where the secondary intension holds necessarily. For all the argument has established analytically, the objector presses, ‘causally structured succession’ could turn out (in some world considered as actual) to pick out something other than Time as §1.4.1 defines it. Scrutability¹² generalizes the pressure well beyond natural-kind terms: wherever a concept is a priori scrutable from a base that does not itself entail the target, a primary/secondary gap is in principle available, and the constitutive identities this argument rests on fall within that scope.

The reply runs on the analytic/a posteriori distinction the 2D framework itself respects, but it must keep two resolutions apart. At role-resolution, a constitutive definition fixes what counts as directed order, relational differentiation, or real bearing. At type-resolution, further arguments identify temporal succession, spatial extension, and causal persistence as the relevant realizers in genuine change. The Modal Identity does not treat the names alone as settling those richer types.

The Kripkean cases 2D was built to accommodate are a posteriori identities: ‘water is H₂O’ was discovered, not defined, and the primary/secondary gap tracks exactly what the discovery added. The structural claim is different at role-resolution: ‘causally structured succession’ specifies the role T plays in a genuine change-realizer rather than reporting a discovery that two independently identified things coincide. Where that role specification is analytic, primary and secondary intensions have no independent residue to track. A critic may still contest the type-level realization, but that is a dispute over the realizer arguments, not a primary/secondary gap generated by the role definition itself.

The 2D matrix is a function from possible worlds to extensions, and functions require a domain. In the present account, real possibility ranges over a fixed global domain W (§2.1). Two-dimensional semantics may offer a different representation of modal evaluation, but it must say whether its primary-intension scenarios range over that same global possibility-domain, a relative accessibility relation, or an epistemic space. Only the first is the book’s alethic operator ◇. The semantic tool therefore does not by itself overturn the fixed-domain account; it either represents the same real possibilities at another level or shifts to a different relative modality. The structural roles become relevant where a scenario is offered as a genuine change-realizer, not merely because any semantic matrix is being described.

The disanalogy with the paradigm case can be stated directly. The water/H₂O primary intension is evaluable without presupposing water or H₂O: the evaluation is built from chemistry, sense-perception, and an ordinary spatiotemporal setup, none of which is the disputed term. No parallel construction is available here. Evaluating the primary intension of ‘causally structured succession’ requires exactly the apparatus in question: a scenario individuated from its alternatives (Space), a reasoner who carries the evaluation from premises to verdict (Time), and content held fixed long enough to be compared against the secondary intension it is meant to diverge from (Substance). The act of running the semantic tool is structured, but that act-side fact does not show that every scenario represented by the tool instantiates the loaded temporal, spatial, and causal types. The object-side question remains whether a scenario is admitted as a genuine change-world or only as an epistemic representation.

Second. The water/H₂O analogy sharpens the pressure. Scientists discovered water is H₂O a posteriori: the identity holds necessarily but was not analytically given. Chalmers and Jackson¹³ generalize the pattern: every necessary a posteriori identity admits a corresponding primary-level contingency, and the identity’s modal weight is confined to the secondary level. If ‘causally structured succession’ and Time stand in this relation, the objector presses, their identity is a discovery rather than a definition, and a discovered identity is necessary only in the restricted secondary sense that leaves the primary-level gap open, which is the gap the theist needs.

The analogy works in the natural-kind case precisely because ‘water’ and ‘H₂O’ pick out different levels of description: surface features on one side, microstructure on the other. The gap opens because a world could present the surface features while hosting different microstructure (Twin Earth’s XYZ). The structural role definitions do not operate at two separable levels. ‘Causally structured succession’ names nothing except asymmetric directed succession among states: the full specification of what Time is. There is no surface/microstructure distinction to exploit, no layer of appearance hiding a different metaphysics underneath. The 2D gap requires exactly the appearance/nature separation the role definitions are designed to eliminate.

The water/H₂O example operates within an ordinary physical setting, but that does not dissolve the semantic distinction it illustrates. The relevant reply is narrower: role specifications and natural-kind identifications have different logical forms. Whether the book’s type-level identifications behave more like definitions or discoveries must be defended by the realizer arguments rather than inferred from the physical setting in which semantic evaluation occurs.

Third. The indexical anchor presses the objection furthest. The primary intension is evaluated from the perspective of a world considered as actual, and ‘considered as actual’ is an indexical that depends on the evaluator’s position rather than any mind-independent fact. The indexicality drives the 2D framework, the objector concludes, and if it survives so does the gap.

The indexical objection establishes an act-side point, not immediately an object-side result. Any concrete act of evaluating a primary intension unfolds through succession, distinction, and persistence. That does not by itself show that every primary-intension scenario instantiates the full temporal, spatial, and causal types. The object-side issue is semantic: whether the scenario is offered as a genuine world in W or merely as an epistemic representation. If it is a world in W, it answers to the Structural Requirement; if it is merely an epistemic representation, its variation does not by itself yield a metaphysical possibility.

Two lines support the claim rather than asserting it. First, Lewis individuates worlds as maximal fusions of spatiotemporally interrelated things (On the Plurality of Worlds,¹⁴ §1.6). That criterion requires the relevant worldmate relations where there are multiple inhabitants, but it does not by itself prove asymmetric temporal ordering in every sparse or changeless world. Lewisian individuation therefore supplies a concrete-world constraint without independently establishing the book’s full T-type. Second, the Chalmers/Jackson scrutability framework requires an idealized a priori reasoner conducting the evaluation exercise from within a world considered as actual (Constructing the World,¹⁵ Ch. 2). That idealization strips cognitive limitations (bounded memory, finite inferential steps) while preserving the structural conditions of reasoning: directed inference from premises to conclusions (T), distinction among propositions being evaluated (S), and persistence of the cognitive state across the derivation (Φ). Any actual token of idealized reasoning is performed through an ordered cognitive practice. That establishes an act-side condition on evaluators; it does not place a temporal coordinate inside every scenario they evaluate or settle whether an abstract ideal inference is itself temporally extended.

Fine’s¹⁶ distinction between essence and necessity sharpens the point. The indexical ‘actual’ is not merely necessarily co-present with a Time coordinate; it is essentially anchored in one. Any attempt to reconstruct an ‘actuality’ perspective from non-temporal resources either covertly reimports the succession from which ‘now’ acquires content, or fails to individuate one world from another at all. Chalmers’ own framework, applied to the structural role definitions, reinstates Time in the act of trying to prise meaning loose from metaphysics.

That the indexical reading of ‘actual’ is itself contested only tightens the anchor rather than loosening it. Van Inwagen’s ‘Indexicality and Actuality’¹⁷ argues the indexical theory is not true as Lewis states it: the words Lewis uses are ambiguous across several readings, on none of which the analysis is both genuinely indexical and correct; relatedly he presses that the indexical reading presupposes we inhabit just one possible world.¹⁸ Adams’s ‘Theories of Actuality’¹⁹ goes further and rejects indexicality outright: actuality is a primitive, absolute property that exactly one world possesses. The structural point is indifferent to which side prevails. If ‘actual’ is indexical, its token use is centered on a speaker or evaluator, but indexicality need not supply a single temporal frame spanning isolated worlds. If actuality is primitive and absolute, its application need not depend on a standpoint inside the actual world. On either account, actual evaluations are structured performances; neither account alone proves that every evaluated world instantiates T, S, and Φ. The bearing is on the theory of actuality, not on the inventory of worlds: nothing here counts the plurality down to one. It shows only that whichever account of ‘actual’ survives presupposes the structure the argument names.

A kindred deflation drops the 2D apparatus entirely for conceptual-role semantics: Ned Block²⁰ and Gilbert Harman²¹ argue meaning is exhausted by inferential role within a conceptual system, and the constitutive identity of §2.6 becomes one inferential-role equivalence among many rather than a modal truth. The reply is the same in thinner vocabulary. Inferential roles require reasoners whose inferences succeed each other (T), operate on distinct premises (S), and carry content across the derivation (Φ). Conceptual-role semantics has not replaced the structural roles; it has renamed them.

The adjacent modal objections are answered elsewhere. The Kripkean challenge (that necessary co-extension is not necessary identity) is handled at §4.1.1, which fixes the modal-epistemic status of the necessity claim and defers to this section the two-dimensional bookkeeping that status raises. The PSR-based modal objection is handled at §4.1.3. The modal-collapse objection (that necessary instantiation flattens modal distinctions) is handled at §4.4.9.

The bookkeeping, made explicit. The modal-epistemic status of □(T∧S∧Φ) is fixed at §4.1.1 as a priori and constitutive, and that section defers to this one the two-dimensional accounting that status invites. Here it is.

Run the derivation through the 2D matrix and the existential datum that change occurs (§2.4) is the clearest contingent input. The role definitions may be analytic at thin resolution, but the role-to-type bridges and the claim that they exhaust genuine change-realizers are substantive metaphysical premises, not analytic consequences of the vocabulary alone. Even that datum does not enter as ordinary empirical discovery; it is Cartesian-certain, surviving the strongest skeptical scenario (§2.2.1), so the lone contingent-primary element is not a posteriori in the water=H₂O way.

And the constitutive identity the datum feeds into (that the possibility of change just is the co-instantiation of the three roles, ◇C ⟷ T∧S∧Φ) has no primary/secondary gap to track. As the replies above establish, ‘causally structured succession’ is not scrutable from an independent base but is the base under another description, so its primary and secondary intensions collapse. That collapse is derived, not stipulated. A primary intension diverges from its secondary only where the concept holds some individuating feature independently of its secondary referent, a handle the primary keeps fixed while the referent varies, as ‘watery stuff’ keeps the surface profile while the microstructure shifts to XYZ.

The Modal Identity removes the handle. What §2.6.6 establishes is that between ‘causally structured succession’ and Time there is no individuating surplus in either direction: every specification of the one is already a specification of the other, with nothing left over. A primary intension therefore has no feature to track independently of the secondary, and no profile of its own with which to diverge. To conceive the referents coming apart, an ideal reasoner would need some feature on which they could differ; the no-surplus result is the absence of any such feature, and where there is nothing to differ on there is nothing left to conceive as different. That no-surplus result is itself fixed structurally at §2.6.6, established by the role definitions of §§1.3–1.4 together with the converse at §2.4.10, and it holds independently of any premise about what is conceivable. Leaning on it here therefore closes the primary-intension gap without circularity. Metaphysical no-surplus is in this way epistemic no-residue, and the gap closes at the level the objection needs it open.

With the fixed-domain modal lift and the constitutive arguments at the full Condition resolution, □(T∧S∧Φ) is argued a priori, not read off the world as an empirical identity. The point is not that S5 alone turns an existential datum into the structural conclusion; the Structural Requirement supplies the necessary conditional, and the type-level realizer arguments supply its content. The resulting necessity is structural and constitutive in the book’s intended sense. The 2D framework, asked to locate where modal weight sits, returns the verdict to the constitutive level it was built to prise open.

What converges. Across all three pressure points the 2D framework either requires the structure or redescribes it in more layered vocabulary. The two-intensions move needs modal space as its evaluation domain (which the roles constitute rather than presuppose) and finds no gap where the target is analytic rather than a posteriori. The water/H₂O analogy needs an appearance/microstructure seam the role definitions do not admit: ‘causally structured succession’ names nothing except what Time is. And the indexical anchor needs a Time coordinate from which ‘actual’ has determinate content at all: any attempt to reconstruct actuality from non-temporal resources either reimports the succession or fails to individuate worlds.

What stands, untouched by any of this, is Chalmers’ narrower point about natural-kind terms; what falls away is the attempt to import it to the constitutive level. The role definitions of §§1.3–1.4 are not natural-kind terms waiting to be pulled apart at a seam between surface and nature; they are role specifications whose content is exhausted by what they require. The 2D framework, turned on constitutive role definitions, does not discover a gap but reinstates the architecture it was designed to distinguish semantic layers from.

4.1.3 Principle of Sufficient Reason

The rationalist demand for sufficient reason is an exacting standard any necessary-existence argument has to meet on its own terms. Applied to T, S, and Φ, the demand takes three forms (a contrast demand, an origin demand, and Pruss’s modal refinement). The replies below grant the Leibnizian point freely. A rationalist argument that terminates in a brute fact has betrayed its own standard, and a conclusion terminating in brute necessity would betray the standard this argument was raised under. The argument does not so terminate. It terminates where the demand itself runs out of contrast.

G. W. Leibniz²² gives the demand its canonical form: nothing exists without a sufficient reason why it is so and not otherwise. Alexander R. Pruss²³ has given it the most rigorous contemporary defense: if every contingent fact has an explanation, the chain must terminate in something whose existence is not contingent. The demand is not parochial; it is the demand that reality be intelligible all the way down.

Applied to this argument, it presses that T∧S∧Φ obtaining necessarily is itself a fact about reality, and facts about reality are subject to the PSR. If the necessity is brute, the argument has violated its own commitment. If it is grounded, whatever grounds it stands outside or beneath structure as the true necessary ground. Three forms of the objection are taken up below.

First. The contrast demand presses that without a reason why the structure obtains rather than not, the argument has merely relocated bruteness from physics to metaphysics, and a brute necessity is not more intelligible than a brute contingency, only more deeply hidden.

The demand presupposes a contrast the argument has already foreclosed. ‘Why structure rather than not?’ supposes that ∅ is a coherent alternative whose non-obtaining requires explanation. As §2.7 establishes, ∅ is not: nothing individuates it, nothing distinguishes its obtaining from its non-obtaining, nothing for any description to attach to. The PSR demands a reason for obtaining rather than not obtaining; where the ‘not obtaining’ side is incoherent, the demand does not find a gap but fails to articulate one. The question presupposes a contrast its own terms cannot sustain.

Second. The origin demand grants the ∅ result but presses for a beginning: even if ∅ is incoherent as a state, the structure must have begun, and there must be some account of why T, S, and Φ obtain now rather than a prior arrangement.

The demand, on this analysis, is not unanswered but unaskable in the form it was issued in. A beginning is an event: a transition from non-existence to existence. Such a transition requires T, S, and Φ to take place: a temporal succession in which the prior state does not contain the structure and the subsequent state does, within a Space that can host both, carried by Substance. To propose a beginning for the structure is to propose a transition before transitions are possible, which is to describe nothing at all. The word ‘prior’ has no application outside the succession it names. The demand uses T, S, and Φ in its own formulation and then asks why they obtain, using the very conditions of the asking.

Third. Pruss’s modal refinement drops the demand for explanation of necessary facts and confines the PSR to contingent facts. The modal PSR does not require ∅ to be coherent as an alternative. It simply notes that every contingent fact admits of explanation, and asks whether the structure’s instantiation in this world is such a contingent fact.

The refinement survives, but it survives by conceding the conclusion. A PSR confined to contingent facts leaves the necessity of T, S, and Φ unchallenged: that is what ‘confined to contingent facts’ means. The structure is not in its scope. What the modal PSR captures are the causal sequences and events within the structure: why this reaction rather than that, why this memory rather than another. The present account (necessary structural conditions plus contingent investigation of everything within them) leaves every such question open to investigation. A theist who adopts the modal PSR is no longer arguing against the structure’s necessity but engaging contingent facts within the structure, the domain the present account is designed to accommodate.

One further engagement is owed at this site. Pruss’s central deployment of the modal PSR is not the abstract form addressed above (why T∧S∧Φ rather than ∅) but the Big Conjunctive Contingent Fact argument (The Principle of Sufficient Reason;²⁴ Necessary Existence with Rasmussen):²⁵ the conjunction of all contingent facts (the BCCF) is itself contingent and by modal PSR requires an explanation outside it.

On the book’s framework the BCCF is the conjunction of all contingent configurations of Nature: this cosmos, these initial conditions, this particular sequence of events. Each contingent fact within Nature has a contingent cause or explanation within Nature’s causal structure; the modal PSR is satisfied at that level. This is compatible with genuine chance and does not quietly presuppose determinism. Where a transition is governed by an objective probability distribution rather than a deterministic law, the distribution itself, fixed by the prior state and Nature’s causal structure, is what the modal PSR is owed. It is a reason why the outcome falls within a determinate range of possibilities, weighted in a determinate way, not a further hidden fact that secretly settles which member of the range is realized. To demand the latter is to demand that chance covertly be non-chance, which begs the question against genuine indeterminism rather than pressing a live PSR violation. This is exactly the room §4.4.9 draws on when it invokes quantum indeterminacy against modal collapse: an open token outcome is not a PSR debt left unpaid but the kind of contingent fact the modal PSR was always going to leave open once explanation is distinguished from determination.

The reply does not, however, rest on the Hume–Edwards principle that explaining each contingent conjunct thereby explains their conjunction, the very inference Pruss’s BCCF argument is built to reject. Its load-bearing move is instead the foreclosure of the external demand itself. The demand for an explanation outside the BCCF commonly motivates a necessary explainer, but the three candidates considered here — ∅, Nature itself, and a necessary being whose free act selects the BCCF — do not exhaust every theory of explanation by logic alone. They organize the principal options engaged by this book. ∅ is foreclosed. Nature does not explain the BCCF necessarily; if it did, the BCCF would be necessary, and contingent facts would collapse into necessities (the modal-collapse risk §4.4.9 addresses). The third candidate is the theist’s actual proposal — a necessary agent whose libertarian choice explains a contingent outcome without necessitating it, escaping the modal-collapse horn. It is engaged, not ignored: that a ground external to Nature is available at all is what §2.7 and the Supernatural Preclusion of Chapter 6 foreclose, while whether the necessary ground is personal and agent-causal is taken up on its own terms at §4.2.1, §5.1, and Chapter 6. What this site settles is narrower — the modal PSR does not by itself convert the demand for an external explainer into a demand for a necessary personal one; that inference is what those sections test, not one the PSR delivers here. Nature explains the structural preconditions under which contingent configurations obtain; the contingent selection among configurations is internal to Nature.

The selection-among-possibilities question (why this configuration rather than another) is addressed at §2.8.4 and §5.1: contingency is a property of configurations within Nature, not of Nature itself, and the demand that selection receive an external necessary explanation posits a grounding framework outside Nature. Section 2.7 forecloses ∅, but it does not by itself refute every proposed external ground; those proposals require the dedicated engagements in Chapters 5–6. This is the disambiguation the reply turns on: the BCCF is not granted an internal sufficient reason for the conjunction-as-a-whole (that would be the Hume–Edwards move Pruss is right to reject) but exempted from the external demand, because that demand has no coherent target. Each contingent conjunct keeps its contingent explanation within Nature’s causal structure; the conjunction as a totality is neither handed a further internal explanation nor left awaiting an external one, since each candidate external explainer is foreclosed as set out above.

The BCCF argument does not by itself establish a necessary personal explainer; further premises are needed to identify the explainer as personal and external to Nature. Conversely, the present section has not shown that no coherent totality-level explanatory demand can be raised. Its narrower reply is that local explanations within Nature and the necessity of Nature do not automatically generate the personal-explainer conclusion.

The dissolution carries further than the deflection of three demands. It shows that the demand was never available as a motivation for the argument to an external necessary ground in the first place. If ∅ is not a coherent alternative to the structure of Reality (§2.7), then the contingency of the structure cannot serve as a starting point for the argument to a necessary divine cause, because the structure has been shown to be not contingent but necessary. Its necessity renders the demand incoherent rather than redirecting it toward a divine answer. The PSR argument for theism is not answered by the Necessity of Nature; it is dissolved at its foundation.

A fourth move deserves naming. The Thomist may press that structural necessity is not the right kind of necessity: that classical theism demands the necessity of essence possessed by a being whose essence is its existence.²⁶ The observation is correct and the counter-construction is live, but it is not a fourth variant of the PSR; it is taken up on its own terms in §4.2.1 (a being outside the structure) and §5 (a being beneath it).

A fifth move presses a structural parallel rather than a PSR variant. If structure is self-grounding, why is the classical God not equally so? The causa sui construction holds that the divine nature is its own sufficient reason, requiring no external explanation, and the suggestion is that the present reading must concede the same status to whatever it proposes as the necessary ground.

The setup is incomplete on its face. Nature does not need an outside ground in the way that objects within Nature need their grounding conditions, because Nature is not an object that occupies anything outside itself. Nature is what occupying-anything consists in. The question ‘what makes Nature possible?’ assumes something outside Nature within which its possibility could be grounded, but that outside does not exist. To be possible is to be permitted by Nature. Nature does not permit itself from some external vantage point; it is what permission consists in.

The parallel fails because the two cases are structurally different in a precise way. Nature is not self-grounding in the sense that it is an entity which happens to satisfy the conditions for grounding from the inside. It is self-grounding in the stronger sense that it just is the conditions under which grounding operates: the logical space within which any explanatory demand, any sufficient reason, any grounding relation first becomes possible. There is no grounding framework outside Nature within which the question of Nature’s own ground could be posed, because the framework is Nature.

The distinction is sharp: God, on the causa sui construction, is proposed as a thing that grounds itself. Nature is not a thing. It is what things are in. A self-grounding thing still requires the conditions in which self-grounding is possible, and those conditions are Nature.

A constructive complement to the three deflections is available once the cyclic-recurrence result of §7.6 is in view. The BCCF paragraph above handles the modal PSR at the contingent-fact level via Nature’s internal causal structure; the cyclic-recurrence result closes two further pressures the contingent-fact reply leaves standing. The Kalām-style press for a first state at which the causal regress would otherwise terminate without a reason is foreclosed twice over. The structural closure of Temporality — □T together with ∅-preclusion and the domain-local reading of negation (§§2.3–2.4, §2.7) — leaves no coherent “before” or “outside” from which a first state could depart. The second closure is Singularity Preclusion (the Corollary of T6 at §7.6); a least moment of the manifold, if there is one, is a boundary with nothing on its far side, not an absolute beginning. This is not ad-hoc denial but the same structural argument applied at cosmological scale. The Leibnizian press for an external explainer of the cycle-as-totality is foreclosed because the cycle’s recurrence under invariant law is the modal necessity of □(T∧S∧Φ) playing out across past-unbounded time, not a further contingent fact requiring external grounding. The PSR is met on rationalist terms at both levels at which it can coherently be pressed against the present account: every event has a sufficient reason within the cycle, and the cycle’s recurrence is the structural necessity playing out rather than a contingency requiring an outside reason.

What closes. Three forms of the demand, and none finds a gap: the contrast demand presupposes ∅ as a coherent alternative §2.7 forecloses; the origin demand uses T, S, and Φ in the very act of asking why they obtain; and Pruss’s modal refinement, confined to contingent facts, leaves structural necessity outside its scope. Two further results follow once the deflections are in place. The dissolution removes the contingency-of-Reality starting point that would otherwise carry the PSR to an external necessary ground, and the causa sui parallel fails on the thing-versus-conditions distinction: structure is not a thing that grounds itself but what grounding consists in.

One last consideration, stated narrowly. The PSR is a live rationalist commitment this argument endorses, not a threat to be neutralized. The narrower Leibnizian point (that contingent facts admit of explanation) is not what the argument denies. What the argument denies is that the structural necessities are contingent facts in need of explanation against an available alternative. Humean contingency is taken up at §4.1.4: whether bare succession can be granted while the productive bond that makes it a succession is denied.

4.1.4 Humean Contingency

Hume’s challenge is a generous and disciplined attempt to thin Time. He grants that states follow states. He just does not see anything holding them together: one state, then the next, and the “then” does no work. No necessary connection, no causal arrow. Just sequence.

The replies here grant his narrowest claim (constant conjunction does not by itself entail necessity) and contest the broader one by testing whether each form can state succession without handing over directional succession through Substance. The structural-level engagement (defending the producing-relation against eliminativist pressure) is carried by §2.6.8; this section takes up the modal-contingency arm.

David Hume²⁷ argues that our impression of causal necessity is a psychological habit projected onto the world, not a feature discovered within it. His contemporary inheritors (Lewis,²⁸ B. Loewer,²⁹ Helen Beebee³⁰) sharpen the point into the Best System Account: laws of nature are the axioms of the simplest, strongest systematization of the pattern of events, not constraints that govern it. Succession without enforcement — order that simply obtains.

Time, as §1.4.1 defines it, is different. Not bare sequence: directional, asymmetric, productive, with the earlier state grounding the later, an arrow running through. Hume grants the sequence and denies the arrow. The three forms we are about to take up each miss the same thing: directional succession through Substance cannot be granted while the arrow is denied. The concession and the denial are the same sentence.

First. The bare-succession form holds that states can follow one another without any carrying medium. States follow states, the ordering is all Time needs, and nothing about Substance makes the next state follow the previous. Indexed order, not enforcement.

That concession is already enough, but note where it is enough. Try thinning succession further. Read “indexed order” as a block universe (B-theory; Mellor,³¹ Real Time II): time-slices arranged by index, all moments equally real, ordered by a tenseless before/after, no privileged now. On this reading, “persistence” just means qualitative similarity across slices; nothing is carried through. Can the thinner reading hold?

Consider what the thinning requires. An ordering of slices by index is one of two things. Either it is a function from integers to qualitative profiles applied to the manifold, in which case the manifold itself has to have enough structure to be indexable at all, an intrinsic distinction of positions that the arithmetic is tracking, not supplying. Or the ordering is free-floating abstract, a function unattached to any manifold, in which case it is not a time-series anymore, just a set of profiles paired with numerals, two disjoint structures sitting next to each other.

The first reading borrows the structural fact it is trying to replace. The second yields an abstract set of profiles paired with indices rather than a concrete temporal manifold. It does not collapse into ∅: the set still has members and formal structure. Its limitation is instead that the abstract pairing does not by itself realize temporal succession or causal persistence. Grant succession at any strength sufficient to state the position, and the Mutual Containment Result (§2.6.5) takes over: succession requires something that persists (Substance) across a structured interval (Space). Whether it is called “produced” or “enforced” is not the hinge; that something is sequenced at all is.

The bare-succession form is the wager that replacement is never more than mere sequence. Put exactly: replacement is not bare succession. It requires something that can be replaced, something that replaces it, and structure in which the replacement counts as replacement rather than mere juxtaposition. Strip the being, the interval, and the relation, and the Humean has not described thinner succession. They have removed the conditions under which succession is describable at all.

Second. The asymmetric-before form grants more ground than the first and tries to hold a thinner line. Sequence is real. Co-instantiation is real. But before-and-after is treated as a brute relational fact between states: the ordering is there, causation is not. Tim Maudlin (The Metaphysics Within Physics)³² goes further, making time’s arrow a fundamental primitive of the manifold: asymmetry that needs no further grounding.

Consider what this does. An asymmetric relation between states is either a label applied from outside (a convention we impose, with nothing in Nature answering to it) or a structural fact about Nature itself. If it is a label, it is a label for one of two things. Either it removes the ordering along with the arrow (leaving co-present states with nothing to index, the ∅ distribution §2.7 forecloses) or it keeps the ordering and imposes only the direction: the direction-conventionalist’s line, on which the series is real but no fact of Nature settles which way it runs. The second reading does not thin Time; it relocates the arrow to a mind-independent gradient, the thermodynamic move §4.2.3 takes up. Neither branch yields an asymmetry that is real, in Nature, and not already Exchange under another name. If it is structural (whether read as a consequence of deeper machinery or as a Maudlin-style primitive) then “asymmetric relation” and “Exchange” are doing exactly the same work. Two names, one structure.

Maudlin’s primitivism (the view that time’s directionality is a fundamental feature of the world, not reducible to anything more basic) refuses reduction in any direction. Time’s passage is not to be analyzed via entropy gradients, nor reduced to tenseless B-theory relations, nor deflated to psychological flow; it is a fundamental directional structure of the world that fundamental physics already presupposes whenever its laws are stated as differential equations evolving an initial state forward (The Metaphysics Within Physics,³³ ch. 4). Take the position on its own terms. The directionality is real, primitive, ontologically irreducible. It still attaches to states standing in directional relation across a manifold that bears the primitive. The primitive is the directionality of something: three structural conditions the primitivism cannot dissolve. Maudlin’s primitive is what T, S, and Φ jointly amount to once a defender insists on irreducibility. The role structure does not refute the primitivism. It states what the primitivism is the structure of.

Third. The Best System Account (BSA) is a sophisticated deflationary form of the challenge, and it is genuinely beautiful as a systematization. On BSA, laws of nature are the axioms of the simplest, strongest way to describe the total pattern, not constraints that govern it. Regularities get described, not enforced. The picture is the Humean mosaic: a distribution of local matters of fact, no necessary connections between them.³⁴

The standard non-Humean replies in the contemporary literature treat laws as relations among universals³⁵ or as primitive necessitating relations between properties.³⁶ The reply here is structurally different. It is not that something enforces the regularity above and beyond the distribution; it is that any vocabulary capable of stating the regularity already runs on the structural conditions the Humean meant to deflate.

Lewis’s doctrine of Humean supervenience says everything supervenes on the distribution of local qualities together with the spatiotemporal relations among their instances. Those relations belong to the supervenience base rather than being derived from the qualities alone. The mosaic is therefore distributed across an already structured manifold, not used to construct that manifold.³⁷ This is the load-bearing concession, and it runs under the whole account.

Consider what a regularity actually is. For events of type A to be reliably followed by events of type B, the Humean has already granted a distribution across a spatiotemporal manifold and a pattern stated over that distribution. That concession is enough. If the regularity is brute, the Humean offers systematization rather than a governing explanation; that is a deliberate feature of the view, not an internal contradiction. If the regularity is grounded in an invariant feature of the manifold, the account moves toward a non-Humean structure. The present argument can therefore claim that BSA presupposes spatiotemporal structure, but it cannot infer a productive causal arrow from regularity alone. A mosaic without a manifold is not a mosaic: it is a set with no ordering.

The BSA does not escape structure. It describes structure in Humean vocabulary, having already conceded the primitive spatiotemporal manifold that is the structure in question. The Humean wants regularity without necessity. Strip Nature and regularity has nothing to regulate.

A kindred deflationary move drops causal vocabulary entirely in favor of symmetric functional relations between variables: B. Russell³⁸ argues causation has no place in mature physics; J. D. Norton³⁹ carries the case forward. The reply is the same, shaped differently: functional relations between variables presuppose the variables’ distinctness (Space) and persistence (Substance), and any functional relation with a privileged direction of dependency presupposes Time’s asymmetry. Thermodynamic variants (that temporal asymmetry emerges from entropy gradients rather than obtaining structurally) come up in §4.2.3.

One narrower point, stated exactly. The self-refutation charge (that a denial of the structural conditions presupposes them in the act of denying) establishes the conditions’ actuality wherever the denial is performed, not their necessity across worlds where no such performance occurs. The move from actuality to necessity is the work of §§3–7 and does not rest here. The result is narrower: the deflationary vocabularies are parasitic on the structure they are meant to replace, wherever they are stated.

One rejoinder remains. A Humean might grant the parasitism and insist it is just how description works: any vocabulary capable of saying anything borrows structure from what it says it about, and that is no refutation of deflationism, just a feature of having a language at all. Look at the words the deflation reaches for. Holds. Runs. Deeper. Reach. Underneath. Ground. Foundation. Each one names the structure the deflation was meant to do without.

The reply runs at a deeper level than vocabulary. Hume published the Enquiry in 1748, and the argument has reached every reader since through a directed causal chain: ink pressed into paper, the marks holding their positions from Edinburgh to here, each transmission running forward and not back, copied and carried across 278 years with the content intact. The content of Hume’s position denies the productive bond; the vehicle by which that content has reached readers is the very causal structure the content denies.

The same reply reaches a closer case. Hume’s account of necessity runs from observation through habit to projected expectation: repeated conjunctions, formed habit, projected necessity. There is a class of cases where that order is reversed: cases where the regularity is established, in a single act, by an agent who specifies it. The Anscombean and interventionist traditions⁴⁰ have pressed a kindred line against the spectator’s epistemology for half a century: an agent who establishes a regularity by an intervention has first-person access to causal efficacy that no abstraction-from-conjunctions can supply. A programmer who specifies a function settles in advance what the function will return on inputs it has never been called on: no series of conjunctions, no time for habit to form, just a stipulation that holds because the system bears it. The constant conjunction is constituted by the stipulation, not abstracted from a history of it. The deflationary epistemology has no machinery for the reversed order, and the cases where the order runs that way are not exotic: they are the cases where regularities are most cleanly established. What the stipulation requires to take effect is the same triple the descriptive case required to be statable: a structural setting, a domain across which the stipulation ranges, and something the regularity holds across. The structure is met from the agent’s side as it was met from the describer’s side.

What converges. Time is the Condition §1.4.1 defines, bearing the succession role §1.1 extracts, and each deflationary form fails to remove it: bare succession cannot be indexed without borrowing the manifold it meant to replace (or thinning into an abstract pairing that realizes no succession at all); asymmetric before-and-after is either a label applied from outside, which removes the ordering or relocates the arrow (§4.2.3), or a structural fact about Nature, which is Exchange under another name; and the Best System Account admits the spatiotemporal manifold into its own supervenience base rather than deriving it, describing the structure in Humean vocabulary rather than replacing it.

Three forms, each conceding the structure they mean to deflate, and one rejoinder undone twice over: by the medium of its own transmission and by the cases its spectator’s epistemology cannot reach. The deflationary move reinstates the structure it meant to remove. The narrower Humean point (that constant conjunction does not entail necessity) is not what the argument contests; it is a constraint the argument accepts and obeys, refusing to read necessity off bare regularity and routing the necessity claim instead through structural analysis and the Cartesian datum that change occurs. That discipline is Hume’s bequest, inherited rather than overturned. The causal-transmission point is narrow and literal: the argument has reached us only through the structure its content tries to deflate. The next section takes up the information-theoretic challenge (§4.1.5): whether reality reduces to bits from which T, S, and Φ emerge rather than hold as structural conditions.

4.1.5 Information-Theoretic Foundationalism

Wheeler’s it from bit proposes that reality is informational all the way down: binary distinctions are fundamental, and spacetime and matter emerge from the relational structure of information processing. The picture has real purchase in physics (Shannon’s⁴¹ formalism, Landauer’s⁴² thermodynamic cost of erasure, information-theoretic reformulations of quantum mechanics) and none of it is in dispute. What is contested is the ontological inversion: the claim that information is more fundamental than the structural roles, that T, S, and Φ are emergent features of an information-processing substrate rather than conditions of what information consists in. T, S, and Φ are the three structural conditions whose necessity §2.8.5 declares, borne by Time, Space, and Substance. Three forms of the inversion follow. The replies test whether each can sustain the inversion without handing over what it meant to deflate.

First. The bit-first move holds that reality at the most fundamental level consists of bits (yes/no distinctions) with spacetime and matter emerging from their relational structure.⁴³

A bit requires difference to be a bit. Bateson’s⁴⁴ phrase names the contrast at the core: information concerns a difference that can matter within some system. But Shannon’s⁴⁵ formalism is deliberately abstract. Its sender, receiver, channel, and signal are formal roles in a communication model; the formalism by itself does not establish that every bit must have a physical bearer, spatial manifold, or temporal transmission. The stronger claim belongs only to physically realized information processing: where a difference is registered, propagated, or acted upon in a concrete system, there must be a persisting realization, distinguishable states, and an ordered process. That is a claim about realization, not a consequence of information theory alone. Information-theoretic foundationalism must therefore show how an abstract informational base is physically realized without already relying on the structural roles it proposes to derive.

Second. The Mathematical Universe Hypothesis⁴⁶ goes further: reality does not resemble mathematics but is a mathematical structure, and all mathematically consistent structures equally exist.

For a mathematical structure to be a structure at all, its elements must stand in determinate relations. That directly supports a role of differentiation. It does not follow from determinate relatedness alone that the relations are asymmetric, temporally ordered, continuous, or borne by a concrete substance. Those further roles enter only when the mathematical structure is claimed to be concretely realized or dynamically operative. Tegmark’s own account requires the mathematical structure to be the kind of thing whose elements stand in determinate relations; this concedes the difference-role at the level of what a mathematical structure consists in. The plenitude reading compounds the concession: every mathematically consistent structure supports that role under abstraction. The Platonist has not escaped differentiation, and the temporal and bearer roles fall due wherever the structure is claimed to be concretely real.

What the section’s account requires Tegmark to give up is not the mathematical-identity thesis as such but its unrestricted plenitude complement. The Mathematical Universe Hypothesis at Level IV holds that all mathematically consistent structures equally exist (Our Mathematical Universe).⁴⁷ If T∧S∧Φ are necessary conditions for a mathematical structure to be concretely real rather than merely abstractly consistent, then the Level IV plenitude is structurally bounded: only T∧S∧Φ-consistent structures are candidates for physical instantiation. The section’s claim is not that MUH is false but that it requires the structural conditions as a constraint on plenitude. Tegmark’s account can absorb T, S, and Φ as specifying what it takes for a consistent structure to be physically instantiated; the cost is giving up the claim that mathematical consistency alone is sufficient for physical existence.

Third. The computational-universe move presses the cleanest form. S. Lloyd⁴⁸ holds that the universe is a quantum computer: physical processes are computations, particles are data, interactions are logic gates. D. Deutsch⁴⁹ and S. Wolfram⁵⁰ develop parallel variants. The common commitment is that computation precedes the structural roles ontologically.

A physically realized computation instantiates each of the three roles at the level of the computing. For a bit to be processed in a concrete machine, there must be an ordered transition between states through a medium capable of realizing it; this supplies the relevant temporal role. An abstract computation, considered only as a mathematical mapping, does not by itself establish physical succession. For information to be held, maintained, and updated, something must persist as the being that bears the informational state, that is Substance. For bits to be distinct from one another, there must be relational non-identity among them, that is Space. A computation lacking any of the three is not a computation but something weaker: a static set of symbols (no Time), a dynamic flux with no being (no Substance), or a homogeneous stream with no distinguishable units (no Space). Landauer’s⁵¹ principle cuts the same way from physics: erasing a bit costs kT ln 2 of energy (the Landauer limit, where k is Boltzmann’s constant and T is temperature, the minimum thermodynamic cost of erasing one bit), and thermodynamic cost requires a physical substrate with temporal evolution, spatial distinctness of states, and persistent beings.

A kindred deflation drops the computational framing entirely for ontic structural realism: J. Ladyman et al.⁵² argue that only relational structure is fundamental and relata are ontologically dispensable. Structure, on this view, goes all the way down, with no bearer beneath it. This form of the inversion has to be met head-on rather than deflected, because the case against it is the case for Φ.

Structure is not a thing; structure is what things have. A pattern is always a pattern of something, as ownership is always ownership of something and a dent is always a dent in something. “Structure all the way down” asks us to keep the having while dismissing everything that has: to keep the floor plan after demolishing the building. But the structure of a building is not a second object standing alongside it; subtract the building and no structure is left over, only a blueprint, and a blueprint is a description of a building rather than a building.

The structural realist has a reply: the relata are nothing but nodes in the structure. A “thing” just is a position in the pattern, with no nature of its own. But a node is still a something. To be a position is to be the position of something positioned; relocating the bearer into the word “node” does not abolish it, it renames it. Nor does multiple realizability help: that one structure can be built in steel, wood, or concrete shows it independent of any particular material, not independent of material as such. A blueprint realizable in many media is still a sheet of paper until it is built in some medium or other. The dispute reduces to a single question: can anything be exhausted by its relations, with no nature of its own beneath how it relates? If not, then under every structure there is at least a minimal what-it-is that the structure is the structure of, and that is Φ. This is the intrinsic nature Russellian monism locates beneath physical structure (§4.3.3), and it is the precise point on which the present account parts from the structuralism of §4.3.9. Relational structure without anything across which relations obtain is ∅ (the Null State, the configuration in which none of T, S, or Φ is instantiated), foreclosed in §2.7.

The Concession Move. One last move remains to the opponent: not denial, but concession, that information is the substrate, and what the argument calls Time, Space, and Substance are already informational through and through. Look at the words information theory reaches for when it needs to name where information lives. Channel. Bearer. Register. Substrate. Carrier. Medium. Every one is a word for something. Information is descriptive. It needs something to describe.

Information requires distinction to be information; transition requires succession to be transition; processing requires persistence to be processing. Strip the roles and what remains is not a more fundamental informational substrate but ∅. Wheeler’s slogan, on the present account, runs in both directions: it comes from bit only where bit is already the output of T, S, and Φ doing their work. The unreasonable effectiveness of mathematics in describing physical structure (which it from bit implicitly trades on) is taken up directly at §4.3.6.

What converges. Five moves, five concessions. A bit requires difference to be a bit: a difference that makes a difference is already three, distinction (Space), being and medium (Substance), and directed registration (Time). A mathematical structure requires determinate relations among its elements and therefore supports the difference-role. The temporal and bearer roles require an additional realization premise; mathematical consistency or plenitude alone does not supply them. Computation instantiates each role at the level of the computing: temporal evolution, persistent beings, distinct units, with Landauer’s principle recording the thermodynamic cost of that instantiation. Ontic structural realism drops the bearer and keeps the pattern, and a pattern is a pattern of something: the node it relocates the bearer into is still a something. The retreat to information-as-substrate names every locale of information in structural vocabulary (channel, bearer, register, medium) and adds a fifth instance of the same concession.

4.1.6 Quinean Modal Skepticism

Quine’s modal skepticism is a thoroughgoing rejection of the modal vocabulary, and the position has not been refuted so much as outlasted by the literatures that find ways to work around it. “Two Dogmas of Empiricism”⁵³ attacks the analytic/synthetic distinction on which much prior modal regimentation depended; “Reference and Modality”⁵⁴ shows that quantifying into modal contexts produces the notorious failures of substitutivity that de re essentialism is supposed to license; Word and Object⁵⁵ turns the screw further with the indeterminacy of translation, on which essential properties cannot be picked out across translation manuals. The proposed cure is austerity: keep the language of regimented science (extensional first-order logic, logic sensitive only to reference and truth-value, not to meaning or modal force, plus the working vocabulary of physics and biology), drop the language of necessity, get on with the work.

Dagfinn Føllesdal⁵⁶ and John P. Burgess⁵⁷ carry the position forward in modal logic; Allan Gibbard⁵⁸ develops the contingent-identity wedge that presses against essentialism from within; Williamson’s Modal Logic as Metaphysics⁵⁹ is the contemporary acknowledgment that the position remains the floor any modal account must clear. Where the present argument needs de re necessities (that the structural roles obtain in virtue of what they are, not in virtue of how we describe them) Quine’s apparatus stands in the way.

The concession runs in full. Quine is right that mid-century modal regimentation imported analyticity as if Carnap had not been answered; right that quantification into modal contexts requires more than its proponents had supplied; right that translation does not pick out essences in any extensionally regimented way. The position closes more loose modal arguments than it opens, and any argument for necessary structural conditions has to clear the Quinean floor before it earns its conclusion. The Quinean tool has tracked something real about how modal vocabulary had been used; the question is whether it tracks something real about the structural conditions §2.6 establishes. Three forms of the Quinean move press the objection at three resolutions.

First. The classical analyticity form runs against the present account directly. Necessities, on this reading, are analytic in disguise: “Time is causally structured succession” is a stipulative truth about how we use “Time,” not a discovery about what any change-permitting world contains. Modal claims trade on hidden synonymies, and once the synonymies are exposed there is nothing left for necessity to track. Gibbard’s⁶⁰ contingent-identity wedge sharpens the pressure: even paradigm de re identities can be analyzed as description-relative if essentialism is given up, and the a posteriori necessities the Kripkean apparatus delivers (water = H₂O) are recast as terminological decisions about which features fix reference.

The reply runs on the distinction between stipulative definition and structural specification. The role definitions of §§1.3–1.4 are not synonymy claims about how the English word “Time” relates to the English phrase “causally structured succession”; they are specifications of what any medium hosting transitions must instantiate, derived from what a transition consists in. The derivation does not depend on the term “Time” picking out anything in particular; it shows that succession, distinction, and exchange are the conditions on transition under any vocabulary in which the conditions can be stated. Quine’s analyticity critique cuts against synonymy claims dressed up as discoveries; it leaves untouched discoveries that hold in any vocabulary capable of stating them. Gibbard’s wedge cuts against essentialism construed as a discovery about which descriptions a thing satisfies; the present account is not in that business. It does not claim that Time picks out the same role across worlds; it shows that any world containing transitions instantiates the role, by whatever name.

Second. The de re / de dicto form presses where the first leaves off. Even granting that the structural conditions are not synonymy claims, the Quinean denies that the de re reading on which they obtain “of the world rather than of a description” is intelligible at all. Quantifying into modal contexts produces the failures Quine⁶¹ catalogues: “necessarily, 9 is greater than 7” appears true, “9 = the number of planets” is true, “necessarily, the number of planets is greater than 7” is false. The de re reading on which a structural condition obtains of the medium, independent of how we describe it, is exactly the reading the substitutivity failures rule out. Føllesdal⁶² and Burgess⁶³ extend the point: modal logic is a useful instrument for organizing inferences but cannot, on Quinean grounds, be read off as picking out language-independent modal facts.

The reply concedes the substitutivity point and presses where it does not reach. The de re readings Quine’s catalogue dismisses are readings on which a contingent description (“the number of planets”) is treated as if it picked out its referent rigidly across modal contexts. The Kripkean apparatus (§4.1.1) answers the catalogue by distinguishing rigid from non-rigid designators; the present account does not depend on the answer. The structural roles are not picked out by a description that could have picked out something else; they are specified by what a transition requires, and the specification is invariant across the descriptions that satisfy it. There is no Quinean substitutivity failure to find because there is no substitution to perform: any term that satisfies the role-specification of Time picks out the same structural condition, and any term that fails to satisfy it picks out something else. The instrumentalist reading Burgess defends is consistent with this: modal logic is a useful instrument, and the structural conditions are what the instrument has to track to be useful. The instrument’s utility on Burgess’s own terms is the data the modal account explains.

Third. The extensional-regimentation form goes deepest. Even granting that the role-specifications hold in any vocabulary, the Quinean denies that the holding is modal in any sense that escapes extensional logic. What looks like necessity is the projectibility of certain predicates within our best science; what looks like de re essentialism is the privileged status of certain natural-kind classifications within that science. The work modal vocabulary did before is now done by what regimented science treats as nomological (law-governed, a matter of the laws of nature) and the question whether that treatment tracks something modally robust is exactly what Quine refuses to answer outside the framework. Williamson⁶⁴ presses the contemporary form: the position survives as a methodological stance even where its first-order claims have been answered, and any necessitarian account has to show that the necessities it posits are not artifacts of the science that names them.

The reply distinguishes scope, and it discharges the burden several times over rather than resting on one bridge. The Quinean is right that much of what passed for modal claim at mid-century was projectibility-talk dressed up; right that natural-kind classifications carry methodological privilege within a science without thereby carrying language-independent modal weight; right that the burden is on the necessitarian. What the reply does not do is the thing the Quinean would be entitled to reject: assert that projectibility simply is structural necessity renamed. Quine has a fully naturalized account of projectibility (entrenched quality-spaces, “Natural Kinds,”);⁶⁵ answering it by relabeling would beg the question. The conditions §2.6 identifies earn their standing instead by what any nomological practice must presuppose to be one at all, and several independent considerations converge on the same commitment.

The transcendental consideration. Actual projection occurs through succession and comparison among distinguishable cases, and Quine’s naturalized story invokes persistence and causal selection. This establishes the structure of the actual scientific practice. It does not show that projectibility requires productive law-like causation rather than Humean regularity, or by itself establish the full loaded types in every possible world.

The indispensability consideration. Quine’s own criterion fixes ontological commitment by what our best science must quantify over (§2.8.3). Any nomological practice quantifies ineliminably over a structured succession of distinguishable, causally interacting items; the structural conditions are not an optional modal overlay but what its bound variables range over. Quinean indispensability can commit the practice to whatever temporal, relational, and physical ontology its best regimentation genuinely quantifies over. Identifying that ontology exactly with the book’s T, S, and Φ requires the independent role and realizer arguments; the criterion alone does not supply the identification.

The invariance consideration. What keeps the conditions from being artifacts of the science that names them (Williamson’s⁶⁶ standing worry) is that every nomological practice presupposes them, not one. Rival sciences, future regimentations, and Quine’s own meta-theory alike must range over them; none can decline them and remain nomological. Projectible predicates vary from science to science; the structural conditions do not, because they are the conditions under which any predicate is projectible at all. This is the thin framework-independence of §4.1.8: not a privileged carving beyond all frameworks, but conditions no framework can refuse.

The three considerations support a cumulative reply; none alone converts actual scientific practice into an unrestricted cross-world necessity. The work necessity did before is now done by what our best science treats as nomological, and what any nomological practice operates on, quantifies over, and cannot vary without ceasing to be nomological, is the structural conditions named.

A kindred move drops Quine’s program for Williamson’s necessitism,⁶⁷ which keeps the modal vocabulary while denying that contingency is fundamental; that position is engaged at §4.4.6. A different skepticism keeps both the vocabulary and ordinary modal knowledge while denying that the modal faculty is reliable at metaphysical distances;⁶⁸ because it concedes everything the present section vindicates, it is engaged separately at §4.1.13. The adjacent deflationary tradition, modal conventionalism (§4.1.7) and metaontological deflationism (§4.1.8), runs parallel deflations at neighboring resolutions, and the structural reply takes the same form across the family. The regimentation of the structural claims in formal terms that satisfy Quine’s extensionalist constraints is carried out at §§2.10.1–2.10.3, showing that the role-specifications are statable without the analytic/synthetic distinction and without quantifying into modal contexts in the way Quine’s (1953) catalogue targets; the bound-variable ontology that licenses the existential commitments is laid out specifically at §2.8.3.

What closes. Three forms of Quinean austerity, and none opens the gap: the classical analyticity form cuts against synonymy dressed up as discovery and leaves structural specifications untouched; the de re / de dicto form catalogues substitutivity failures for contingent descriptions and leaves role-specifications, which admit no substitution, intact; and the extensional-regimentation form relocates necessity into nomological practice, reinstating the structural conditions as what any such practice operates on. The Quinean refuses the modal vocabulary; the present account shows what the vocabulary picks out, and what it picks out is unaltered by the refusal. However the austerity is run (as analyticity-skepticism, de re skepticism, or extensional regimentation) what each form contests is □(T∧S∧Φ) at the a priori / constitutive register, supported by the fixed-domain S5 lift from the Cartesian witness together with the separate Structural Requirement and role-to-type premises, the T5/N register line §4.1.1 records and §4.1.2 makes explicit.

The austerity has done genuine work, as the concession above records. What the austerity has not produced is a configuration of reality in which the structural conditions fail to obtain, and what the catalogue has not produced is a substitution failure that touches role-specification rather than description. The skepticism, taken seriously, narrows to the claim that necessity is an unnecessary word. The argument concedes the possibility and asks what remains when the word is gone. What remains is structure.

4.1.7 Modal Conventionalism

Alan Sidelle’s Necessity, Essence, and Individuation⁶⁹ develops the most systematic contemporary defense of the view that necessity is a feature of how we describe the world rather than of the world itself. The argument descends from logical positivism (Ayer’s Language, Truth and Logic⁷⁰ on analyticity, Carnap’s “Empiricism, Semantics, and Ontology”⁷¹ on framework-relativity, Quine’s⁷² “Two Dogmas” as the critical pressure the tradition had to absorb) and Sidelle’s contribution is to extend the analyticity treatment to de re modal claims (necessities about a thing itself rather than about its description). Saying water is necessarily H₂O is, on this picture, a decision about how to use “water” as a natural-kind term rather than a discovery about water.

The position has been pressed in different directions since. Sidelle’s “Modality and Objects”⁷³ develops the view in response to its critics. Ross Cameron’s “On the Source of Necessity”⁷⁴ and his subsequent work argue that necessity is grounded in linguistic and conceptual practice without being reducible to it. Amie Thomasson’s Norms and Necessity⁷⁵ defends an adjacent modal normativism on which modal claims express norms governing concept-use rather than describing language-independent facts. Against the conventionalist tradition, Bob Hale’s Necessary Beings⁷⁶ develops the most sustained contemporary critique, and Jonathan Schaffer’s necessitarian counter (“Laws for Metaphysical Explanation,”)⁷⁷ presses from a different direction.

The argument here does not contest the analyticity of selected modal claims. It contests the claim that all of them are. The structural conditions T, S, and Φ are not stipulated into necessity by linguistic decision; they are the conditions any stipulator must operate within in order to stipulate. (The formal regimentation of these structural claims in terms that satisfy Quine’s extensionalist constraints is carried out at §§2.10.1–2.10.3; the bound-variable ontology is at §2.8.3.) Any actual proposal and comparison is a structured cognitive and linguistic performance. That secures an act-side point about actual convention-making, not yet the object-side conclusion that every possible world instantiates the full temporal, spatial, and causal types. The global conventionalist still owes an account of the non-conventional resources used to formulate and apply the thesis. Three forms of the conventionalist move sharpen the reply at three resolutions.

First. The classical Sidelle⁷⁸ form treats de re modal claims as analytic in disguise. “Water is necessarily H₂O” is not a discovery about water; it is a stipulation about how “water” is to be used in modal contexts, with the analyticity hidden by the de re surface grammar. The argument generalizes: every necessity is, on examination, a hidden analyticity, and the appearance of language-independent modal fact is the product of unrecognized linguistic decision.

The reply is internal. Stipulations are themselves entities, sequences of marks, sounds, or mental events whose individuation depends on extension, persistence, and contrast with other stipulations. Conventions are configurations within Reality, not grounds prior to it. To say “necessity is conventional” is to use a medium whose conditions are not conventional in any sense the position can recover. Sidelle anticipates the move and replies that the medium itself is conventional all the way down, but the regress closes the position rather than vindicating it. If “medium” is conventional, then the convention that fixes “medium” as conventional is itself a configuration, requiring the very conditions it purports to deflate. Quine’s regress presses that conventions cannot alone generate all logical consequence, since applying them requires inferential resources. The structural account may use this against a global conventional reduction, but specifying those resources as the full T, S, and Φ types requires further argument rather than following from the regress itself. Hale’s Necessary Beings⁷⁹ presses this point at length: the conventionalist account requires modal vocabulary to state, and the modal vocabulary it requires is not stateable in the conventionalist’s preferred terms.

Second. Cameron’s⁸⁰ refinement concedes the medium’s structural conditions and presses the thesis at a higher level. Necessity is grounded in linguistic and conceptual practice in the sense that what makes modal claims true is the practice of using modal vocabulary in the way we use it, not in the sense that modal claims are reducible to claims about the practice. The grounding relation is asymmetric: modal facts hold in virtue of practices, but modal facts are not identical to facts about practices. The position survives the regress because it does not deny that practices have structural conditions; it grounds modality in practices and is silent on the structural conditions of practices.

The concession is real. The reply is that Cameron’s position succeeds at the cost of being indistinguishable from the present account at the level where the present argument’s claim applies. A concrete discursive practice ordinarily involves continuing practitioners, distinguishable acts, and temporal ordering. Those are conditions on such practices; calling them unrestricted modal conditions, and identifying them with the full loaded types, depends on the book’s separate structural derivation. The grounding Cameron offers (modality grounded in practice) terminates at conditions on practice that are themselves necessary in every world in which practice exists. The present account names those conditions; Cameron’s account leaves them implicit. The remaining disagreement is about whether the structural conditions are best described as conditions on linguistic practice in particular (Cameron) or on any configuration capable of hosting practice or anything else (the present argument).

Third. Thomasson’s⁸¹ modal normativism is a refined version. Modal claims do not describe modal facts; they express norms governing concept-use. “Water is necessarily H₂O” does not say anything about water; it endorses a norm about how the concept water is to be applied across counterfactual cases. The position is genuinely deflationary about modal facts without being committed to the regress-prone claim that modal claims are reducible to claims about language.

The reply tracks Thomasson’s own framework. Actual norm-governed practices distinguish applications and often maintain standards across time. An abstract norm need not itself persist as a physical bearer or be applied in temporal sequence, so the practice-side structure should not be transferred automatically to the norm as an abstract object. Thomasson’s normativism does not deny these conditions; it is silent on them, treating them as background to the normative analysis rather than as the locus of modal commitment. The present account makes them explicit. The conditions of a norm’s being in force (proposed, applied, maintained in a practice) are what the structural roles secure, and any normativist account that disowns those conditions disowns the apparatus the account requires.

A kindred move presses from outside conventionalism entirely. Schaffer’s⁸² necessitarianism argues that the laws governing metaphysical explanation are themselves necessary, but the necessity is not conventional: it is grounded in the structure of explanation itself. The present account agrees with Schaffer that the necessity is not conventional, and adds a constructive specification of what the necessity is grounded in: the structural conditions any configuration capable of being something must instantiate. Where Schaffer leaves the grounding at “the structure of explanation,” the present account specifies the structure.

Two adjacent positions live elsewhere. Carnapian framework-deflationism is engaged at §4.1.8 (Metaontological Deflationism). Modal fictionalism, which deflates modal commitment by relocating it to a fiction rather than to convention, is engaged at §4.4.4.

What converges. The conventionalist tradition has refined itself across thirty-seven years, from Sidelle’s classical analyticity-extension (unstable on its own articulation, since stipulations have structural conditions it cannot recover) through Cameron’s grounding-asymmetry (consistent with the structural account at the level of conditions on practice) to Thomasson’s normativism (consistent at the level of conditions on norms). However the conventionalism is pitched (as analyticity-extension, grounding-asymmetry, or normativism) each form locates the source of □(T∧S∧Φ) in a derivative register (convention, practice, norm), while the register at which the modality holds is a priori / constitutive, supported by the fixed-domain S5 lift from the Cartesian witness together with the independent Structural Requirement and role-to-type premises, the T5/N register line §4.1.1 records and §4.1.2 makes explicit.

Sidelle’s 1989 book was written when the analyticity treatment of modality looked, after Quine’s “Two Dogmas,” like a project that had not survived the mid-century. The intervening decades have been a long argument that selected analyticities can be defended even if Quine’s general critique stands. The argument has succeeded in part: there are conventional necessities, and the analyticity of selected modal claims is a stable position. What the long argument has not produced is a global conventionalism that survives examination of its own articulation. The conditions for the convention are not conventional. The structural roles are what any convention presupposes.

4.1.8 Metaontological Deflationism

A family of contemporary metaontological positions argues that disputes in fundamental ontology are verbal rather than substantive. Different ontological vocabularies are different languages, equally adequate to the world; the choice between them is pragmatic, not factual.

The lineage runs from Rudolf Carnap’s “Empiricism, Semantics, and Ontology”⁸³ (distinguishing internal questions answerable within a linguistic framework from external questions of pragmatic framework-choice) through Eli Hirsch’s quantifier variance (the view that different ontological frameworks have different quantifier-meanings, each equally adequate to reality; Quantifier Variance and Realism,⁸⁴ collecting work from 2002 onward) to Amie Thomasson’s easy ontology (Ontology Made Easy).⁸⁵ David Chalmers’s “Ontological Anti-Realism” (in Metametaphysics: New Essays on the Foundations of Ontology, ed. Chalmers, Manley and Wasserman)⁸⁶ develops a more refined version on which there is no fact of the matter about absolute existence, while internal-to-framework existence is unproblematic. Matti Eklund’s “The Picture of Reality as an Amorphous Lump” (in Sider, Hawthorne and Zimmerman’s Contemporary Debates in Metaphysics) defends a related but distinct position on which different carving-schemes are equivalent descriptions of a structurally underdetermined reality. On this picture, asking whether T, S, and Φ are necessary is asking the wrong kind of question; there is no language-independent fact of the matter for an answer to track.

The deflationist’s neutrality is not neutral. Different ontological vocabularies must be vocabularies of something: different ways of carving up a shared reality. If there were no common reality, there would be nothing for distinct vocabularies to be alternative descriptions of. The deflationist’s framework presupposes that this reality has whatever conditions are required for distinct descriptions to be applied to the same thing: extension over which carvings can vary, persistence for the same matter to be re-described, distinction between description and described. These are the structural conditions under different names. (“Configuration” throughout this reply names whatever jointly instantiates the three structural roles T, S, and Φ as derived at §2.6; it is shorthand for that joint-instantiation condition rather than an additional ontological category. The formal regimentation of the structural claims in terms that satisfy Quine’s extensionalist constraints is carried out at §§2.10.1–2.10.3.) Three forms of the deflationary move sharpen what each version commits to.

First. The Carnapian framework form treats ontological questions as internal to a chosen linguistic framework. “Are there numbers?” is well-formed inside the framework of arithmetic and answerable trivially in the affirmative; outside the framework it is a pragmatic question about whether to adopt the framework. Hirsch’s quantifier variance generalizes the move: different frameworks have different quantifier-meanings, each equally adequate to reality. There is no privileged sense of “exists” against which framework-choice can be assessed.

The reply is internal to the position’s own apparatus. For two frameworks to be alternative carvings of a shared reality, the shared reality must have whatever properties license distinct alternative carvings: at minimum, sufficient extension to be carved differently (S), persistence across the alternative carvings (Φ), and the temporal ordering of carving-events that lets one framework be adopted, then another, then evaluated comparatively (T). Hirsch concedes that frameworks must be alternatives of something but holds that what they are alternatives of can be specified only within some framework. The concession is enough. If the something has whatever properties are required to be the something the frameworks are alternatives of, those properties are the structural conditions, and they obtain whether or not any framework names them.

This is the thin claim, not the heavyweight one: not that some framework-transcendent carving is metaphysically privileged, but that whatever a framework must quantify over in order to be a framework at all already instantiates the roles. On Quine’s bound-variable criterion (§2.8.3; with the later formal discussion at §2.10.2), the commitment is read off what any framework-comparison must range over; the conditions are framework-independent in the sense that no framework can decline them, not in the sense of a privileged carving lying behind every framework and reachable from none.

That the roles are Reality’s and not the framework’s is earned below, on the deflationist’s own criterion, once the deepest form of the position is in view. Carnap’s “external questions” are pragmatic only on the assumption that there is no fact of the matter outside the framework. The assumption is the position’s central claim, and the assumption is what the present argument contests. It is a contest whose founding form is Quine’s, who rejects any principled internal/external distinction in “On Carnap’s Views on Ontology”⁸⁷ on the strength of the ontological-commitment criterion of “On What There Is.”⁸⁸ What the present account adds to that Quinean realism is the specification of the conditions a substrate must satisfy to be the something rival frameworks are alternatives of.

Second. Thomasson’s⁸⁹ easy ontology is the most carefully developed contemporary version. Existence questions are answerable trivially given conceptual competence: “Are there tables?” follows from “There are particles arranged tablewise” plus the application conditions of “table”; “Are there numbers?” follows from “There are five apples” plus the application conditions of “number”. The position is deflationary about ontology while remaining committed to the trivial truth of ordinary existence claims. There are no hard ontological questions because the apparent hardness reflects only confusion about application conditions.

The reply tracks Thomasson’s own framework. Application conditions are themselves features that must be specified: a concept’s application conditions persist across applications (Φ), distinguish cases of application from cases of non-application (S), and operate across the temporal succession of applications (T). Thomasson’s account is consistent with the present account at the level where the present account’s claim applies: the application conditions are themselves features of a configured reality, and the conditions of their being application conditions are the structural roles. Thomasson’s deflationism succeeds at deflating ontology in the sense she means; it does not deflate the conditions on which the deflationary apparatus operates. The remaining disagreement is whether the structural conditions are themselves substantive ontological commitments (the present view) or further easy commitments derivable from concept-use (a possible Thomassonian reply). The present argument takes the Triconditional machinery of §2.6 (especially the Mutual Containment Result at §2.6.5) as the case for the first reading: the structural roles mutually require one another and cannot be derived from any prior conceptual base.

Third. The Chalmers⁹⁰ and Eklund⁹¹ positions occupy philosophically refined ground. Chalmers’s ontological anti-realism distinguishes lightweight quantifiers (used in ordinary discourse, indifferent to deep ontological commitment) from heavyweight quantifiers (used in fundamental ontology, subject to framework-relativity). The position holds that heavyweight ontological questions have no determinate answer while lightweight ones have trivial answers. Eklund’s position treats reality as a structurally underdetermined “amorphous lump” that admits of multiple equivalent carvings, with no carving privileged over others. Both positions sharpen the deflationary tradition by making it consistent with the trivial truth of ordinary existence claims.

The reply concedes the lightweight/heavyweight distinction and presses where the heavyweight claim sits. Chalmers’s position requires that there be lightweight-but-real configurations (otherwise lightweight existence claims would not be trivially true) and that no heavyweight fact about absolute existence obtains. The present account’s claim is not a heavyweight claim about a privileged absolute carving; it is a claim about what any configuration capable of being something (at any weight) must instantiate. If Chalmers’s lightweight configurations include extension, persistence, and temporal succession (and they must, on pain of being indistinguishable from no configuration at all), then the structural conditions obtain at the lightweight level Chalmers grants. Eklund’s amorphous lump faces the same pressure. The more charitable reading of the position is that the lump is underdetermined by privileged carvings rather than structureless: reality admits multiple equivalent carving-schemes, none of which is the metaphysically privileged one. The reply does not require treating the lump as structureless. Even reality underdetermined-by-privileged-carvings must still support three things. It must support the framework-comparison the position invokes (different carvings must be distinguishable from one another and from the lump they carve). It must support the application of each carving to the lump (each scheme must operate on something that persists as what it operates on). And it must support the temporal succession of carving-acts by which one scheme is adopted, evaluated, and replaced by another. Those three are the structural conditions under different names, and they obtain at whichever level reality is configured to admit alternative carvings at all. The ∅ fork then sharpens what the position itself commits to: a lump structureless enough to admit any carving whatever is a lump indistinguishable from ∅ (§2.7); a lump with sufficient structure to admit some carvings but not others is a lump with the structural conditions named.

The position can be sharpened across all three forms into the claim that every ontological commitment is internal to a framework. The sharpening is unstable. The claim itself is either internal to a framework (in which case it is one ontological commitment among others, not the metalevel verdict it presents itself as) or external, in which case it asserts a framework-independent fact about the relation between frameworks and the world, contradicting its own deflationism. There is no third position from which the verdict can be issued without using the apparatus the verdict denies. Thomasson’s reply⁹² (that the metaontological claim is itself trivially true given the application conditions of metaontological vocabulary) relocates the instability rather than dissolving it: the application conditions of metaontological vocabulary are themselves features of a configured reality, and the conditions of their being application conditions are what the present argument names.

The retreat to the framework. One reply remains. Grant everything so far: that no framework can decline the three roles, that every carving and every comparison of carvings already deploys succession, distinction, and exchange. The deflationist can still say this fixes only a fact about frameworks (about what any scheme must presuppose to be a scheme) and not a fact about Reality. That the roles are presupposed by all our descriptions is one thing; that they are Reality’s own structure, owing nothing to the describing, is another, and the transcendental result delivers only the first. The line is Stroud’s: that we cannot think past a structure shows how we must think, not how things stand.

The retreat has no ground to stand on, and the reason is internal to the position once more. The deflationary apparatus rests on a single criterion of what it is for something to be objective rather than framework-relative: to be objective is to be invariant under the choice of framework, and to be merely framework-relative is to vary with it.

Quantifier variance is the thesis that what exists varies with the quantifier-meaning a framework supplies; the lightweight/heavyweight distinction is the distinction between what holds inside a chosen scheme and what would hold across all of them; the verdict that no carving is privileged just is the claim that every candidate joint varies across admissible frameworks. Strip the position of the invariance-criterion and it cannot state its own thesis: “privileged” means framework-transcendent, and framework-transcendence is measured as invariance. The criterion is not optional equipment; it is what gives the deflationary verdict its content.

Apply it. The three roles do not vary across frameworks; they are what every framework presupposes, the fixed point under the operation adopt a different scheme. Framework-invariance supplies evidence that the roles are not artifacts of any one scheme. It does not, without a further bridge, entail that they are mind-independent features of Reality; a deflationist may treat them as constraints on every admissible framework. The dispute therefore narrows to whether invariance is merely transcendental or genuinely ontological. The present account argues for the latter, but it should not present that bridge as forced by deflationism alone. The transcendental fact, read through the criterion the deflationist must hold in order to speak at all, counts as transcendent by the deflationist’s own measure.

This is the heavyweight claim in the only sense the opening allows it. Reality is fixed at §1.1 as a structural-role term (what any content presupposes, its reference set by what the role requires and not by which thing fills it). The constitutive language there names the structural shape of the primitive modal field, not an essence prior to it. The present conclusion claims exactly that and no more: the roles belong to Reality in that they are invariant across every scheme in which anything is configured at all, not in the sense of an essence lying behind all configuration and reachable from none.

The deflationist who demands the second sense before conceding the first asks for a target §1.1 already declines to posit. The demand purchases nothing: a structure that does no work in any framework and is reachable from none is, by the ∅ fork, indistinguishable from the absence of structure (§2.7). What §1.1.1 leaves conditional is supported here but not discharged by invariance alone: what is invariant across languages is internal to no single language, yet may still be characterized by a deflationist as a universal condition on representation rather than as mind-independent ontology.

One regress is owed an answer. The deflationist can lift the retreat a level: the claim that the roles are invariant across all frameworks is itself advanced within a framework (a meta-framework that individuates schemes and ranges over them) so what is shown is invariance within that meta-framework, not invariance as such. But the trilemma is indifferent to level. The meta-claim that one has shown only invariance-within-M is itself either internal to some further scheme, in which case it varies across schemes and is, by the criterion, non-objective, and so cannot deliver the objective verdict it pronounces; or it is external, a framework-transcendent fact about the relation between meta-frameworks and the roles, which is the thing the position forbids itself.

Nor does the regress escape the roles by climbing: any meta-framework that individuates frameworks and ranges over them is itself carving. It requires distinction to hold its frameworks apart, exchange to compare them, succession to adopt one and then another. Each ascent instantiates the roles it is meant to rise above. The invariance is therefore not invariance at a level but invariance the hierarchy cannot exit; the regress, far from undercutting the result, is its demonstration.

A kindred move worth marking. Modal conventionalism (§4.1.7) runs a parallel deflationary argument at the level of modal claims rather than existence claims; the structural reply takes the same form. Modal fictionalism (§4.4.4) deflates modal commitment by relocating it to a fiction; again the structural conditions return at the level of the storytelling. The deflationary family across these three sections shares a common structure and a common vulnerability: each version succeeds at deflating its target only by leaving the conditions on its deflationary apparatus implicit, and each, on examination, presupposes the structural roles at the level its deflation operates.

What converges. Three forms of metaontological deflationism (Carnapian framework-relativity, Thomassonian easy ontology, Chalmers–Eklund refined deflationism), and each, fully developed, is consistent with the present account at the level its claim applies: the Carnapian form requires a substrate frameworks are alternatives of, Thomasson’s form requires application conditions, and the Chalmers–Eklund form requires lightweight-but-real configurations or a structurally non-trivial lump. Each of these already has the structural conditions. However the deflationism is pitched, each form denies □(T∧S∧Φ) a framework-transcendent register, and each is answered at the same fixed point: the register is a priori / constitutive, the register line fixed at §4.1.1–§4.1.2.

The deflationist who presses anyway operates within the structural conditions while declaring them undecidable. The declaration is a configuration within Reality, and so is the framework that hosts it. Both fall within the scope of the argument they contest.

4.1.9 Wittgensteinian Quietism

Wittgensteinian quietism comes in distinct varieties that share a common stance: philosophical theses, including metaphysical theses about the structural ground of reality, are not properly to be argued for or against but to be diagnosed. The varieties differ in what diagnosis amounts to.

The Tractarian Wittgenstein,⁹³ and the resolute reading developed by Cora Diamond⁹⁴ and James Conant⁹⁵ on which the Tractatus’s philosophical sentences are themselves nonsense to be climbed and discarded, holds that metaphysical sentences fail to express propositions. They are not false; they lack truth-conditions. The proper response is silence: to climb the ladder of philosophical analysis and then throw it away.

The later, Investigations-derived varieties operate differently. Hacker⁹⁶ defends the orthodox Wittgensteinian view that metaphysical claims are grammatical confusions: when we ask whether time is real, we have misread the grammar of real and projected an answerable question where none lives. John McDowell⁹⁷ develops a therapeutic quietism in epistemology and philosophy of mind: philosophy’s task is not to construct theories but to dissolve the appearance of problems by clarifying the conceptual framework that generated them. Stanley Cavell⁹⁸ goes further: skepticism cannot be refuted by argument but only acknowledged, and the criteria for our concepts come from forms of life rather than from any standpoint outside them. On all three readings, a sustained metaphysical argument like the present one is not engaging a question the quietist takes to be live; it is exhibiting a confusion the quietist takes its task to clarify.

The challenge is sharper than ordinary skepticism. The quietist does not argue that the present argument is wrong. The quietist argues that it is not the kind of thing that can be wrong, because it is not the kind of thing that can be right.

The reply has three parts, addressing the varieties in turn.

First. The Tractarian framework distinguishes saying from showing: certain features of the world cannot be stated within language but can be exhibited by what language does. Time, Space, and Substance, on the present argument, are precisely such features. Every act of language exhibits them, ‘exhibits’ here in the sense fixed by the ∅-fork at §2.7: any act that is not ∅ instantiates the structural conditions, their joint instantiation being what counts as there being an act at all. The argument is not the proposition T, S, and Φ are necessary; the argument is the structural exhibition of what utterance, distinction, and persistence require. A Tractarian who grants showing has granted what the argument needs; the resolute reader, who discards showing along with the rest of the ladder, is met not here but by the performative point below, since the climbing and discarding are themselves enacted.

Second. The Hacker–Baker line says metaphysical questions are misuses of language; McDowell says they are appearances generated by a conceptual framework whose clarification dissolves them. Both readings require a coherent distinction between misuse and use, or between confused appearance and clarified perception. Drawing the distinction is itself an activity. It occurs over time, differentiates one form of words from another, and persists through the diagnostician. The diagnostic activity exhibits T, S, and Φ no less than any first-order metaphysical claim does. The therapy does not undo what makes the therapy possible.

Third. Cavell’s deepest move is that philosophical claims rest on criteria that come from forms of life, and that a form of life is not a thesis but a way of being-in-language. The present argument does not contest this. It identifies what any form of life (any way of using criteria, any practice of acknowledgment) requires. The structural conditions are not criteria offered from a standpoint outside the form of life. They are what occupying any form of life involves. Cavell’s acknowledgment of skepticism is itself a sustained activity; the activity has the structure the argument identifies; the acknowledgment does not float free of what makes acknowledgment possible.

The grammatical reading, owned. The quietist presses one step further: if the only thing that would count against ‘every act exhibits T, S, and Φ’ is ∅ (the absence of any act at all) then the claim has no contrastive use among actual cases, and a claim with no contrastive use is grammatical rather than substantive. The response is not to resist this but to accept it, on one condition. The claim is indeed constitutive: it does not report a feature some acts have and others lack; it states what counts as there being an act at all. But this concedes nothing to quietism, whose charge is that metaphysical claims are confusions, pseudo-questions to be dissolved. A constitutive truth whose only alternative is ∅ is the opposite of a confusion. The grammar is not idle. It marks the one boundary that does not fall between cases but between there being cases and there being nothing at all: the line between something and ∅. To call that “merely grammatical” is to mistake the floor of the world for a turn of phrase. The claim earns a constitutive register precisely because its proposed countercase is absolute structural absence rather than another ordinary act. Still, the exhibition of the roles in actual speech or diagnosis establishes an act-side result. The cross-world □(T∧S∧Φ) conclusion depends on the modal and role-to-type arguments developed elsewhere; quietist performance alone does not supply it.

What converges. The dialectical point extends across all the varieties considered here. Where refusing to articulate an objection is itself a performed act of declination, it involves persistence, distinction, and an ordered episode. Mere absence of speech, however, is not automatically such an act and does not by itself instantiate the full triad. The performative point applies to enacted quietism, not to silence considered simply as no tokened response. Quietism, enacted, is the position that exhibits structure most economically.

4.1.10 The Source of Possibility

The PSR demand is met at §4.1.3 by showing that Nature is not a thing that grounds itself but what grounding consists in. A residual question remains: can an entity within the modal domain be explanatorily prior to real possibility, conferring content on the global range W that ◇ quantifies over? The answer runs by disjunctive elimination on the candidate’s own modal status. It is deliberately narrower than the claim that every possible essence-prior or primitive grounding account has been refuted by contradiction.

The position pressed against is the family of grounding accounts that take possibility as a derivative phenomenon: neo-Fregean abstractionism on modal grounds,⁹⁹ Lewisian reduction of modality to concrete possible worlds,¹⁰⁰ Williamsonian necessitism,¹⁰¹ dispositionalist grounding of possibility in the potentialities of actual objects,¹⁰² engaged on its own terms at §4.1.12, and the theistic moves that locate modality in a divine intellect or divine nature;¹⁰³ the last of these, pressed as an argument from modality’s need of a ground to God’s existence, is engaged on its own terms at §5.1.8, whereas the result established here runs purely at the level of any candidate’s modal status.

Definition. Let source of possibility name any X such that ◇p (for some non-empty range of p) holds in virtue of X. The “in virtue of” relation is left maximally general: causal, metaphysical, conceptual, intellectual, or explanatory grounding all qualify. The argument does not turn on which grounding relation is in view.

Modal Exclusivity · Premises

1. If X is a source of possibility, then ◇p holds in virtue of X (Definition).

2. Every candidate X has exactly one modal status: ¬◇X (impossible: true in no world), ◇X ∧ ¬□X (contingent: true in at least one world but not in every world), or □X (necessary: true in every world). The three cases are mutually exclusive and jointly exhaustive.

The impossible case. ¬◇X. If X is impossible, X has no instances and no relata. Nothing is the case in virtue of what is not the case. The impossible is not a source but the absence of one.

The contingent case. ◇X ∧ ¬□X. If X is possible without being necessary, ◇X is itself a fact within the modal field. The proposed source of ◇ presupposes ◇ in the specification of its own modal status. The field is not grounded by one of its own features.

The necessary case. □X. This horn must be met in its strongest form, the one that grants the most. That X is necessary does not by itself bar it from being explanatorily prior: grounding priority and modal status are distinct, and an item may be necessary without being explanatorily inert — essence need not reduce to any modal fact,¹⁰⁴ and the necessitist may hold X necessary yet grounding.¹⁰⁵ Concede the orthogonality and the modal classification alone does not close the case. Describing X as necessary uses □, but it does not follow that X’s nature or grounding activity is ontologically posterior to the modal field; a Finean essentialist may hold that □X is true because of X’s essence. What the classification shows is narrower: the modal description □X cannot itself explain the content of ◇ without circularity. Excluding X as an essence-prior source requires an additional theory of grounding, not the modal trilemma alone. The point is not that necessity precludes priority, but that priority to ◇ is unavailable to anything whose own status is already modal.

Modal Exclusivity, scoped. No candidate’s modal status by itself explains the modal domain: impossible and contingent candidates fail as global sources, while a necessary candidate requires an additional, non-modal account of how its essence or grounding activity confers modal content. The trilemma exposes a circularity burden; it does not refute every essence-prior source by modal classification alone.

The self-grounding rejoinder. A defender may concede the trilemma and posit that possibility grounds itself, that ◇ is its own ground. Two readings exhaust the move. On the deflationary reading, “self-grounding” is a notational variant of primitivism: ◇ is not derived from anything more basic, and the Principle stands. On the substantive reading, grounding is a reflexive relation on ◇, and the field grounds itself. Here the analytic literature on ground divides the relation in two.¹⁰⁶ Strict ground is irreflexive: nothing strictly grounds itself.¹⁰⁷ So if the rejoinder means that ◇ strictly grounds ◇, irreflexivity rules it out directly. Weak ground is reflexive by construction (every fact weakly grounds itself) so ◇ does weakly ground ◇; but weak self-grounding is just the absence of any strict ground, which is primitivism stated in the grounding idiom rather than an alternative to it. Either way ◇ has no strict source. This is also what accommodates the minority dissent that admits reflexive or non-well-founded grounding:¹⁰⁸ that dissent is a thesis about weak ground and the failure of well-foundedness, not a denial of strict-ground irreflexivity, so it is absorbed by the fork rather than set aside as a bare commitment. The substantive reading does not escape the trilemma; it renames its floor.

The classical-theism rejoinder. Aquinas-style classical theism does not propose God as a source of possibility distinct from God; on divine simplicity, God is the modal field (ipsum esse subsistens applied to modality). The move sidesteps the trilemma by collapsing the source/sourced distinction. The collapse delivers one of two results. On the relabeling reading, God renames “the primitive modal field,” in which case the Principle stands and the relabeling is theological vocabulary rather than a modal-foundations counterexample. On the predicate-bearing reading, the modal field carries predicates classical theism attributes to God (will, intellect, agency), which is the engagement §5.2 and §5.1 take up, not a counterexample to the present Principle.

The transcendence rejoinder. A defender may place X beyond modality altogether: neither possible, necessary, nor impossible. That proposal is not refuted merely by repeating the trilemma, because the trilemma classifies candidates whose modal status is already in view. The present result instead marks a burden. If X is said to explain W, the account must explain how an entity outside real possibility bears an explanatory relation to the domain of real possibilities without thereby being specified as one more item in that relation. A classical-theist or Finean essentialist may regard that relation as primitive or essence-prior. The chapter does not close that option by contradiction. It shows that it is no longer a modal-semantic explanation of possibility: it is an additional metaphysical posit whose credentials must be argued in its own theory of grounding. The theological version is engaged at §5.1.8; the essence-priority question remains the live gloss identified below.

The same closure bears on the question §2.6.1 and §3.5.3 defer to this section: whether the realizer biconditionals, □(T↔︎(S∧Φ)) and □(Φ↔︎(T∧S)), state a structural-functional necessity or a Finean real definition under another name. The two readings must be prised apart from what depends on them. The type-upgrades of §3.5.3 need only that the biconditionals are necessary, and they are necessary on either reading — a real definition is necessary if anything is — so □(T↔︎(S∧Φ)) transmits its necessity to the spatial and persistence types under modus ponens whichever reading is correct. The type-layer therefore closes even for a reader who keeps the Finean essence-prior gloss; nothing in the structural conclusion is hostage to the outcome.

What the elimination above underwrites is the further claim the modal-monist commitment (§2.8.6) makes: that this necessity is not seated in the essence of an item standing prior to the modal field. That claim is not secured by declaring every standing a modal standing — the Finean is entitled to hold that essence, were it prior, would ground modal facts rather than stand among them. The ‘occupying a role is itself a possibility’ reply used against a source of ◇ above does not carry over here: there the role at issue was the sourcing of ◇ itself, so standing in it was intra-modal by construction, whereas a realizer biconditional makes no claim to source ◇, and importing the move would beg the question. It is secured only to the extent the trilemma applies: any candidate put forward as the essence-bearing ground has its own modal status (□X, ◇X, or ¬◇X) and is to that extent specified from within the field rather than before it, leaving the essence-prior reading with no definiendum the elimination has not already located inside ◇.

A committed essentialist can decline that last step; if so, the disagreement is over the gloss on a necessity both sides grant, not over the necessity the biconditionals transmit. The structural-functional reading is the one this account adopts and the one the modal monism of §2.8.6 requires: the biconditionals record what any medium hosting succession must be, the structural feature reality is rather than an essence some entity has.

Nor does the presupposition improve on iteration. Let the defender universalize: every possibility is sourced. Then the sourcing (itself a modal fact) is a further possibility, and by the defender’s own thesis it requires a source; so, in turn, does the possibility of that. The chain forks. Either some link goes un-sourced, which concedes the Principle outright, or none does, and the sourcing relation never steps outside the field it was meant to ground. The regress is vicious not because it fails to terminate but because no stage of it is ever prior to ◇; the explanatory debt the account incurred is never discharged. Its one exit (that ◇ grounds its own possibility, closing the chain in a loop) is already foreclosed: a strict-grounding loop is barred by the irreflexivity of strict ground, and a weak self-grounding loop collapses into primitivism, as the self-grounding rejoinder established. What stands is a dilemma whose floor is the one-step presupposition: primitivism, or a regress that never leaves the field, with self-grounding shut.

The exotic placement of X buys no exemption from this. The via negativa is addressed at §5.2.2, on its own register. The Principle holds at the relation level against any candidate placeable at any level.

What is established. No candidate whose own status is stated in modal terms can be a thing-style source of the modal domain it already inhabits. The impossible and contingent horns fail directly; the necessary horn remains inside the domain whose source it was meant to supply. The result blocks internal modal grounds — divine intellect, concrete worlds, abstractions, or necessary entities specified as modal occupants. It does not settle whether a committed essentialist or transcendence theorist can posit an essence-prior ground; that is a further metaphysical cost and a live dispute, not a contradiction the present trilemma alone supplies. The Necessity of Nature identifies the structural content of the global domain; Modal Exclusivity establishes that no modal item within it confers that domain from outside. Its lateral counterpart, Modal Closure (§2.8.4), records that no unnecessary alternative stands alongside it from outside.

The impossibility pole. The same elimination runs on any candidate source of impossibility. Let a source of impossibility be any X such that ¬◇p holds in virtue of X; the trilemma on X’s own modal status is unchanged. An impossible X has no relata in virtue of which anything could hold. A contingent X is itself a fact within the field and presupposes the modality it would confer. A necessary X cannot explain impossibility merely by being described with □; whether its essence can ground modal exclusion remains the same live grounding question as on the possibility side. Modal Exclusivity is therefore bilateral only as a diagnostic: modal status alone does not source either ◇ or ¬◇. The stronger claim that both poles are primitive requires the structural-functional account adopted here, not the trilemma by itself. The field is sourceless at both poles. This closes the residual voluntarist retreat that grants possibility its primitiveness but relocates the necessary truths, or the standing of the impossible, in a divine intellect or will (the created-truths tradition; theistic activism on the necessary truths, Leftow¹⁰⁹); that inference is met on its own terms at §5.1.8, while the result here holds at the level of any candidate’s modal status. What Way–Possibility Convertibility (§3.2.5) records on the possible side has its dual on the impossible side: to be impossible is for there to be no way to be, and the absence of a way is not conferred by an antecedent ground but is, like its positive twin, a structural feature reality is rather than has.

4.1.11 Pizzi’s Impossibility of Nothingness

David Pizzi’s The Impossibility of Nothingness: A Foundational Argument for Necessary Existence¹¹⁰ is the closest contemporary cousin of the impossibility-driven recovery this argument runs across the lines of §3.2–§3.3 and the convergence tallied at §3.5, and at §5.1.1. The dialectical situation is symmetric and instructive: Pizzi and the present account converge on the same modal floor (that absolute nothingness is impossible and that something must exist necessarily) by closely related performative and self-refutational machinery. They diverge on what sits at that floor. Pizzi concludes that something exists necessarily (□∃x x) and explicitly declines to arbitrate among the candidates that could occupy the necessary slot. The present account concludes that the structural Triconditions T ∧ S ∧ Φ exist necessarily (□(T ∧ S ∧ Φ)) and arbitrates among the candidates Pizzi leaves open. The same final move (closing modal space against ∅) yields a bare-existential conclusion in Pizzi and a determinate-content conclusion here.

Pizzi’s argument runs three reductios against the metaphysical possibility of total absence and bridges them through a single theorem. The first reductio targets the truth-maker logic of nothing exists: given truthmaker maximalism, if that proposition were possibly true it would need a truth-maker; the proposed absence cannot stand in the truth-making relation without being something. The result therefore depends on the truthmaker requirement rather than following from the proposition alone. Quantification over absences therefore collapses. The second reductio targets representations of the proposition: actual assertion and formal modeling presuppose syntactic and semantic resources. That is a strong act-side or model-side result, though an opponent may deny that resources used to represent a scenario must exist inside the represented scenario. The third reductio targets empty-domain semantics. Whether actuality operators force non-emptiness depends on the quantified modal semantics adopted, so this line constrains particular formalisms rather than settling every empty-world model by definition. The bridging theorem assembles the three: since each path forecloses the possibility that nothing exists, by the modal duality ¬◇¬X ≡ □X it follows that something exists necessarily. The argument runs at the same logical strength as the convergence of lines catalogued at §3.2–§3.3 and tallied at §3.5: classical modal duality, no S5, no positive theology, no conceivability premise outside the structural-self-defeat asymmetry the conceivability-inversion line develops at §3.3.4.

A note on formalization for the Pizzi-mode reader. Pizzi’s three reductios run in formal modal-logical mode and converge on □∃x x via classical modal duality. That result is significant but deliberately thin: it establishes necessary existence without identifying the necessary content.

The present argument reaches □(T∧S∧Φ) by a different line. It needs the fixed-domain modal theorem, the Cartesian witness of change, and the Structural Requirement that any change-permitting world realizes the Conditions. The move from □∃x x to □◇C is not licensed by modal duality alone, and the present account should not attribute such a T-completeness lemma to Pizzi’s result. A nonempty world might contain abstract or otherwise thin content without thereby realizing genuine change. The structural conclusion therefore requires the independent role- and type-level arguments Chapters 2–3 develop.

The boundary is worth marking in the object language as well. The □ in □(T∧S∧Φ) ranges over worlds, while T is a within-world structural condition, so the box governs whether the Triconditions obtain at each world, not any temporal ordering internal to one of them.

The methodological convergence is close enough to require an account of what the present argument does that Pizzi does not. Two answers. First, Pizzi’s three paths each target the assertion or modeling of ∅: what happens when one tries to state or represent the empty case. The negation line of §3.3.2 closes a parallel gap by targeting the operation that would produce ∅ from a populated case: stripping structure away. This line is not among Pizzi’s three; he does not run an introduction-by-subtraction argument, since his framework treats ∅ as a candidate state whose semantics must be evaluated, not as a state that would have to be reached by subtracting structure from somewhere. The negation line therefore complements rather than duplicates Pizzi’s three. Second, Pizzi’s three reductios establish ¬◇∅ at the level of the bare-existential conclusion; the present argument runs that same closure across the convergent lines of §3.2–§3.3 (tallied at §3.5) and then names the necessary content. The naming is what closes Pizzi’s §7.

Pizzi’s §7 surveys the candidates for the necessary existent his argument has established: abstracta (Platonism), substratum (priority monism), modal-realist plurality, theistic models, and Vedāntic Brahman. He notes that the argument is silent on which candidate occupies the slot, citing this silence as a virtue of the argument’s neutrality. The present account agrees that the argument to □∃x x is silent on the slot’s occupant, and disagrees that the silence is virtuous: there is a further argument to run, and §5 is where it runs.

The candidates require different verdicts. Abstracta may be coherent as abstracta; the present argument does not infer from their atemporality that they are impossible. The relevant question is whether they can ground the global domain of real possibility or the concrete, change-realizing world. That requires an additional grounding relation, and the abstracta theorist must supply its content. Priority monism may identify the cosmos as a partless whole prior to its parts. The present account can grant that priority while asking whether the whole, insofar as it realizes change and grounds configurations, instantiates the structural roles. The verdict is not that every candidate simply collapses into T∧S∧Φ, but that none receives automatic entitlement to the grounding office merely from Pizzi’s □∃x x conclusion.

Modal-realist plurality supplies concrete worlds subject to Lewis’s worldmate criterion. Change-rich Lewis worlds realize the structural roles, while sparse or changeless candidates remain tests for the book’s independent necessity and type-level arguments; and the fictionalism line in §3.3.3 picks this up against modal fictionalism’s milder cousin. Theistic models split on the classical/temporal distinction taken up at §5.2.1 and §5.3.3. A classical model that denies the loaded types may still posit determinate grounding or dependence roles; Chapter 5 must test whether those can remain wholly outside the structure rather than simply defining the model as ∅. Temporal, process, and open-theist models more directly instantiate the roles. Vedāntic Brahman likewise divides into nirguna and saguna readings at §5.2.2; the burden is to determine whether the former is genuinely contentless or instead carries thin determinacy, while the latter more openly carries structure. In each case the structural account adds a further question Pizzi’s result leaves open: does the candidate realize, ground, or merely coexist with the structural roles required by a concrete change-realizing Reality? Some candidates may be compatible abstracta or metaphysical pictures; none is selected as the necessary ground by □∃x x alone. The Chapter 5 engagements supply differentiated verdicts rather than a simple reduction of every candidate to one option.

The closing chapter of The Impossibility of Nothingness (§8.3) places Pizzi’s conclusion in a theist-friendly axiological frame, citing Adams¹¹¹ on the goodness of the existence of God as the contemporary anchor for the move from “something exists necessarily” to “the necessary existent is divine.” Pizzi flags this as a gesture rather than an argument, and is explicit at §8.2 that the formal result of the book is silent on theism. The present account takes the disclaimer at face value. Already engaged at §5.1.1 is the contemporary necessary-existence cluster¹¹² as a methodologically continuous neighborhood that nonetheless does not reach this account’s conclusion. The axiological gesture is a separate move from the modal-existence argument, and one this account does not run. T ∧ S ∧ Φ are not maximally great in the value-theoretic sense; they are necessary because nothing else is a something. The axiological closing is available to anyone who has reached □∃x x and wants to identify the necessary existent with a value-maximal being; the present account reaches the stronger and value-neutral □(T ∧ S ∧ Φ), and the axiological move does not gain traction on it.

What is established. Pizzi’s dialectical placement is therefore distinctive. He sits here among the modal and semantic foundations rather than among the §5 engagements, because those sections take up positions that disagree with the present account about the form of the necessary ground: classical theism (§5.2.1), Swinburne’s personal-God theism (§5.1.4), Spinozism (§5.3.1), Neoplatonism (§5.3.2), Plantinga’s modal ontological argument (§5.1.7), and so on. Pizzi disagrees with none of the formal commitments. He runs the same kind of argument to a strictly weaker conclusion and leaves the further work undone. The reply is correspondingly cooperative: this section catalogs the convergence, names where Pizzi stops, and points to the sections that pick up the work he sets down. Necessary Existence¹¹³ shares Pizzi’s bare-existential terminus and runs by formal-modal apparatus. Rasmussen’s How Reason Can Lead to God,¹¹⁴ engaged at §5.1.1, pushes past the bare-existential terminus by stage-by-stage construction toward a personal necessary being. The cluster is the closest the contemporary literature comes to the present argument, and the gap between □∃x x and □(T ∧ S ∧ Φ) is Chapter 5’s central piece of work.

4.1.12 Dispositionalism about Modality

Dispositionalism is the nearest rival the account has, and it is nearest because it agrees with the account about the thing most other positions deny. It holds that possibility is grounded in actual reality rather than conferred by a plurality of worlds. Where Lewis¹¹⁵ makes ◇p true by the existence of a concrete world at which p obtains, the dispositionalist makes ◇p true by a property of something actual: a potentiality, a power, a disposition. Vetter¹¹⁶ gives the position its book-length statement, grounding modality in object-level potentialities and dispensing with worlds as explanatorily basic; Borghini, Andrea; Williams, Neil E.¹¹⁷ and Jonathan D. Jacobs¹¹⁸ run the article-length versions; Pruss¹¹⁹ develops a related actualist grounding. Behind them stands the dispositional essentialism of Alexander Bird¹²⁰ and Stephen Mumford,¹²¹ with Mumford, Stephen; Anjum, Rani Lill¹²² on the causal side.

A distinction has to be drawn before the engagement starts, because two different theses travel under one name. Dispositionalism about laws holds that laws are the structural expressions of the dispositional essences of natural kinds. That thesis is engaged at §2.8.1, where it is one of four law-accounts the constitutive claim survives, and the engagement there is complete. Dispositionalism about modality is a different and more ambitious thesis: not that laws are grounded in powers, but that ◇ itself is. It is that thesis this section takes up, and it presses at the point §4.1.10 is defended.

The concession runs in full. The dispositionalist is right that a plurality of concrete worlds is an extravagant truthmaker for a modest claim; right that modal truths should answer to what is actually the case rather than to what is elsewhere the case; right that the possible-worlds apparatus is a formal convenience whose metaphysical reading was never obligatory. On the negative program the present account and the dispositionalist are allies, and §4.4.1 is where that alliance does its work. The disagreement is entirely about the positive one.

First. The reduction form runs directly at the Convertibility. Its claim is that ◇p holds if and only if some actual object has a potentiality for p, so possibility is not a structural feature of reality but a summary of what things can do. The Way–Possibility Convertibility (§3.2.5) is then a redescription of the powers facts rather than a rival to them, and the structural triad is read off potentialities rather than presupposed by them.

The reply is that a potentiality already exhibits the triad, and exhibits it in the form the decomposition specifies. A potentiality is a directedness: it runs from a bearer, through a medium of manifestation, to an outcome that is not yet the case. That is an ordering with an asymmetry (the T-role), a distinctness between latency and manifestation without which nothing has been directed at anything (the S-role), and something that has the potentiality rather than the potentiality free-floating (the Φ-role). §2.8.1 already makes the point on the laws side: a disposition without a persisting bearer, without relational extension through which it manifests, and without an interval between latency and activation is not a disposition. The dispositionalist has not produced an alternative to the way-structure. A potentiality is a way, individuated as one, and taking it as primitive does not remove what individuates it.

The reply is deliberately not the stronger one available. It does not say that necessitation is directional production, or that the manifestation relation supplies the content of T. That would give the structural conditions a second, analogical sense, which §4.2.4 forecloses. The claim is the deflationary one: the potentiality is described in the structural vocabulary because it instantiates the structure, not because the structure is a species of causal production.

Second. The primitiveness form concedes the exhibition and contests the direction of explanation. Grant that potentialities display direction, difference, and bearing; the dispositionalist may hold that potentiality is metaphysically primitive and the triad is a true description of primitives rather than a condition on them. Description does not establish priority.

The reply asks which side survives the removal of the other. A potentiality stripped of its directedness is not an undirected potentiality but no potentiality; strip the distinctness of latency from manifestation and there is nothing the potentiality is for; strip the bearer and there is nothing whose potentiality it is. The triad is therefore not a description that happens to fit but the individuation-condition of the primitive, and an individuation-condition is not posterior to what it individuates. The structural claim also takes a weaker form than the dispositionalist’s own. The Convertibility runs as a convertibility, with neither possibility nor way reducing to the other, whereas dispositionalism is a reduction in one direction and therefore owes an account of its reduction base. What sits in that base is a property of an actual object, and a property of an actual object is specified from within the modal field in exactly the sense the Modal Exclusivity trilemma at §4.1.10 identifies.

Third. The coverage form is where the position is under most pressure from inside its own literature, and where the disagreement with the present account becomes visible rather than verbal. Jennifer Wang¹²³ presses that dispositionalism cannot reach possibilities for which nothing actual seats the relevant potentiality: alien properties, radically different physical regimes, possibilities concerning how things stood before there was anything with powers of the requisite kind. Vetter¹²⁴ answers with iterated and joint potentialities, extending the base by chaining what actual things can bring about.

The reply is that the internal dispute is settled, whichever way it goes, by asking whether there is something actual to seat the potentiality, and that question concedes the structural point rather than answering it. Both parties agree that possibility must be seated in actual reality. They differ only over whether the seat is a property of an object or the conditions any object-involving state already instantiates. At ordinary scope the difference is hard to see, because objects are available to bear the powers. At global scope the difference is decisive.

The Bearer Problem · Premises

1. Dispositionalism about modality: for any p, ◇p holds in virtue of the potentiality of some actual object (Reduction).

2. A potentiality is borne. There is no potentiality that is not the potentiality of something (Bearing).

3. Consider the possibility that there be objects at all.

4. By (1) and (2), that possibility holds in virtue of the potentiality of an object.

5. So the possibility that there be bearers presupposes a bearer, and the account has supplied its own manifestation-condition as its ground.

6. The dispositionalist must therefore either accept the circle at (5) or exempt the limit case from (1), which concedes that not all possibility is dispositionally grounded.

Neither horn is available at low cost, and the structural reading takes the second one without paying. What seats the possibility that there be objects is not a further object but the conditions any object-involving state instantiates: succession, distinction, and exchange are not properties an item has but the structure any item-involving state is. The dispositionalist’s seat runs out at precisely the scope the argument of Chapter 2 operates at, and it runs out for a reason internal to the position rather than imported against it: powers are borne, and the global case has nothing to bear them.

A rejoinder is available and should be recorded. The dispositionalist may deny that there being objects at all is a genuine possibility requiring a ground, holding instead that it is necessary, so that no potentiality need seat it. That reply is not obviously wrong, but it is a concession in the relevant sense: it grants that something is necessary without being dispositionally grounded, and once one such necessity is admitted the reduction is no longer general. It also moves the dispositionalist onto the terrain of §2.8.4 and §4.1.10, where the question is no longer what powers ground but what a necessity that grounds nothing further can be. Williamson¹²⁵ reaches a structurally similar place from the opposite direction, and the answer §4.4.6 gives applies here.

What closes. Dispositionalism is not refuted as an account of laws, and §2.8.1 already grants it what it is owed there. As an account of modality it does not bypass the structural requirement but instantiates it, since a potentiality is individuated as a directed, differentiated, borne thing and is therefore a way rather than an alternative to ways. The order of explanation runs from the individuation-condition to the primitive, not the reverse. And the reduction does not reach the global case, because the possibility of there being bearers cannot itself be seated in a bearer. What the position establishes, and what the present account keeps, is that modality answers to actual structure rather than to a plurality of worlds. Where it locates that structure in the powers of things, the account locates it in what any thing with powers already presupposes.

4.1.13 Modal Epistemology and the Remoteness Objection

Quine denies that the modal vocabulary is intelligible. Van Inwagen grants it and denies something else: that the faculty by which we reach modal verdicts is reliable at the distances philosophy asks it to work at. Inwagen¹²⁶ is the canonical statement. Ordinary modal knowledge is secure, on this view, because it operates by small variation on the actual: the table could have stood two feet to the left, the meeting could have run long. Remote modal claims (that there could be a being than which none greater can be conceived, that there could be a physical duplicate of a person lacking consciousness, that transparent iron is possible) are cases where the faculty has been carried far past anything that calibrates it, and the confident verdicts philosophers return are not knowledge.

This is not the skepticism of §4.1.6, and the two must not be run together. The Quinean refuses the vocabulary. Van Inwagen keeps the vocabulary, keeps ordinary modal knowledge, and refuses the extension. A reply that vindicates the intelligibility of necessarily leaves his objection entirely untouched, which is why it needs its own address. Its compositional half is discharged at §4.3.11; the half taken up here is the one that section explicitly defers.

The objection lands with more force here than almost anywhere, because □(T∧S∧Φ) is about as remote as a modal claim gets. It quantifies over every world, including worlds bearing no resemblance to anything experienced, and it returns a verdict of necessity rather than mere pervasiveness. By van Inwagen’s own criterion of similarity to ordinary life, no claim in the book is further from the calibrated range.

The concession runs in full, and it is larger than the concessions elsewhere in §4.1. Van Inwagen is right that a great deal of analytic metaphysics has run on brute intuitions about distant scenarios; right that no one has produced a credible account of a faculty that would deliver such verdicts reliably; right that the modal ontological argument and the zombie argument both turn on exactly this move; right that the burden falls on whoever wants a remote modal conclusion. §3.3.4 already agrees, and the agreement there is not tactical. The book’s own treatment of conceivability is closer to van Inwagen’s than to his opponents’.

First. The remoteness form runs directly. Ordinary modal knowledge proceeds by imaginative variation from the actual, the reach of that method is short, and □(T∧S∧Φ) lies far outside it. So whatever warrant the conclusion has, it cannot be the warrant ordinary modal claims have.

The reply is that the conclusion is not an output of that method, and the point is methodological rather than defensive. The argument nowhere asks the reader to imagine a world and report what is found there. It asks the reader to specify one and observe what the specification requires. Those are different operations, and only the second is publicly checkable. §2.1.5 is the clearest case: it does not invite the objector to picture a world lacking the structural conditions, it invites the objector to describe one, and the argument turns on what the description turns out to contain. An invitation to describe can be accepted or declined in a way an invitation to imagine cannot, and its results can be inspected by someone who does not share the imaginer’s phenomenology. Van Inwagen’s skepticism is calibrated against a faculty; the derivation does not run on that faculty.

The inversion is worth stating plainly, because it is the book’s general shape. Where the standard method treats I can conceive of it as evidence for possibility, the method here treats the attempt to specify it consumes what it denies as evidence for impossibility. The first is a report about the person; the second is a result about the attempt.

Second. The parity form presses that the reply has not escaped anything. The premises of the derivation are themselves modal claims (that a way must have direction, difference, and bearer; that no world can lack the conditions), and modal claims are what the skepticism targets. Relocating the warrant from imagination to specification changes the vocabulary of the defense without changing what is being defended.

The reply distinguishes two ways a modal verdict can be reached. One surveys possibilities and reports what is found among them, and that method requires epistemic access to the possibilities surveyed, which at remote distances is exactly what van Inwagen denies is available. The other attempts to construct a specification and finds that the construction cannot be completed, and that method requires no access to anything remote. The attempt is made here, with materials available here, and its failure is observed here. That no way and impossible are two names for one verdict is the surface expression of the point: a no-way result is reached by trying and failing, not by looking and not finding. The premises the parity objection targets are of the second kind. They are not reports about distant worlds but results about what a description must contain in order to be a description of anything at all, and their test is performable at zero distance.

Third. The reliability form is the strongest, and it comes from both directions. Sonia Roca-Royes¹²⁷ presses that conceivability-based methods do not deliver de re modal knowledge of the kind essentialist conclusions require. From the other side Peter Hawke¹²⁸ argues that van Inwagen’s own similarity-to-ordinary-life criterion is unstable, either permitting too much or forbidding the ordinary knowledge it means to protect. Pressed at the present argument, the form asks why an inability to complete a specification should indicate impossibility rather than a limit on the specifier. Every argument from failure-to-construct risks being an argument from ignorance.

The worry is granted in general and answered in the particular. Inability to specify is defeasible evidence, never proof, and an argument that rested on bare inability would be worthless. What distinguishes the cases the book relies on is that the failure is diagnosed rather than merely encountered. When the objector attempts to specify a world lacking the structural conditions, the attempt does not simply stall; it proceeds by individuating that world from others, locating it as a state that obtains, and contrasting it with the worlds it differs from, which is to deploy distinction, succession, and bearing in the act of denying them. That is a demonstrated self-defeat with a stated mechanism, not an absence of imagination. The distinction between I cannot think how and the attempt consumes what it denies is the whole of the reply, and only the second carries weight.

The Specification Test · Standard

1. For a remote candidate p, the question is not whether ◇p is imaginable but whether p can be specified.

2. A specification is adequate only if it does not deploy what p denies.

3. Where every attempted specification deploys what p denies, and the deployment can be exhibited, the failure is diagnosed rather than merely encountered.

4. A diagnosed failure is evidence of impossibility. An undiagnosed failure is evidence only about the specifier.

5. Negative results are claimed here only where (3) holds, and are to be marked as conjectural where it does not.

The standard cuts against the book as well as for it, which is the reason to adopt it. It commits the argument to marking which of its negative results are diagnosed and which rest on nothing better than an unproductive search, and any result of the second kind is owed a downgrade rather than a defense.

That marking has been carried out rather than promised. Across the load-bearing negative results of Chapters 2 and 3 — the Negation Seal at §2.1.5, the null-state reductio and the subtraction reply at §3.2.6, the fictionalist escape at §3.3.3, the masking test at §2.4.12, the conceivability trap at §3.3.4 — and across every verdict in the Chapter 4 ledger, the results the argument relies on are diagnosed in the sense clause 3 requires. One exception was found and corrected. A verdict on strict temporal unreality had been recorded as an incoherence result when the section supporting it claims only an act-side burden and leaves the object-side error theory open (§4.2.5); it now states the weaker result. The exception is reported here rather than silently repaired, because a standard whose first application finds nothing is not being applied. The pattern behind it is worth naming as well: the failure was not an undiagnosed argument but a summary that claimed more than the section beneath it, which is where this standard is likely to be broken again.

An ally is available and should be recorded, though the argument does not depend on it. Timothy Williamson¹²⁹ argues that modal knowledge is not the deliverance of a special faculty at all but a byproduct of the ordinary capacity to evaluate counterfactuals, which is exercised constantly and is not obviously less reliable at metaphysical distances than at everyday ones. If that account is right, van Inwagen’s skepticism loses its target, because there is no distinct modal faculty whose calibration range can be exceeded. The present reply is deliberately weaker and does not require Williamson’s account to be correct. It requires only that specification is a different operation from imagination, and that its failures can be diagnosed.

What closes. Van Inwagen’s challenge is not answered by showing that modal vocabulary is meaningful, which is why §4.1.6 does not answer it, and it is not answered by producing a confident intuition at the relevant distance, which is what the challenge forbids. It is answered by declining the method the challenge is calibrated against. The conclusion is reached by attempted specification and diagnosed failure rather than by imaginative survey, and those failures are exhibited rather than reported. What remains of the objection is a standing demand, correctly placed: that the book distinguish the results whose failures it can diagnose from the results where it has not succeeded in finding a counterexample. The demand is accepted. Van Inwagen was right about the faculty, and right that most of what has been built on it should be given up; the argument here is built on something else.

4.1.14 Impossible Worlds and Relative Modality

Two pressures are often filed together and should be separated, because they attack different parts of the same sentence. One enlarges the space of worlds to include impossible ones.¹³⁰ The other relativizes the quantifier to a contextually supplied background.¹³¹ The first asks whether □(T∧S∧Φ) has been quantifying over too little. The second asks whether its □ means anything absolute at all.

First. The non-normal worlds form is the technically sharpest objection in this section, and it is aimed at the S5 derivation rather than at the structural conclusion. Non-normal worlds are the standard device for invalidating the rule of necessitation: at such a world, logical truths may fail, and modal formulas are evaluated by stipulation rather than by the accessibility clause. Admit them and the characteristic S5 axiom is no longer guaranteed, since ◇C may hold at the actual world while □◇C fails at a non-normal one. A semantics that includes them yields a non-normal modal logic in which the lift the argument uses is simply unavailable.

The reply has two stages, and the first is a redirection rather than a defense. The book does not need the lift. §3.5.4 and §3.5.5 reach the conclusion along a route that does not run through S5 at all, and a critic who blocks the lift by admitting non-normal worlds still arrives at the same place. The S5 derivation is one line among several, and the semantic objection reaches one line.

The second stage is the substantive one. An impossible world is admitted as a representation of an inconsistent content, not as a genuine alternative way for reality to be. Nolan¹³² makes the point in defending the modest approach: impossible worlds earn their keep as representational devices for reasoning about inconsistent scenarios, and admitting them carries no commitment to their being possibilities. But a representational device is an item with content, and content is structured. The reply developed against linguistic ersatzism at §4.4.2 applies here without modification: distinct representations must be distinct (the S-role), must be evaluated in an ordered succession of acts (the T-role), and must be borne by something that carries the content (the Φ-role). Enlarging the space of representations enlarges the space of representations. The □ in □(T∧S∧Φ) ranges over genuine alternatives, and no number of additional representations adds one.

A further point closes the limit case. An impossible world at which T, S, and Φ all fail is not merely an inconsistent scenario but the null state, and the argument of §3.2.6 is that the null state is unspecifiable rather than merely inconsistent. The distinction matters exactly here, because the impossible-worlds theorist is professionally in the business of hosting inconsistent contents. A round square is hostable: it has parts to describe, and the description can be given and then found contradictory. The null state has no parts to describe, and §3.3.3 makes the corresponding point against the fictionalist. An impossible-worlds semantics can host any content that can be stated. What it cannot do is supply a content where none can be stated, and that is the case the necessity claim turns on.

Second. The hyperintensionality form is the actual motivation behind most of the literature, and the honest response is concession rather than reply. Berto, Francesco; Jago, Mark¹³³ argue that necessity is too coarse a sieve for many of the distinctions metaphysics needs: necessarily equivalent contents are not thereby the same content, and a framework whose finest grain is necessary equivalence will misdescribe the differences between them.

That is granted without reservation, and nothing in the argument resists it. The claim defended is that T, S, and Φ obtain at every genuine alternative. It is not the claim that they are constitutive in a hyperintensional sense, nor that the propositions expressing them are individuated by anything finer than necessity. Whether a further, more discriminating notion is required to state what the structural conditions are, as opposed to where they obtain, is the essence-prior question, and it is engaged at §4.1.10 rather than settled here. Hyperintensional distinctions cut below the necessity claim without cutting into it.

Third. The relative-modality form runs from linguistic semantics. On the standard account,¹³⁴ a modal expression is interpreted relative to a contextually supplied modal base together with an ordering source, and the many flavours of necessity (epistemic, deontic, nomic, metaphysical) are one machine supplied with different backgrounds. If that is right, metaphysical necessity is not a privileged absolute but one setting among many, and □(T∧S∧Φ) reduces to a claim about what follows given a particular background. Another background, another verdict.

The reply is that context-relativity in this framework operates on the restriction, not on the domain. A modal base selects a subset of the worlds to quantify over; it does not introduce worlds that were not there. The structural claim is about what every member of the domain instantiates, and a subset of a set whose members all instantiate the conditions has no member that fails to. Varying the background varies which necessities are salient and which are worth asserting. It cannot vary what a world is.

Restriction Adds Nothing · Premises

1. A modal base restricts the worlds quantified over to some subset of the global domain (Kratzer semantics).

2. □(T∧S∧Φ) states that every member of the global domain instantiates the structural conditions.

3. Every subset of the global domain has only members of the global domain.

4. So no choice of modal base delivers a world lacking the conditions.

5. Context-relativity therefore varies which necessities are asserted, not what any world is.

The conclusion generalizes past Kratzer’s framework specifically. Any relativization that proceeds by restricting a domain inherits the argument, because restriction is a subset operation and a subset operation cannot manufacture a member. A relativization that instead replaces the domain, supplying worlds the global domain lacks, is no longer a restriction and owes an account of where the new worlds came from and in what sense they are alternatives. That account is the one §4.4.1 and §4.4.2 already press for, and the demand is unchanged by relabeling the framework as semantic rather than metaphysical.

One clarification of scope is owed. Nothing above denies that there are genuinely relative modalities, or that most ordinary uses of must and can are relative in exactly Kratzer’s way. The §4.4 overview grants this already. The claim is narrower: that a relative modality is a restriction on the field rather than a rival to it, and that the field is what the structural conclusion describes.

What closes. Impossible worlds and relative modality both appear to widen or shift the quantifier and neither does. Non-normal worlds block one derivation of S5, which the argument can afford because §3.5.4 supplies a route that never used it, and they otherwise multiply representations rather than alternatives; the representations are themselves structured, so the multiplication runs the ersatzist’s problem again. Hyperintensionality is conceded outright and leaves the necessity claim standing at its own grain. Kratzer-style relativization restricts a domain it does not populate, and a restriction cannot produce a world the domain lacked. What all three share is a move from the formal apparatus to the metaphysics, and in each case the apparatus turns out to be describing the field rather than adding to it.

4.2 Time, Change, and the Changeless

This cluster asks whether change can be absent from a real world. Each candidate is tested in the same way: does it genuinely describe a reality without change, or does it restore the conditions that make change possible?

The challenge bears most directly on the proof’s claim that any reality permits change and on the identification of change-possibility with T∧S∧Φ. The subsections test whether a supposedly changeless account can do more than redescribe Time, Space, and Substance under another vocabulary.

A preliminary note before the named engagements. Two general intuitions support the positions taken up below. The first is that change might be denied at the limit: that some level beneath, above, or beside the changing world is itself unchanging, and that the unchanging level is the truer one. The second is that the appearance of change might be misleading: what looks like succession may be juxtaposition, what looks like persistence may be pattern. The argument grants the intuitions as intuitions and contests what they license: the structural conditions are derived at §§2.4–2.7 from the actuality, possibility, and impossibility profile of change. Clearing the candidates below turns on the permission side alone: each is introduced as a reality from which change is absent, and each, on inspection, instantiates the conditions it was introduced to lack. That a candidate permits change suffices to answer it, and that sufficiency is the load-bearing claim for the cluster. The Frozen-World Test at §2.4.11 argues that fully typed structure makes change possible; it does not show that every world must contain an actual transition or that every changeless interval is ∅. A frozen or silent world may retain the capacities and role-structure required for change. The cluster therefore tests whether the candidates remove those capacities. The subsections press the weaker, sufficient point and leave the stronger result standing behind it.

4.2.1 The Changeless Ground

The classical theist proposes the most important exception to the claim that change pervades reality: a changeless ground (atemporal, aspatial, immaterial) exempt from all of it. Not frozen in time but outside time altogether. The doctrine has two tightly linked expressions. The first is static-being: God as timeless in the sense of possessing being without temporal instantiation. The second is eternal-present: God as timelessly simultaneous with every temporal moment, possessing life tota simul, whole and all at once.

Thomas Aquinas¹³⁵ develops both: God is actus purus, pure act without potency, entirely without temporal or spatial extension; God is also eternal in the Boethian sense, possessing illimitable life in complete simultaneity. Boethius¹³⁶ introduces the eternal-present doctrine to resolve the foreknowledge problem: God knows all temporal events not by foreseeing them but by perceiving them as simultaneously present. Plantinga¹³⁷ defends a modern version: the Maximally Great Being is a necessary, immaterial mind whose changelessness grounds rather than instantiates the changing world. Eleonore Stump and Norman Kretzmann¹³⁸ refine eternity through their doctrine of ET-simultaneity (eternity–time simultaneity), a non-temporal simultaneity relation between God and temporal events. Brian Leftow¹³⁹ develops a quasi-temporal eternity (QTE) variant. Julian Barbour¹⁴⁰ proposes a physical analogue: Platonia, a timeless configuration space from which the experience of time arises as an illusion. The intuition is ancient, the defenders are serious. If any such being is coherent, it grounds structure without being subject to it.

The counter-construction bears on the argument’s most ambitious reach. The argument claims that wherever change is possible, Time, Space, and Substance obtain. The classical theist counter-proposes a being for which change is not possible (for which T, S, and Φ are strictly inapplicable) and which grounds structure rather than participating in it. If such a being can exist, then □(T∧S∧Φ) as a claim about all of reality fails. Structure obtains wherever there is change, but something else obtains changelessly, and that something else is the ground.

A clarification before the moves. The argument’s intermediate conclusion is □◇C: Reality necessarily permits change, not that change must always be occurring. The static-being objection treats these as different claims. They are not. Permission is not a fact about whether change happens in a being. It is a fact about what Reality structurally allows. A static being that exists within Reality exists in a Reality that already permits change, and that permission is structure. The objector is proposing something stronger: a being that exists outside Reality-as-structured. The shared structural failure has a name in advance. Changeless is a parasitic negation of Change, a quantifier internal to the domain of transitions, not a state outside it (§2.1.5, Parasitic Negation). A ground specified as changeless either reduces to that quantifier and grounds nothing, since a quantifier is not a substrate, or smuggles in determinate features that instantiate the very domain the construction was meant to escape. Five forms of the move test that dilemma.

First. The Maximally Great Being move proposes a being atemporal, aspatial, and immaterial that possesses being without instantiating any of the three roles, grounding the structure without being part of it. A changeless being cannot be identified with ∅ merely from atemporality, aspatiality, or immateriality. At most, its being determinate rather than nothing supplies a thin difference-role. The further step from qualitative or logical distinctness to Spatiality’s loaded type — relational extension within a manifold — requires the independent bridge argued elsewhere (§2.4.7); it cannot be assumed against the classical theist at the outset.

Now consider grounding: the static being is proposed as the ground of existence, something that produces or sustains structure. Any grounding relation, whether construed as causal production or as metaphysical/explanatory dependence, requires determinate relata, asymmetric direction between them, and real obtaining. An atemporal grounding theorist will deny that it must persist through a temporal interval; the simultaneous-causation move is taken up in detail at §5.1.3. The local conclusion is therefore role-level: a wholly indeterminate ground cannot ground, while a determinate grounding relation instantiates difference, direction, and bearing. Whether those roles require the fully temporal and physical types is the separate role-to-type question. The proposal seeks a ground with none of the Conditions that grounding requires, and so leaves itself with no determinate way to specify that ground at all.

Second. The Aristotelian non-temporal priority move refines the claim: the static being is prior to the structure in the order of explanation, not in the order of time. Formal causes (what a thing essentially is) and final causes (what a thing is for) involve non-temporal ordering; a changeless being can stand in formal or final priority to the structure without standing in temporal priority. Non-temporal priority is still structural priority. Formal priority requires the ground to be structurally distinct from what it grounds, requiring the difference-role to hold ground and grounded apart. Final priority is not temporal succession, and the Aristotelian point should be granted in full. But it is still an asymmetric ordering relation: source and end, potency and actuality, orientation and fulfillment. The structural argument does not need to redescribe final causality as clock-time. It needs only the weaker claim that final priority is not structureless priority. To ground structure, the static being must stand to what it grounds in some determinate ordering relation (formal, final, logical, or temporal) and a determinate ordering relation already requires distinguishable relata, asymmetric priority, and a stable relation between them. The defining feature of the static being, its non-instantiation of the roles, is exactly what disqualifies it from standing in any grounding relation at all.

Third. The Boethian eternal present presses a different line: God exists not in time but in an eternal present, a timeless simultaneity from which the whole of temporal history is present at once, as a landscape is present to an observer on a hill. From eternity, all temporal moments are equally present, not as a series God remembers or anticipates but as a totality God possesses, tota simul. This is not static immobility but complete, simultaneous, perfect possession of illimitable life.

The eternal-present view should not be caricatured as requiring God to track a moving present from inside an A-series (McTaggart’s term for the tensed past/present/future ordering, distinct from the tenseless B-series of earlier/later relations; §4.2.5 develops the distinction). Its point is exactly the opposite: God knows each temporal event in its temporal location without occupying one moment rather than another. The pressure arises when the doctrine is asked to preserve more than tenseless total knowledge. Classical theism also wants divine responsiveness, providence, answer to prayer, and knowledge of creaturely events as events to which God stands in determinate relation. Those commitments require more than the passive possession of a static array. They require asymmetric dependence between prayer and answer, between creaturely event and divine knowledge, between what is asked and what is given.

If the Boethian account gives only B-series access to the whole temporal order, it has not yet supplied response in the operative sense; prayer and answer stand as correlated elements of the array rather than in a response relation. If it supplies genuine response, it has introduced ordered asymmetry, distinguishable relata, and a stable relation between them, the structural roles in the very place the doctrine was meant to exempt from them. The objection is not that God must know “currently” in the creaturely mode. It is that the theological predicates the doctrine wants to preserve require a structured relation that mere eternal co-presence does not by itself provide.

Fourth. Stump and Kretzmann’s ET-simultaneity offers a technically refined version: a non-temporal simultaneity relation between God and temporal events. God is simultaneous with every temporal moment, not in the sense of being at the same time as each, but in the sense of standing in a “simultaneity in the mode of eternity” relation to each. This relation is neither temporal nor non-existent; it is a third kind, definable precisely.

ET-simultaneity replaces ordinary temporal simultaneity with a sui generis relation between eternity and temporal events. That relation supplies determinate relatedness, but relatedness alone does not entail spatial extension. The live objection is instead whether ET-simultaneity can support asymmetric response and agency without importing temporal succession; the doctrine should not be rejected simply by redescribing every relation as Space. J. M. E. McTaggart¹⁴¹ argues that genuine change requires the A-series; agency, which involves change, inherits the requirement, and ET-simultaneity forgoes that agency in the trade.

Fifth. Barbour’s Platonia offers the most radical version, physical rather than theistic: a timeless configuration space with the experience of time an illusion arising from “time capsules”, configurations that happen to encode apparent memories. No being, just the static configuration space.

Platonia is a world whose description requires time in the describing. Configurations must be individuated (each is this configuration rather than that) requiring relational non-identity (Space). The configuration space must subsist as one structural whole that bears its configurations (Substance). Distinguishing one configuration from another requires cognitive succession from one to the next (Time). The act of describing Platonia occurs in time, but that establishes only an act-side fact about the describer. It does not by itself show that Platonia’s represented content is temporal. The object-side question is whether the configuration space contains an intrinsic ordering or dynamical structure adequate to recover time; that must be argued from the model rather than inferred from the description event.

A kindred softening worth naming is Leftow’s¹⁴² QTE (quasi-temporal eternity) which allows God to stand in quasi-temporal relations to creatures while remaining intrinsically timeless. QTE is the weakest version of the doctrine, and it concedes most explicitly that relations to temporal creatures require some structural analog of Time. The position is honest about the cost: intrinsic atemporality is preserved only by distinguishing intrinsic from relational structural features, and the relational features do the work that the structural roles require. Open-theist softenings (Hasker, Sanders) concede still more (temporal duration, genuine response, real relation to creatures) and concede the structural commitments most explicitly.

What converges. The five forms exhaust the outside-proposal, and each fails the same way: reaching outside structure only to find structure already there. The Maximally Great Being has no foothold for specification (a wholly indeterminate ground cannot ground, and a determinate grounding relation instantiates difference, direction, and bearing); non-temporal priority is still structural priority, since formal and final precedence require the very conditions they were meant to escape; the Boethian eternal present either supplies only B-series co-presence, which is not yet response, or supplies genuine response and with it the ordered asymmetry it was meant to exempt; ET-simultaneity supplies determinate relatedness but forgoes asymmetric response and agency in the trade; and Platonia’s configurations are individuated in a space that bears them, while the succession its description deploys poses a question the model has yet to answer.

The static being is proposed from within the structure it is supposed to precede: a present is a position in a sequence, and with the sequence removed, “present” names nothing.

Sibling theistic counter-constructions: the simultaneous-causation mechanism classical theism uses to connect the eternal God to the temporal world is answered in §5.1.3, and the divine-simplicity construction of the eternal God’s internal metaphysics in §5.2.1.

4.2.2 Shoemaker’s Time Without Change

Sydney Shoemaker, writing at Cornell in 1969, constructs a universe built on local freezes: a region in which nothing moves, decays, or interacts for a stretch of time, then resumes exactly where it left off, displaced from the ongoing world by the length of the freeze. He multiplies the case into three rooms subject to periodic freezes on staggered intervals, a universe built to let observers in one region infer the freezes of another by clock comparison. Run the arithmetic long enough and the periodicities coincide: once every sixty years, all three regions freeze at once. During that year, by Shoemaker’s construction, nothing changes anywhere in the universe. And yet, the argument runs, time still passes.

The inference is staged with such care that the paper became the canonical modern defense of time-without-change in the analytic literature, the standing objection to every Aristotelian who thinks time requires motion. The objection is serious. Pressed to the global limit, however, it collapses under its own construction.

Shoemaker (“Time Without Change,” The Journal of Philosophy)¹⁴³ offers the most carefully constructed defense of time without change in the analytic literature. Local cessation of change is coherent: a region of a universe can freeze while the rest continues, and the frozen region can be re-integrated into the ongoing temporal order when it resumes. That is true, and the present account grants it without cost. What is contested is whether the scenario extends to a universe-wide cessation of change while time continues to pass: whether Shoemaker’s construction, pressed to the global case, describes a coherent modal possibility or collapses into incoherence when the apparatus that supports the inference is removed.

The literature on time-without-change runs back to Aristotle’s Physics,¹⁴⁴ where time is defined as the number of motion with respect to the before-and-after, and an unchanging universe therefore fails the conditions for time’s obtaining. Newton’s Principia¹⁴⁵ develops the opposing substantivalist position: absolute time flows uniformly without regard to anything external, and its passage is independent of whatever change happens to take place within it. McTaggart (“The Unreality of Time,”)¹⁴⁶ poses the foundational question of what change itself consists in, the A-series/B-series distinction that structures the modern debate. Prior (“Changes in Events and Changes in Things,”)¹⁴⁷ sharpens the change-theoretic tradition. Shoemaker’s 1969 paper is the modern inheritor of Newtonian substantivalism (time as an independent substance whose passage does not depend on any change within it) and the argument’s most rigorously worked-out form. Newton-Smith (The Structure of Time)¹⁴⁸ defends the Shoemaker conclusion at length.

Against the substantivalist line: Le Poidevin (Change, Cause, and Contradiction)¹⁴⁹ presses the verificationist worry about the global-freeze extension; Lombard (Events: A Metaphysical Study)¹⁵⁰ develops an events-based metaphysics on which change is constitutive of time; Graeme Forbes¹⁵¹ challenges the inferential structure of Shoemaker’s thought experiment; Mellor (Real Time II)¹⁵² and Craig (The Tensed Theory of Time)¹⁵³ defend change-theoretic views from B-theoretic and tensed angles respectively; Horwich (Asymmetries in Time)¹⁵⁴ works the change-primitive side of the debate; Sider (Four-Dimensionalism)¹⁵⁵ develops the perdurantist alternative on which change is cashed out through temporal parts; Dainton (Time and Space, 2nd ed.)¹⁵⁶ gives the standard survey; Bardon (A Brief History of the Philosophy of Time)¹⁵⁷ places Shoemaker’s contribution in its long historical arc. The present reply sits within the change-theoretic tradition, closest to the Mellor–Le Poidevin wing, and adds what neither specifies on its own: Shoemaker’s inference runs on the very persistence the scenario was trying to eliminate, and that persistence is named explicitly by the structural-role apparatus of §2.6.

The challenge bears directly on §2.6. The Modal Identity established there is bidirectional: necessarily, change-possibility obtains if and only if T ∧ S ∧ Φ obtains, and the two are one structural fact under two descriptions. Shoemaker’s argument exhibits, if successful, time during an interval in which no change is occurring. That does not by itself break the book’s Modal Identity, whose left side is the possibility of change rather than its continuous actuality. A globally frozen interval bracketed by change may still belong to a world in which change is possible. The relevant question is therefore whether the case preserves change-capacity and structural continuity, not whether a transition occurs during every interval. What follows is narrow by design. The reply locates the substantivalist move as a relocation rather than an elimination.

A clarification before the moves. The dispute is not with Shoemaker’s local scenario: rooms that freeze while others continue are coherent, and the argument of §2.6 does not contest them. Nor is the dispute with Shoemaker’s inferential apparatus considered in isolation: given three regions freezing on staggered intervals, one can indeed calculate when they coincide, and the calculation is sound. What is contested is the step from the local scenario, where change continues in the unfrozen regions, to the global scenario, where the universe as a whole freezes. Three forms of the move take up the objection at three depths: the inferential-structure form, the level-invariance form, and the extension-to-the-limit form. The space they canvass is the space Shoemaker’s argument can occupy. The reply argues that each confirms the Modal Identity rather than breaking it.

First. The inference is always from observed change, not to time-without-change. Shoemaker’s observers reconstruct A’s frozen interval by reference to B’s and C’s ongoing change: by counting B’s cycles while A is frozen, by watching the clock. When all three regions freeze simultaneously, no such reconstruction is available from inside the freeze. The claim that time still passes during the global freeze is not a direct measurement but an extrapolation from the periodicities on either side of the gap, underwritten by the assumption that the pattern continues across it. That assumption is Substance (the persistence condition §2.6 names) carrying the identity of the periodicity across the interval. Without Φ operating across the freeze, the post-freeze regions are not the pre-freeze regions resumed; they are new regions with coincidentally similar patterns, and the inference collapses into correlation. Shoemaker’s reasoning works because the universe stays the same universe across the pause. A universe that does not stay the same across the pause cannot underwrite the reasoning. The scenario describes a gap inside a universe whose Substance remains intact across it. That persistence is precisely what the structural conditions name.

Second. Level-invariance (§2.1.7) applies to the containing universe, not just to each region. The regions are genuinely distinct (S). The freezes occur in a determinate order relative to one another (T). The containing frame persists across the freezes to underwrite each region’s re-emergence with preserved identity (Φ). The simultaneous freeze is a local phenomenon within a structured universe, a phenomenon the universe’s structural conditions make describable in the first place. Shoemaker has described a universe with change in which certain regions undergo periodic cessation, not a universe without change. The time-without-change the argument purports to exhibit is always the frozen region rather than the universe, and the scale of the freeze is not the scale of the claim.

Third. Pushed to the global limit, the scenario contains no internal clock or observer during the freeze. That creates an epistemic and verification problem, not an immediate metaphysical contradiction: a thought experiment may specify unobserved structure, and observers before and after may infer a duration from the wider history. The structural reply should therefore not depend on the absence of a thinker. Its narrower point is that the construction preserves cross-freeze identity, periodic dispositions, and the possibility of resumed change; it is a frozen interval within a change-permitting world, not a world lacking the conditions of change. The fully extended Shoemaker scenario is therefore not a universe with time but no change; it is a universe with change containing a region where nothing happens, which is exactly the case the first two replies already handle. The bidirectional identity survives because the scenario, followed through, never actually breaks it. The Aristotelian position is recovered by the structural apparatus: time is not the number of motion merely, but motion’s structural conditions are internal to what time consists in. Strip them and what remains is not empty duration but ∅.

A kindred move drops the Shoemaker apparatus for classical Newtonian absolutism: time is an independent substance whose passage does not depend on any change within it. Newton’s Principia¹⁵⁸ articulates the position most clearly, and Newton-Smith¹⁵⁹ carries it forward. The reply is the same in older vocabulary, with one dependency made explicit. An absolute time independent of change must still be individuated (this moment of absolute time distinct from that moment), and, on the change-theoretic conception of Temporality defended at §1.4.1 (Directionality as structural, not imposed) and abductively at §7.2, that individuation is the structural fact §2.6 identifies. Absolute-time substantivalism relocates the structural roles into the substance of time itself rather than eliminating them; the substance of time contains distinctness of moments (S), directed ordering (T), and persistence across the ordering (Φ) as the conditions of its being a time at all.¹⁶⁰

Another kindred: a deflationary verificationist move¹⁶¹ that the global-freeze scenario is meaningless because no possible observation could confirm or disconfirm it. The structural reply accepts the verificationist point for the specific purpose for which it was introduced while declining the broader verificationism: the scenario is not meaningless; it is underspecified in a way that dissolves under analysis, which is a stronger claim than verificationism alone supports.

What closes. Three depths and none lets the global extension stand. The inference from local freeze to global freeze runs on Substance (the very persistence condition Φ specifies) carrying the identity of the periodicity across the interval; without that persistence the post-freeze regions are new regions with coincidentally similar patterns, and the inference collapses into correlation. Level-invariance: the scenario describes a structured universe containing regions of freeze, not an unstructured universe frozen entire. The fully extended scenario preserves cross-freeze identity, periodic dispositions, and the possibility of resumed change; it is a frozen interval within a change-permitting world, which is exactly the case the first two replies already handle.

The emergent-time objection, that time is derivative rather than fundamental, is §4.2.3. The momentariness objection, that persistence is illusory, is §4.2.8. The changeless-ground objection, that a timeless being grounds the structure from outside, is §4.2.1. The Shoemaker objection occupies a distinct position relative to these three: it grants that time is real and that persistence obtains, and contests only the bidirectional linkage of change and time. The reply is therefore correspondingly narrow: the target is the global extension of the local scenario, not the coherence of time or of persistence in general.

What Shoemaker’s argument supports is stronger than merely local cessation: it gives a serious model of an interval, potentially global, during which no change occurs while temporal structure is retained. The structural account can grant this because its modal claim concerns permission for change, not uninterrupted occurrence. The claim of §2.6 is that the fundamental ontological structure cannot instantiate Time without instantiating change, not that every subregion must undergo change at every moment. Shoemaker has produced a counterexample to a claim the argument does not make.

One clarification for the Shoemakerian reader. The argument addressed here is the 1969 time-without-change paper, not Shoemaker’s signature causal theory of properties (Causality and Properties, 1980; Identity, Cause, and Mind, 1984/2003), on which properties are individuated by their causal powers. That broader thesis is not a further counterexample to §2.6 but an instance of it: causal individuation presupposes the very structural roles the account names, since a causal power is exercised across an ordering (T), among distinct relata (S), and is borne by what persists long enough to exercise it (Φ). The omission of the causal theory from the reply above reflects the scope of the time-without-change objection, not a narrowing of the account.

What the paper actually gives its readers, followed through, is not the universe-without-change it seems to promise but something better: the sharpest available demonstration that inferring time-without-change requires a standpoint where change is occurring.

4.2.3 Emergent Time

Suppose time is not fundamental. Suppose what we call time is a feature that emerges from a deeper, timeless level of reality: the way temperature emerges from molecular motion, or the way solidity emerges from atomic lattices. The fundamental description of the world, on this proposal, makes no reference to time; time is a useful approximation at our scale, a coarse-grained feature that drops out of the deeper physics in the appropriate limit. If this is right, then the structural account of §2.6 is at best a description of the emergent layer, not of fundamental reality. The classical theist, sensing again an opening, can locate the changeless ground at the timeless fundamental level and the structural roles at the emergent surface.

The move is not idle. The Wheeler–DeWitt equation, the canonical equation of quantum gravity, contains no time variable; the wavefunction of the universe is timeless, and the appearance of time in our experience is to be recovered (somehow) from a fundamentally timeless formalism. Don N. Page and William K. Wootters¹⁶² develop the most influential proposal: time emerges as a correlation between subsystems of a timeless universe, with one subsystem serving as a clock for the rest. Barbour’s The End of Time¹⁶³ develops Platonia: a timeless configuration space from which temporal experience emerges as “time capsules,” configurations that happen to encode apparent memories. Rovelli’s The Order of Time¹⁶⁴ offers a relational account: time is the structure of correlations between physical systems, not a fundamental backdrop. Smolin’s Time Reborn¹⁶⁵ takes the opposite line, defending time’s fundamentality, but the position the present reply targets is the timeless-fundamental position, not Smolin’s.

In the philosophy of time, kindred views include Albert’s Time and Chance,¹⁶⁶ which derives temporal asymmetry from the past-hypothesis and statistical-mechanical considerations; Price’s Time’s Arrow and Archimedes’ Point,¹⁶⁷ which argues that the apparent directedness of time is a perspectival feature; Reichenbach’s The Direction of Time,¹⁶⁸ the classical mid-century treatment of time’s arrow; Maudlin’s The Metaphysics Within Physics,¹⁶⁹ which defends time’s fundamentality against eliminativist readings. Anderson’s The Problem of Time¹⁷⁰ gives the technical state of play in canonical quantum gravity. The literature is vast and the technical questions are subtle. The present reply targets the structural claim: that time is derivative rather than fundamental, and therefore that the structural conditions of §2.6 obtain at most at an emergent layer.

Three forms of the move develop the position at three depths: the strict-emergence form, the correlational-time form, and the relational-time form. The reply argues that each fails for the same structural reason at increasing levels of refinement.

First. The strict-emergence form claims that the fundamental level is timeless and that time is a phenomenon of the emergent layer only, in the way that temperature is a phenomenon of bulk thermodynamics and not of individual molecules. The Wheeler–DeWitt equation, on this reading, describes a fundamentally timeless universe, and the appearance of time at our scale is a coarse-grained illusion.

The analogy with temperature does not carry the weight the position needs. Temperature is emergent in the sense that it is a statistical property of an underlying dynamics that is itself temporal: the molecules move, collide, and exchange energy in time, and the bulk property of temperature is a statistical summary of that temporal dynamics. The molecular dynamics is not itself temperature, but it is itself temporal; the emergence is from a more fundamental temporal description to a less fundamental temporal description. For time itself to be emergent in the analogous sense would require a more fundamental description that is itself non-temporal but supports the emergence of time, and this is exactly what the strict-emergence form needs and exactly what the position cannot supply. The Wheeler–DeWitt equation does not describe a non-temporal dynamics from which time emerges; it describes a wavefunction satisfying a constraint, and the recovery of time from the wavefunction is the open problem the field has been working on for decades. To call the formalism “timeless” is to describe the constraint, not the dynamics; the dynamics, if there is to be one, is what the recovery program is meant to produce.

The deeper structural point: any proposed fundamental level that supports the emergence of time must contain the resources for time to emerge from it. Those resources are the structural roles §2.6 names. Distinctness of fundamental states (S), an ordering across them sufficient to support the emergent direction (T), and a persistence spanning the ordering sufficient to underwrite the emergent identity of objects across the emergent time (Φ). The fundamental level can be called “timeless” only in the restricted sense that it does not contain time as we recognize time at our scale. It cannot be timeless in the sense of lacking the structural roles: if it lacked them, no time could emerge from it.

Second. The Page–Wootters correlational form is more refined: time emerges as the correlation between two subsystems of a timeless universe, with one subsystem (the clock) serving as a reference relative to which the other subsystem (the rest) appears to evolve. Mathematically, the global wavefunction is timeless; the conditional wavefunction (conditioning on the clock state) evolves in the clock variable. The conditional evolution is what we observe as time.

The correlation between subsystems requires distinctness of subsystems (S), an ordering of clock states sufficient to support the conditional evolution (T), and a persistence of the universal wavefunction across the conditional structure (Φ). The Page–Wootters construction is internally coherent and structurally informative: it shows that something that looks like time at one level can be derived from a timeless-looking formalism at another level. What it does not show is that the structural roles are absent at the fundamental level. The clock subsystem must be distinct from the rest (S obtains). The clock states must stand in an ordering that the conditional evolution tracks (T obtains in the clock structure). The universal wavefunction must persist across the conditional construction (Φ obtains globally). The construction is a relabeling of where T is located (in the clock variable rather than in a global parameter), not an elimination of T.

Third. The relational-time form, developed most influentially by Rovelli, claims that time is not a backdrop in which physical systems evolve but a structure of correlations between physical systems. There is no absolute time; there is only the relative evolution of one system with respect to another. Each system is a clock for the others, and the appearance of a universal time is the coarse-grained agreement of many such relative clocks.

Relational time is a thesis about the metaphysics of time, not its elimination. Rovelli’s view is closer to Leibnizian relationism than to eliminativism: time consists in relations between physical systems rather than in an absolute backdrop, but the relations are real and structural and do the work that absolute time was supposed to do. The reply, accordingly, is not that relationism is incoherent (it is not) but that relationism is not a refutation of the structural conditions: it is a relocation of where the conditions are instantiated. The relations are between distinct systems (S), they support directed evolution (T), and they persist across the systems’ interactions (Φ). Relational time is the structural conditions of §2.6 in a relationist metaphysics.

A kindred move is the past-hypothesis style derivation (Albert, Price): the apparent directedness of time is derived from a low-entropy initial condition plus statistical mechanics, with no fundamental time-asymmetric law. The reply is the same. The derivation requires distinctness of microstates (S), an ordering across them sufficient to define entropy increase (T), and a dynamical evolution that preserves the relevant structural facts across the increase (Φ). The asymmetry of time is derived from the past-hypothesis, but the symmetric temporal structure is presupposed by the derivation, not eliminated by it.

Another kindred is the eternalist/B-theoretic position that the appearance of temporal becoming is a perspectival feature of conscious experience while the world itself is a four-dimensional manifold without genuine becoming. This is treated separately at §4.2.6 (the B-theoretic block). The B-theoretic eternalist accepts T (the four-dimensional manifold is temporally extended) and contests only the metaphysical reality of A-series becoming. The dispute is therefore not about whether T obtains but about its metaphysical character.

What converges. Three forms and none lets the elimination of T at the fundamental level stand. Strict emergence cannot supply a non-temporal fundamental dynamics; the Wheeler–DeWitt ‘timelessness’ is a constraint on the wavefunction, not a non-temporal dynamics from which time can emerge. The Page–Wootters construction relocates T from a global parameter to a clock subsystem; relocation is not elimination. Relational time is a metaphysics of time as relations between systems, not an elimination of the structural role time plays. Each move produces a redescription of where the structural conditions are located, not a description of a world in which they are absent. The shared failure is structural: timeless and never are parasitic negations of Time, quantifiers internal to the temporal domain rather than substrates outside it (§2.1.5, Parasitic Negation). A ¬∃t (the assertion that there is no t, no temporal coordinate, no instant) cannot ground an emergence of time, since a quantifier is not a level. The three forms collapse into relocations because there is nothing else coherent for them to be.

Sibling counter-constructions: the changeless-ground objection, on which a timeless being grounds the structure from outside, is §4.2.1. The Shoemaker objection, on which time can obtain without change, is §4.2.2. The momentariness objection, on which persistence is illusory, is §4.2.8. The B-theoretic block objection, on which the temporal manifold lacks genuine becoming, is §4.2.6.

What the emergent-time literature genuinely establishes is worth preserving: the time of our experience is not the time of fundamental physics in any direct mapping. The clock-based time of relativity is not the conditional time of canonical quantum gravity; the entropic arrow is not the relativistic light-cone structure; the phenomenal flow of conscious experience is not the parameter of the Schrödinger equation. There are many time-like structures in physics, and the relations between them are subtle and partially open. The structural account of §2.6 is consistent with this plurality. What the account requires is that any proposed fundamental level that supports the emergence or relocation or relabeling of time-like structure contains the resources (distinctness, ordering, persistence) for that emergence or relocation or relabeling to be possible. The structural roles are not the time of any particular physical theory; they are the conditions any theory must instantiate to have time-like structure at all.

The research program around the Wheeler–DeWitt equation has occupied physicists for over half a century; no consensus on the recovery of time has emerged. What has emerged is a more careful articulation of what the structural conditions of time actually are, and the present reply is a beneficiary of that articulation. The fundamental level may turn out to be very different from the time of our experience. It will not turn out to be without distinctness, ordering, and persistence, because if it were, no time could emerge from it, and the recovery program would have nothing to recover.

4.2.4 Platonism

What of mathematics? Numbers seem to be the paradigm of changeless objects. The number 7 is not somewhere; it does not move; it does not decay; it does not depend on anything for its existence; it stands in eternal mathematical relations with the other numbers and with the structures it figures in. If anything is a counterexample to the present account, it is the realm of mathematical objects: abstract, atemporal, aspatial, immaterial, and indubitably real (or at least real enough that mathematical practice across cultures and millennia has converged on remarkably stable conclusions about them). The Platonist about mathematics has the most natural case for objects exempt from the structural conditions of §2.6.

The move is venerable. Plato’s Forms, the original Platonist proposal, are paradigmatic of changeless reality: the Form of the Triangle, of Justice, of the Good, each eternal and unchanging, the temporal world of becoming a degraded copy of the timeless world of being. Frege’s Grundlagen der Arithmetik¹⁷¹ revives mathematical Platonism in modern dress: numbers are objective abstract objects, and arithmetic is the science of their structure. Gödel¹⁷² defends a robust Platonism on which mathematical objects are perceived through a faculty analogous to sense perception. Quine and Putnam’s indispensability arguments¹⁷³ make mathematical realism respectable to naturalists who would otherwise have shied away from abstract objects. Shapiro’s Philosophy of Mathematics: Structure and Ontology¹⁷⁴ develops structuralism, on which mathematics is the science of structures rather than of particular objects. Maddy’s Defending the Axioms¹⁷⁵ and Naturalism in Mathematics¹⁷⁶ defend a naturalized Platonism. Balaguer’s Platonism and Anti-Platonism in Mathematics¹⁷⁷ gives the comprehensive survey. The classical theist sometimes adopts a Platonism-plus theology: God’s mind contains the abstract objects, and the realm of Forms is the divine intellect.

The target of the present reply is the structural claim: that mathematical objects are abstract entities exempt from the structural conditions of §2.6, and therefore that the Modal Identity of §2.6.6 (the result that the necessary possibility of change entails the necessary co-instantiation of Time, Space, and Substance) holds at most for the concrete world while a wider reality contains entities that violate it. The reply is not that Platonism is incoherent. One available response is nominalist or modal-structuralist: mathematical objects can be treated as descriptions of what structured reality, actual and possible, affords. But that is a substantive philosophy of mathematics, not something established merely by the structural argument. A robust Platonist may instead retain abstract objects while denying that the book’s concrete change-realizing types apply to them.

Four forms of the move develop the position at four depths: the classical-Platonist form, the structuralist form, the naturalized form, and the plenitudinous form. Each fails for the same structural reason at increasing levels of refinement.

A note on the shape of the reply before the four forms. The reply does not claim that mathematical objects are abstract entities possessing their own time-like, space-like, and substance-like structure in some rarefied mode. That move would only relocate the Platonist’s exempt object and quietly stretch Time, Space, and Substance into a second, analogical sense, trading the determinacy of the concrete bearers for a metaphor. The reply runs the other way. There is no number 7 standing apart from things, awaiting structural analysis; there are pluralities that can be counted, structures that can be instantiated, and 7 is the description of a feature they have. Mathematical objects are descriptions of what structured reality affords, not residents of a separate realm that must be shown to satisfy the conditions after all. Because the conditions govern the possible and not merely the actual (§2.6.6), the descriptions reach as far as the possible reaches (to collections larger than anything the actual world contains, and to the infinite) without positing a single abstract object.

First. Classical Platonism (the Plato–Frege–Gödel line) holds that mathematical objects are particular abstract entities, each numerically distinct from the others, standing in eternal relations of mathematical structure. The number 7 is a particular object, distinct from the number 6 and from the number 8; the natural numbers stand in the successor relation; the relation is eternal and changeless.

Platonic Forms presuppose form. The word form (English form, Latin forma, Greek εἶδος) carries both the verbal sense (to take shape, to come into being) and the nominal sense (the eternal pattern, the abstract universal). εἶδος descends from ἰδεῖν, to see: it named originally what something looks like as it takes shape, the seen aspect of something undergoing process. Plato’s move is to read the seen-aspect as an eternal pattern that prescribes the shape the process takes. We observe trees taking shape, mammals developing, planets accreting; the Platonist infers Forms of Tree, Mammal, and Planet prescribing what those processes converge on. The inference trades on the equivocation between the verbal sense (formation) and the nominal sense (the eternal Form): what follows from the observation is that formation-processes occur, and that the conditions of reality are such as to permit them, not that there is an eternal pattern prior to them prescribing what they must produce.

Three pressures on the classical-Platonist line are worth naming before the section turns to the math-Platonist refinements. The natural-kind pressure. Trees vary across species, climates, and life-stages; mammals span the spectrum from blue whale to shrew; the Form of Tree or Form of Mammal that would be at the same explanatory level as the Form of the Triangle becomes harder to specify the closer the inspection goes. Plato himself hesitates on this point in the Parmenides (130c–d), where Socrates flinches over whether there are Forms of hair, mud, and dirt. The modern reactivation runs through Wittgenstein’s family-resemblance critique,¹⁷⁸ Dupré’s species pluralism,¹⁷⁹ and Boyd’s homeostatic-property-cluster account of natural kinds,¹⁸⁰ each pressing the point that the kinds the world contains are not the kinds the Platonist’s Form-talk requires.

The process pressure. A tree is not a static form: it is seed, sapling, mature, senescent, snag, and across seasons evergreen or deciduous. The Form of Tree would have to be either the form of a static cross-section (which is not what a tree is) or the form of the developmental sequence (which makes the Form temporal, conceding the present account). Aristotle presses this against his teacher in Metaphysics A.9, 991a8–b9 (where the Forms are shown useless for change and coming-to-be) and develops the immanent, hylomorphic alternative in De Anima II.1, 412a19–28; the process-philosophical tradition from Heraclitus through Whitehead¹⁸¹ and Bergson¹⁸² develops the point systematically.

The reflexive-substrate pressure. The Platonist standing on Earth is held by what they describe (a planet) and owes an account of the Form of Earth that supports them. Retreat to property-universals (Armstrong, Tooley) abandons Forms of individuals; retreat to the Timaeus elemental geometries (53c–55c) confuses earth-the-element with Earth-the-planet; retreat to Plotinian emanation (Enneads V.9, VI.9) substitutes a Neoplatonist theology for the Platonist analysis; retreat to Lewisian modal realism¹⁸³ changes the ontology under discussion. Each retreat concedes that the Form of the individual thing supporting the Platonist was not available.

The pressures share a structural diagnosis. The Platonist treats Forms as prescriptive: the Form of Tree explains why trees take the shape they do; the Form of Triangle explains why triangles have the angle-sum they have; the Form of Justice explains what just acts have in common. The structural account inverts the direction. That trees take the shape they do follows from the structural conditions of §2.6 being such as to permit trees: the necessary co-instantiation of Time, Space, and Substance yields a Reality in which tree-formation is among the possible processes. The descriptive content the Platonist reads as the Form of Tree is the description of what those conditions afford for that kind of configuration. The 180° angle-sum of a Euclidean triangle is not prescribed by a Form of Triangle; it is a structural consequence of the metric specification of the relevant space (§2.4.7 develops the metric-as-contingent-specification point) and spherical and hyperbolic geometries produce different angle-sums under different metric specifications, all licensed within the structural conditions. What the Platonist calls a Form runs downstream of actuality plus structural conditions, never upstream. This is the polarity inversion: what looks like a prior pattern prescribing actuality is a description of what actuality, under the structural conditions, produces.

To be a Form, at all, is to be a determinate this rather than that: the Form of Justice, not Beauty; the Form of the Triangle, not the Circle. But that determinacy is borrowed, not native. The Form of the Triangle is determinate because triangles (actual or possible) are determinate; its distinctness from the Form of the Circle is the distinctness of the configurations it describes. The determinacy the Platonist reads as a property of the Form is the determinacy of the structured (possible) reality the Form is a description of. The Form is not a further object that must itself be located in the order of distinctness, placement, and persistence; it is the description that order makes available.

Take the number 7. What is 7 apart from seven of something? The classical Platonist’s answer is: a particular abstract object, distinct from 6 and 8, sitting in the eternal successor relation. The present answer is that 7 is the count-structure of any seven-membered plurality there could be, the way any such collection is when it is enumerated. Its distinctness from 6 and 8 is the distinctness of those collections (one fewer thing, one more); the successor ordering is the ordering of the collections under the operation of adding a unit; the stability of “7 is the successor of 6” across every time, person, and culture is the stability of what any possible seven-membered collection would be. These can be represented as facts about actual or possible pluralities. Actual concrete pluralities have concrete members, but merely possible, infinite, or abstract pluralities cannot simply be assumed to occupy Space, stand in temporal order, or be made of Substance. The modal-structuralist analysis must explain those cases rather than treating concreteness as automatic. The number needs no private time, space, and substance of its own, because it was never an object over and above the things; it is the description of their count. And because the count is a count of possible collections, it reaches numbers larger than all the things that happen to exist, and the transfinite, without any object’s being conjured to bear it. Classical Platonism does not exhibit a counterexample to the account developed here; once the number is seen as the description of structured possibility rather than as an object exempt from it, the apparent counterexample dissolves.

Two homely cases make the reading concrete. First, measure. No one is tempted to house “1.5 meters” in a realm of abstracta; it is how long a thing is, a magnitude borne by the rod that has it. Number belongs to the same family: 7 is a how many in the way a length is a how long, a feature of a plurality rather than an object standing beside it. The Platonist’s temptation is to treat the one magnitude we happen to reach by counting as though it were unlike every other magnitude, a thing rather than a measure. Second, matching. A shepherd who cannot count can still keep his flock exactly: he drops a pebble into a sack as each sheep leaves the fold and removes one as each returns, and a pebble left over at dusk is a sheep unaccounted for. He has tracked the number without ever naming it, and the sheep and the pebbles share nothing beyond standing in one-to-one correspondence with none left over. That correspondence is what 7 records, and it is why collections with nothing else in common can be identical in number: sameness of number is sameness under pairing, not joint participation in a shared object. This is the content of Hume’s principle:¹⁸⁴ the number of Fs equals the number of Gs just when the Fs and Gs correspond one-to-one; on the present reading the principle fixes a fact about pluralities rather than introducing a name for an abstract particular.

Second. Structuralism (Shapiro, Resnik, Hellman) refines the classical position by holding that mathematical objects are not particular entities but positions in structures. The number 7 is not a particular abstract object but the seventh position in any ω-sequence (any infinite sequence ordered like the natural numbers, with a first element and a successor for every element); what mathematics studies is the structure of ω-sequences, not particular sequences. This is meant to handle the multiple-reductions problem (3 can be reduced to {{{{}}}} or to {{},{{}},{{},{{}}}}) by holding that any reduction that preserves the structure preserves what 7 is.

Structuralism makes the structural-conditions reply more direct, not less. The structures mathematics studies are constituted by determinate elements or positions and relations, directly supporting a difference-role. Some structures include asymmetric order; many do not. Their abstract stability does not by itself amount to temporal persistence or physical bearing. Upgrading the mathematical roles to T and Φ therefore requires a realization premise. Shapiro’s account of structures as patterns whose positions are individuated by their relations to other positions in the same structure is a description, in the mathematical case, of the distinctness, ordering, and persistence that §2.6 names. The reply is not that Shapiro is wrong about mathematical structure; the reply is that Shapiro’s account of mathematical structure describes what §2.6 names. The structural-conditions account is, on the present reading, structuralism extended to non-mathematical reality, the same insight applied to the wider domain of which mathematics is one structured region.

Third. The naturalized form (Maddy, Quine, Putnam) drops the strong realism about abstract objects and holds instead that mathematical objects are whatever the indispensability argument requires, the entities that are quantified over in our best scientific theories. The ontological commitment is to whatever is needed to make the science work; the metaphysics of the entities is left open.

The naturalized form does not aim to exhibit a counterexample to the structural account; it aims to deflate the metaphysical question. As deflation, it is irrelevant to the structural reply: whatever the indispensable mathematical entities turn out to be, they will be whatever the science requires, and the science requires distinctness, ordering, and persistence of mathematical facts. If the naturalist insists on a stronger reading (that mathematical entities have some metaphysical status that exempts them from the structural conditions) then the reply to the classical Platonist applies. If the naturalist remains deflationary, then the position is not in tension with the structural account.

Fourth. Plenitudinous Platonism (Balaguer) holds that every consistent mathematical theory describes a real domain of mathematical objects: the universe is mathematically plenitudinous, with every possible mathematical structure instantiated somewhere. This is meant to dissolve the access problem: we do not need a faculty for picking out particular abstract objects, because every consistent description picks out something.

Plenitude is a stronger claim than the classical Platonist makes, and the structural reply applies more straightforwardly. The plenitude of mathematical structures is not itself a further structured object but a description, at maximal extent, of what the structural conditions afford. To speak of the totality of consistent theories is to describe how such theories stand: individuated by their consistency conditions, ordered by relations of inclusion or interpretation, stable across all considerations of which structures exist. Each of those is a description of the distinctness, ordering, and persistence any possible structure would have, not a Time, Space, or Substance instantiated by some super-object ‘the totality.’ If the present account is right, plenitudinous Platonism is its mathematical correlate, not its refutation.

A kindred move is the divine-intellect Platonism: the abstract objects exist in God’s mind, and their atemporality is the atemporality of the divine intellect. The reply combines the present reply with §4.2.1 (the changeless-ground reply): the divine intellect that contains the structured realm of abstracts is itself structured, and the structuredness is what §2.6 names. The classical theist’s Platonism-plus-theology faces the structural reply twice over.

The modal-structuralist position (Hellman) is, on the present reading, not one more Platonism to absorb but the closest mathematical statement of the reply itself: mathematics is not about actual abstract objects but about possible structures. That is precisely the account given here: 7 is the structure of any possible seven-membered collection, and arithmetic is the study of what such possible structures are like. The only adjustment is the ground: the modality is not a primitive realm of its own but is fixed by the structural conditions of §2.6, which govern the whole space of the possible (§2.6.6). Mathematics describes the possible; the possible is structured by Time, Space, and Substance; so mathematics, far from naming exceptions to the conditions, is one of the most general descriptions of what they afford.

What converges. Four forms and none exhibits the counterexample the position needs: classical Platonism’s particular abstract objects resolve into descriptions of structured possibility, not exceptions to it; structuralism’s structures are constituted by exactly the relational facts the account names; the naturalized form is either deflationary (and not in tension) or stronger-than-deflationary (and faces the classical reply); and plenitudinous Platonism is the most extensive description of what structure affords, not a counterexample to it. Across the four forms, no mathematical view by itself supplies a counterexample to the necessity of structure in concrete change-realizing worlds. The stronger nominalist conclusion, that mathematical objects are not objects at all but descriptions of possible structure, remains one proposed interpretation, not a result forced on Platonism by §2.6 alone.

Sibling counter-constructions: the changeless-ground objection, on which a timeless being grounds the structure from outside, is §4.2.1. The B-theoretic block objection, on which the temporal manifold is a static four-dimensional reality, is §4.2.6. The McTaggart objection, on which time itself is unreal, is §4.2.5.

What the Platonist tradition genuinely captures is worth preserving. There is a sense in which mathematical truths are not merely conventional, not merely useful, not merely about marks on paper or sounds in the air. The persistence of mathematical truth across cultures and millennia, the convergence of independent investigations on the same theorems, the resistance of mathematical structure to revision in ways that empirical theories are revised, each speaks to something real about the subject matter of mathematics. The account of §2.6 is consistent with that sense. Mathematical structures are real structures; their reality is the reality of what the structural conditions afford, the count and pattern of every possible collection, described maximally and changelessly. The Platonist who reads the changelessness as exemption from §2.6 has misread the changelessness; what does not change is not a timeless object floating free of the conditions but the description of what every possible collection, every possible structure, would be, and that description does not change because the possibilities it describes do not. The stability is the stability of the possible, not the serenity of an abstract object.

Frege’s Grundlagen (1884) was printed on paper that has decayed in the intervening century and a half. The arithmetic he was investigating did not decay. What did not decay, on the present reading, is not a timeless object exempt from §2.6 but a truth about every possible collection: that seven of anything is the successor of six of it. The number 7 is not somewhere because it is not the kind of thing that could be anywhere: it is the count any seven-membered plurality has, and pluralities are where their members are. It does not move or decay because descriptions of what is possible do not move or decay; the possibilities they describe are fixed by the structural conditions that govern the possible. The Platonist’s number is not a counterexample to the present account, and it is not a hidden instance of it either. It is a description of what reality, under the conditions of §2.6, makes countable.

4.2.5 The Unreality of Time

In 1908 J. M. E. McTaggart published a paper in Mind titled “The Unreality of Time.” The argument is short, the conclusion is shocking, and the influence has been deep. Time, McTaggart claims, is unreal. The paper became one of the most cited works in twentieth-century philosophy of time, with well over a century of subsequent literature responding to it. If McTaggart was right, the present account of §2.6 has, in its first condition, named something unreal, and the bidirectional identity collapses on the temporal side.

The argument is concise enough to state in a paragraph. McTaggart distinguishes two ways of describing temporal events: the A-series (past, present, future, with the present moving) and the B-series (earlier than, simultaneous with, later than, fixed across the manifold). Time, McTaggart argues, requires the A-series: without becoming, there is no genuine temporality, only a static array of events with relations of earlier and later. But the A-series is contradictory. Every event is past, present, and future (these properties are mutually incompatible) and any attempt to remove the contradiction by relativizing (event E is future relative to time t1, present relative to t2, past relative to t3) generates a regress: the relativizing times are themselves events, each of which is past, present, and future, and so on. The A-series is incoherent; the B-series alone is insufficient for time; therefore time is unreal.

The argument has divided the philosophy of time ever since. Broad’s An Examination of McTaggart’s Philosophy (¹⁸⁵) gives the first sustained reply, defending A-series becoming against the contradiction charge; the moving-spotlight theory (the view that a four-dimensional manifold of events is timelessly arrayed and a privileged ‘now’ moves across it) is the image Broad had earlier introduced in Scientific Thought (1923) only to reject it. Prior’s Past, Present, and Future (¹⁸⁶) develops tense logic to formalize A-series claims. Mellor’s Real Time (¹⁸⁷) and Real Time II (¹⁸⁸) accept the contradiction in the A-series and develop a B-theoretic alternative on which tensed claims are reduced to tenseless truth-conditions. Tooley’s Time, Tense, and Causation (¹⁸⁹) defends a growing-block theory (on which the past and present are real but the future is not, with the block growing by accretion as time passes). Dummett (,¹⁹⁰ in Truth and Other Enigmas) reads McTaggart as a reductio of the assumption that the description of reality must be perspective-independent. Dainton (Time and Space, 2nd ed.)¹⁹¹ gives the standard survey. The contemporary positions (presentism, eternalism, growing-block, moving spotlight) are largely structured by which side of McTaggart’s argument they accept and which side they reject.

The target of the present reply is McTaggart’s conclusion: that time is unreal. The A-series/B-series question, at the level it operates within the philosophy of time, is largely orthogonal to the present account. The dispute is already bracketed at the structural level by §§2.6–2.7: the present account requires temporal ordering of states sufficient to support change, and §7.2.2 develops the composition view that resolves the dispute by holding A-series passage and B-series dimension together as two features Temporality requires rather than rival specifications of one. The reply here is narrower: McTaggart’s argument, granting its details, does not establish what its conclusion claims. It establishes at most a thesis about how to describe time, not a thesis about whether the structural conditions of temporality obtain.

Three forms of the move develop the position at three depths: the strict-unreality form, the A-series-incoherence form, and the perspective-dependence form. Each fails for the same structural reason at increasing levels of refinement.

First. The strict-unreality reading: McTaggart’s argument shows that time itself is unreal, full stop. There is no time, only the appearance of time, generated by some non-temporal feature of reality. This is the most extreme reading and the one McTaggart himself sometimes seems to endorse.

McTaggart’s writing and publication occurred within the apparent temporal order, so the performance of the argument instantiates temporal structure at the act level. The 1908 paper is itself an event with structural conditions: it is distinct from the 1907 draft and from the 1909 reception (S), it stands in temporal relations to those events (T), and it persists as a paper across the century of subsequent commentary (Φ). Its publication in Mind presupposes that the paper was first written and then published, not that the writing and the publication are non-temporally co-present in some atemporal mode. What that establishes is narrower than the date 1908 might suggest: the argument cannot dispense with temporal representation in its articulation. It does not by itself refute an error theory on which the apparent order is grounded in non-temporal reality. The object-side account must be tested separately, and the burden it carries is to explain the indispensable structure without real time.

The deeper structural point: the A-series/B-series dispute is internal to the analysis of T. If one of the two series is incoherent, then T must be analyzed in terms of the other series, or in terms of some third structure. McTaggart’s own C-series (the permanent, direction-neutral ordering of positions that survives the A-series’ collapse) is one such third structure, and the structural account partially vindicates it: the C-series’ bare ordering is real, the genuine residue of McTaggart’s positive view. What the structural T-role adds is precisely what the C-series withholds (asymmetric direction) supplied not by the contradictory A-series but by the self-exclusion the account grounds independently. So the affinity is the ordering and the difference is the asymmetry: the account preserves McTaggart’s positive residue while supplying the direction his negative argument leaves ungrounded. What the dispute does not show is that T fails to obtain. The structural condition T (directed ordering of states) is satisfied by either an A-series tensed analysis or a B-series tenseless analysis or some combination, and the choice between them is the topic of the philosophy of time, not a refutation of the existence of T.

Second. The A-series-incoherence form is the more defensible reading. McTaggart’s argument shows only that the A-series is incoherent, and a B-theoretic analysis of time is the proper response: time consists in the B-series, and the apparent reality of becoming is to be explained as a perspectival or psychological feature rather than as a metaphysical fact. Mellor’s Real Time II develops this position with full rigor.

The retorsion-from-1908 above closes the strict-unreality reading only; the serious B-theoretic tradition (Mellor, Williams, Sider) operates at this level and is engaged here and at §4.2.6. The B-series (earlier than, later than, simultaneous with) supplies the dimension feature of Temporality. The four-dimensional manifold of B-series events is distinguishable into events (S), ordered (T), and persists across the manifold as a unified four-dimensional whole (Φ). On the composition view at §7.2.2, Temporality requires both the dimension and the passage; the B-theorist who supplies the dimension cleanly and reduces passage to indexical content gives one feature where the composition requires two, and the disagreement lives there. The Mellor reading is one specification of Temporality’s dimension feature, not the complete metaphysics of T.

Third. The perspective-dependence reading (Dummett, Sider, the early-twentieth-century Cambridge reading) takes McTaggart’s argument as showing not that time is unreal but that any complete description of reality must be from no perspective (perspective-independent in the strong sense) and that the A-series, which is irreducibly perspectival, cannot figure in such a description. This is a thesis about the metaphysics of description rather than about the metaphysics of time directly.

The perspective-dependence reading is structurally interesting and partly correct. The A-series claims are perspective-dependent in a sense the B-series claims are not: ‘the present moment’ picks out different moments depending on when the claim is made; ‘earlier than 1908’ picks out the same moments regardless of when the claim is made. What the present account requires is that whatever specification of T is offered, the structural fact of directed ordering obtains. Whether the metaphysical description that captures the ordering is perspectival or perspective-independent is a separate question. The B-theorist offers the perspective-independent version; the A-theorist offers the perspectival version; both supply T. McTaggart on the present reading, like Dummett, raises a question about the metaphysics of description; the question is genuinely open, and the present account is consistent with either resolution.

A kindred move is the radical eliminativism (some readings of Parfit, some readings of Buddhist analyses): time is unreal because the relations it requires (between selves, between events, between moments) are themselves unreal. The reply combines the present reply with §4.2.8 (the momentariness reply): the relations are exactly the structural facts §2.6 names, and their elimination either collapses on its own description or relabels the structural facts under another vocabulary.

Another kindred is the McTaggart-via-presentism position: only the present is real, and the past and future are not, so the A-series cannot generate the contradiction because there is only the present moment. Presentism is a metaphysical thesis with its own challenges (the truthmaker problem, the cross-time relations problem, the problem of historical reference); the present account is consistent with presentism in the sense that the structural conditions can be specified at each present moment as a fact about what reality contains. Whether presentism is the right metaphysics of time is a separate question that the present account does not adjudicate.

What converges. Three forms clarify the dispute: strict unreality faces an act/object burden because its argument uses temporal representation but may still propose a non-temporal ground of appearance; A-series-incoherence is consistent with the present account, which admits B-theoretic specification of T; and perspective-dependence is a thesis about the metaphysics of description, not about whether T obtains. Across the three forms, McTaggart’s argument either fails on its own terms or is consistent with the present account.

Sibling counter-constructions: the B-theoretic block universe is treated separately at §4.2.6. The temporary-intrinsics objection, on which apparent change is the difference of intrinsic properties at different times rather than change of a single intrinsic, is §4.2.7. The momentariness objection is §4.2.8.

McTaggart’s argument belonged to a broader idealist program, and the natural worry is that the present account’s necessary, empirically unencounterable structure is itself a covert idealism rather than a neutral structural claim. The question is taken up directly at §4.3.5: whatever is taken as fundamental (a cosmic mind included) instantiates T, S, and Φ once it is structured enough to do the work the position needs of it, so the structural conditions are mind-compatible without being mind-derived. The idealist setting is engaged there, not set aside here.

What the McTaggart literature genuinely establishes is worth preserving: the conceptual analysis of time is harder than the pre-analytic vocabulary suggests, and the relation between A-series description and B-series description is one of the deepest structural questions in metaphysics. The present account does not adjudicate that question. It makes a weaker claim: the structural fact of directed ordering obtains, with the A/B dispute resolved by the composition view of §7.2.2. Hybrid views (growing-block and moving-spotlight) and presentism each carry the structural conditions in different vocabulary, with the metaphysical commitments developed there. McTaggart’s contribution to twentieth-century metaphysics is to have sharpened the question of how T is to be specified. The contribution does not extend, on the present analysis, to dissolving T altogether.

The 1908 argument survived its tradition: the British Hegelian idealism that gave it its broader setting was largely abandoned by the 1920s, but the distinction between A-series and B-series became a permanent landmark of analytic philosophy of time. The argument survives because the question it raises is real, even if the conclusion it draws is not. Time is real. Its analysis is hard. McTaggart helped make the analysis harder, and the present account of §2.6 is, on the present reading, one beneficiary of the long literature his paper provoked.

4.2.6 Block Universe

The B-theorist of time (Mellor, Smart, Williams, Sider) holds that the world is a four-dimensional manifold in which all events, past, present, and future, are equally real. There is no objective becoming; the apparent flow of time is a feature of conscious experience or of the indexical content of tensed claims, not a metaphysical fact about the manifold. The block-universe view is an influential metaphysical interpretation of relativity: the relativity of simultaneity makes a privileged present problematic, and the four-dimensional manifold is a natural setting for relativistic physics, though the interpretation remains contested. If the B-theorist is right, the present account of §2.6 must be reinterpreted: T is the temporal extension of the block, not the directed flow of becoming, and the structural conditions obtain across the block as a whole rather than as a feature of a moving present.

The move is influential and the position is widely held. Smart’s “The River of Time”¹⁹² and Philosophy and Scientific Realism¹⁹³ give the position its mid-century analytic articulation. Williams’s “The Myth of Passage”¹⁹⁴ is the canonical statement of the case against objective becoming. Mellor’s Real Time¹⁹⁵ and Real Time II¹⁹⁶ develop the most careful contemporary defense, on which tensed claims are reduced to tenseless truth-conditions and the apparent flow of time is a perspectival feature. Sider’s Four-Dimensionalism¹⁹⁷ defends the block as the proper setting for ordinary objects, with persistence cashed out as having temporal parts at different times. Putnam’s “Time and Physical Geometry”¹⁹⁸ and Rietdijk’s¹⁹⁹ argue that special relativity forces the block. Maudlin’s The Metaphysics Within Physics²⁰⁰ defends time’s fundamentality against the standard B-theoretic readings. Price’s Time’s Arrow and Archimedes’ Point²⁰¹ argues for a thoroughgoing block view on which temporal asymmetry is a perspectival illusion.

The target of the present reply is the structural claim: that the block universe is a static four-dimensional manifold without genuine becoming, and therefore that the structural conditions §2.6 names (if they obtain at all) obtain in a sense compatible with the absence of becoming, and possibly in a sense weaker than the present account requires. The reply is not that the B-theory is incoherent (it is not) but that the B-theoretic block instantiates the structural conditions of §2.6 in the sense the account requires. The block is not a counterexample to the present account; it is one of the available metaphysics of how the structural conditions are instantiated.

Three forms of the move develop the position at three depths: the strict-static-block form, the relativistic form, and the perspectival-becoming form. Each is consistent with the present account at the level the account operates.

First. The strict-static-block reading: the four-dimensional manifold is static in the sense that no genuine becoming obtains. The events of the manifold are timelessly arrayed, and the appearance of becoming is illusory. This is the most extreme reading and is sometimes endorsed by readers of Williams or by readers who take the relativistic argument to its eliminativist limit.

The strict-static-block reading must distinguish two senses of ‘static.’ In one sense, the block is static in that the manifold as a whole does not itself change: there is no super-time in which the four-dimensional manifold is arrayed and across which it evolves. In this sense, ‘static’ is correct and uncontroversial. In a second sense, ‘static’ would mean that the events within the manifold do not stand in temporal relations: that no event is earlier or later than any other event, that there is no directed ordering. In this second sense, the block is not static; the temporal extension of the manifold is what makes it a four-dimensional manifold rather than a three-dimensional one. The events within the block stand in B-series relations (earlier than, later than, simultaneous with) and these relations supply the directed ordering T requires. The B-theorist who denies becoming in the metaphysical sense (no objective A-series flow) does not deny temporal relations in the structural sense (B-series ordering across the manifold). The present account requires the latter, not the former.

The events within the block are distinguishable from one another (S). They stand in temporal relations of earlier and later (T). The manifold is a unified four-dimensional whole, but it does not itself persist through a further time. Within it, worldlines, temporal parts, or other cross-time relations may realize the bearing and continuity role associated with Φ. The block’s unity should not be redescribed as temporal persistence of the whole. The B-theoretic block instantiates the dimension feature of T cleanly. On the composition view developed at §7.2.2, Temporality requires both the dimension and the passage as two features held together; the block-only reading without objective passage gives one feature where the composition requires two. The B-theorist who reduces passage to indexical content (Mellor, Williams, the perspectival-becoming form below) is in disagreement with the composition view on the metaphysical status of passage rather than consistent with it; the disagreement lives there. The B-theorist who allows objective passage through the manifold is the composition view in B-vocabulary.

Second. The relativistic form is the central motivation for the block. Special relativity’s relativity of simultaneity is incompatible with a privileged objective present: different inertial frames disagree on which events are simultaneous, and there is no frame-independent fact of the matter. The four-dimensional manifold of Minkowski spacetime is the natural setting; events stand in the light-cone structure (causal past, causal future, elsewhere) rather than in a frame-independent A-series (the tensed past/present/future ordering with the present moving); the block is what relativity gives us.

The relativistic argument is correct as far as it goes, and the present account accommodates it without revision. T (directed ordering of states) does not require a frame-independent global present. It requires that events stand in temporal relations sufficient to support change, and the light-cone structure of Minkowski spacetime supplies exactly that: every event has a determinate causal past (events that can have influenced it) and a determinate causal future (events it can influence). The ordering is partial rather than total (events in each other’s elsewhere do not stand in a determinate temporal relation) but partial ordering is sufficient for the present account. Change is local; the structural conditions §2.6 names are conditions on local processes of change, not on the existence of a global frame-independent present. Relativity makes the present account more careful, not less correct.

A note on the contested status of the inference itself. The move from special relativity to the four-dimensional block is not a settled physics result. Neo-Lorentzian interpretations, John S. Bell,²⁰² Maudlin’s Quantum Non-Locality and Relativity,²⁰³ and Craig’s Time and the Metaphysics of Relativity²⁰⁴ most directly, preserve a privileged frame consistently with the empirical content by treating Lorentz contraction and time dilation as dynamical effects on rods and clocks rather than as features of the spacetime structure itself; the Putnam–Stein exchange²⁰⁵ and Balashov–Janssen²⁰⁶ press the orthodox four-dimensionalist reading against these alternatives. The dispute is genuine, and the present account does not contest either reading: on the orthodox reading T is supplied by light-cone partial ordering across the manifold, on the neo-Lorentzian reading T is supplied by a privileged dynamical frame, and on either reading the directed ordering the structural conditions §2.6 name is instantiated.

Third. The perspectival-becoming form (Mellor’s mature position; Williams’s “Myth of Passage”; Price’s perspectival arrow) accepts the block at the metaphysical level and locates becoming at the perspectival level: the apparent flow of time is a feature of conscious experience, of indexical content, of the perspective of an observer embedded within the block, not a fact about the manifold itself.

The perspectival reading is consistent with the present account on the dimension feature of T and in tension with it on the passage feature. The structural conditions are instantiated by the manifold; the Mellor account preserves the dimension feature of T (B-series ordering), preserves S (event-distinctness), preserves a cross-time continuity or temporal-parts structure capable of realizing the Φ-role, and offers a story about how perspectival experience of becoming arises from the structural facts. On the composition view at §7.2.2, the passage feature of Temporality is metaphysically real rather than indexical, and Mellor’s reduction is what the composition view declines. The story about perspectival becoming is consistent with the present account where it explains how the metaphysically-real passage registers from within a configuration that tracks its own position; it is not consistent where the becoming is the whole of what passage names. The disagreement is local to the passage feature; the rest of the Mellor architecture stands.

A kindred move is the growing-block theory (Tooley, Broad’s earlier position): the past and present are real, the future is not, and the block grows by accretion as time passes. Growing-block is structurally a hybrid: it accepts B-series relations among the past-and-present events, accepts a privileged present (the leading edge of the block), and posits an objective process of accretion. The present account is consistent with growing-block in the sense that the growing block instantiates T (across the past-and-present), S, and Φ; the disputes between growing-block and the strict-block (and between both and presentism, on which only the present moment is real, the past having ceased and the future not yet existing) are internal to the metaphysics of which structures are instantiated, not whether the structural conditions obtain.

Another kindred is the moving-spotlight theory (Cameron, Skow): the block is timelessly there, and a privileged present moves across it. The present account is consistent with the moving-spotlight in the sense that the spotlight obtains within an already-structured block; the spotlight is one structural feature among others, supported by the underlying structure.

What converges. Three forms and each is consistent with the present account at the level it operates: the strict-static-block reading distinguishes two senses of ‘static’ and yields the structural conditions in the sense the account requires; the relativistic form replaces a global present with light-cone structure and supplies T as a partial ordering; and the perspectival-becoming form locates becoming in conscious experience while preserving the structural conditions at the metaphysical level. Across the three forms, the B-theoretic block instantiates T ∧ S ∧ Φ, sometimes more cleanly than the A-theoretic alternatives.

Sibling counter-constructions: the McTaggart-unreality objection is §4.2.5. The temporary-intrinsics objection, on which apparent change is the difference of intrinsic properties at different times rather than change of a single intrinsic, is §4.2.7. The emergent-time objection, on which time is derivative rather than fundamental, is §4.2.3.

A consequence for the Kalām engagement (§5.1.2). On the composition view at §7.2.2, the passage feature of Temporality is metaphysically real; began to exist therefore has determinate structural content. What the composition view marks is two readings of the begin-question: a quantitative reading (how far back the passage runs along the dimension) engaged at §5.1.2 and a categorial reading (a moment-prior-to-T from which T began) foreclosed by Parasitic Negation. The strict-B-theoretic reduction of passage to indexical content is a separable issue that does not absorb the begin-question in the way Craig’s pressure assumes; the composition view’s commitment to objective passage is what the §5.1.2 engagement runs on.

What the B-theoretic literature genuinely establishes is worth preserving: it gives powerful reasons to question naïve assumptions about a frame-independent present, objective passage, and the metaphysical privilege of the present moment. B-theorists commonly locate apparent flow and presentness in conscious experience or the indexical content of tensed claims; A-theoretic and neo-Lorentzian rivals contest that interpretation. These are live philosophical disputes, and the present account is designed to state what it preserves across them. What the present account adds is that the four-dimensional manifold the B-theorist describes is not an alternative to T ∧ S ∧ Φ but a particular specification of it. The events of the manifold are distinct (S), the manifold has temporal extension (T), and the manifold’s structure underwrites the cross-time identity of objects (Φ). The block is what the dimension feature of T looks like when it is described by the metaphysics of relativistic physics.

The block is the standard backdrop now in a way it was not in McTaggart’s day. The present account, on the present reading, is what the block is when it is described not as a metaphysics of time but as a metaphysics of the structural conditions that any time must instantiate. The block’s events are distinct and ordered, and its cross-time structure bears the persistence Φ names; that is what the block consists in, and that is what §2.6 names.

4.2.7 The Problem of Temporary Intrinsics

David Lewis (On the Plurality of Worlds)²⁰⁷ names the problem of temporary intrinsics; a simple case makes it vivid: a banana is green at t1 and yellow at t2, and the banana cannot be both green and yellow simpliciter (in the unqualified, unrelativized sense), since those properties are incompatible. Some way of relativizing the predication is needed, and the analysis one chooses for it is largely determinative of one’s metaphysics of persistence and change.

The problem is sharp because the available solutions all carry costs. Presentism (only the present is real) avoids the contradiction by holding that the banana is yellow simpliciter at t2 (t2 is the present) and there is no greenness for the contradiction to obtain with. But presentism faces difficulties about cross-time relations and historical truth. Four-dimensionalism (perdurance) holds that the banana is a four-dimensional object with temporal parts; the t1-stage is green, the t2-stage is yellow, and there is no contradiction because the green and yellow are properties of distinct stages. But four-dimensionalism faces the standard objection (Lewis himself raises it) that ordinary objects do not seem to have temporal parts in the way they have spatial parts. The relational-properties account holds that ‘green at t1’ and ‘yellow at t2’ are time-relativized properties that the banana has wholly and simpliciter; the contradiction is dissolved because the properties have time-indices built in. But the relational-properties account makes intrinsic properties relational, which Lewis argues is to misdescribe what ‘intrinsic’ means.

The target of the present reply is the structural claim that the temporary-intrinsics analysis, in any of its three forms, undermines the account of §2.6. The argument the objector might run: change, on every available analysis, turns out to be either the difference of intrinsic properties at different times (presentism, with cross-time relations as a separate problem), or the difference of intrinsic properties of different temporal parts (perdurance), or the relational property of the same object at different times (relationalism). On none of these analyses is change a single thing standing in different conditions across time, the picture §2.6 names. So the present account is, at best, a description of one of these analyses, and likely a description of none of them.

The reply is that the present account is consistent with all three analyses of temporary intrinsics, and indeed is the structural fact each of them is an analysis of. The structural conditions §2.6 names, Time (T), Space (S), and Substance (Φ), generating the structural features of ordering, distinctness, and persistence, are at a level above the metaphysics of how persistence is realized; they are conditions any analysis of persistence must satisfy. Each of the three analyses satisfies them in a distinct way. The dispute among the three is internal to the metaphysics of persistence; the present account is what the disputes are disputes about.

Three forms of the move develop the position at three depths: the presentist form, the perdurantist form, and the relationalist form. Each is consistent with the present account.

First. Presentism (Prior, Bigelow, Markosian) holds that only the present is real. The banana is yellow simpliciter at the present moment t2; the banana was green at the past moment t1, but that past moment is not now real, only its having-been-real is now real. Tensed truths are taken as primitive: ‘the banana was green’ is true now, but its truth-maker is not a present-tense fact about a past entity but a primitive past-tense fact.

Presentism is consistent with the present account at each present moment. The banana at t2 is distinct from the seed it grew from and from the compost it will become (S). Presentism cannot straightforwardly relate the present banana to presently nonexistent past and future stages. It instead invokes primitive tensed facts, ersatz times, or another truthmaking account. Those resources may realize the ordering role, and the reply assumes no literal relations to non-real relata. The banana persists across changes (the banana that is now yellow is the same banana that was green) and the persistence is the structural fact that the cross-time identity obtains, however the metaphysics of cross-time relations is specified (Φ). Presentism’s specific contribution to the metaphysics of temporary intrinsics is to hold that the cross-time relations are tensed primitives rather than relations to wholly-real past or future entities.

Second. Four-dimensionalism (Lewis, Sider, Heller, Hudson) holds that the banana is a four-dimensional object with temporal parts. The t1-stage of the banana is green; the t2-stage of the banana is yellow; the four-dimensional banana itself, the whole, is partly-green-and-partly-yellow in the way a striped flag is partly-red-and-partly-white. There is no contradiction because the green and yellow are properties of distinct temporal parts.

Four-dimensionalism is structurally a maximally clean specification of T ∧ S ∧ Φ. The temporal parts are distinct (S). They stand in temporal relations of earlier and later (T). The four-dimensional whole does not endure from one time to another; it has temporal parts at different times. Its perdurance and the relations unifying those parts offer one realization of the Φ-role, but calling the timelessly extended whole persistence in the strongest sense would beg the question against endurantism. The four-dimensionalist has not eliminated persistence; they have given the most explicit possible analysis of it.

Third. Relationalism (Mellor, Lowe in some moods, Wasserman) holds that ‘green’ and ‘yellow’ are not strictly intrinsic but covert relations to times: ‘the banana is green-at-t1 and yellow-at-t2.’ The contradiction is dissolved because ‘green-at-t1’ and ‘yellow-at-t2’ are compatible properties; the banana has them both wholly and simpliciter, in the way it has the property of being-three-feet-from-the-window and the property of being-five-feet-from-the-door wholly and simpliciter without contradiction.

Relationalism is consistent with the present account. The relations to times require times for the relations to be relations to (T). The relations distinguish the banana’s condition at t1 from its condition at t2 (S in the relational structure). The persistence of the banana (the same banana having both relations) is the structural fact that one entity stands in both (Φ). The relationalist’s specific cost (making intrinsic properties relational) is a cost internal to the metaphysics of properties, not a cost to the structural account.

A kindred move is the adverbialist account (van Inwagen, in “Four-Dimensional Objects”): the banana is green at t1 in the sense of being t1-ly green, where ‘green’ is modified adverbially by the time. The adverbialist hopes to keep intrinsic properties intrinsic by relocating the time-relativization from the property to the manner of having the property. The reply is that the structural conditions are satisfied either way: the manner of having the property at t1 is distinct from the manner of having the property at t2 (S in the manners), the manners stand in temporal relations (T), and the entity persists across the manners (Φ).

Another kindred is the soft-presentist combination (Crisp, Markosian’s later work): only the present is real, and tensed truths are primitive, but the structural fact that the world has been and will be in determinate ways is itself a primitive feature of present reality. The reply is that the structural facts the soft-presentist takes as primitive are the structural facts §2.6 names, taken as primitive. The present account is consistent with primitivism about the structural facts.

What converges. Three forms and each instantiates the structural conditions of §2.6 at a level the account operates above. Presentism instantiates them at each present moment with cross-time relations specified by tensed-fact primitivism. Four-dimensionalism instantiates them in the timelessly-extant four-dimensional manifold. Relationalism instantiates them in the relational structure of properties indexed to times. Across the three forms, the temporary-intrinsics problem is a problem about the metaphysics of persistence, and the present account is what the metaphysics is the metaphysics of. The level-above placement is a deliberate trade, worth stating in Lewis’s own cost-benefit idiom. What it buys: neutrality, the structural conditions are secured without adjudicating the persistence dispute, and what all three rivals share is made explicit rather than left implicit. What it costs: discriminating power, the account does not, and does not try to, settle which of presentism, perdurance, or relationalism is correct, and it commits to structural facts of ordering, distinctness, and persistence whose own individuation is specified at §2.6 rather than here. For the present argument the purchase is the right one: the thesis of §2.6 needs only what every analysis of persistence must satisfy, so buying neutrality at the price of adjudication costs the argument nothing it required and secures exactly what it did.

Sibling counter-constructions: the changeless-ground objection, on which a timeless being grounds the structure from outside, is §4.2.1. The B-theoretic block, on which the four-dimensional manifold is a static reality, is §4.2.6. The momentariness objection, on which persistence is illusory, is §4.2.8. The McTaggart-unreality objection is §4.2.5.

What the temporary-intrinsics literature genuinely establishes is worth preserving: ordinary change is harder to analyze than the pre-analytic vocabulary suggests. The banana that ripens is not in the simplest possible relation to its earlier self; the relation requires either a metaphysics of temporal parts, or a metaphysics of relational properties, or a metaphysics of tensed primitives, or some other specification of how cross-time identity is sustained. The present account does not adjudicate among these. It makes a weaker claim: whatever specification of cross-time identity is correct, the cross-time identity itself is a structural fact of the kind Φ names. The banana that ripens is analyzed differently across the views: as one enduring object, as a four-dimensional whole with differently qualified parts, or as appropriately related stages. What survives neutrally is an ordered continuity structure, not strict numerical identity on every account. The present account is the level at which the same is sameness regardless of which specification is correct.

Lewis introduced the temporary-intrinsics problem in 1986, in the brief passage on endurance and perdurance in On the Plurality of Worlds (pp. 202–204) that has occupied subsequent metaphysics for decades. Lewis’s own resolution was perdurance, but the problem outlived his answer; presentism, perdurance, and relationalism remain the three dominant analyses, and the dispute among them is one of the central debates of contemporary metaphysics of persistence. The banana ripens, on the present reading, by being one banana standing in successive conditions across an ordering of times, with the metaphysics of how the standing-across is realized to be specified separately. The standing-across is the structural fact. The metaphysics of its specification is the topic of the dispute. The former is named by §2.6; the philosophy of persistence elaborates the latter.

4.2.8 Stage Theory

What if there is no persistence? What if the world is a sequence of momentary flashes: each instant a complete fact unto itself, each one followed by another that is not it, each connected to the next by nothing but resemblance and causal succession? The Yogācāra Buddhists call this kṣaṇabhaṅga, momentary destruction. The contemporary analytic version is stage theory, or four-dimensionalism with temporal counterparts. Different traditions, different vocabularies, one structural proposal: persistence is illusory, and what we call a continuing thing is a series of distinct stages bound only by relations of similarity and causal continuity.

The move is radical and the consequences are drastic. If it succeeds, the structural account of §2.6 collapses at Φ, the persistence condition. Without persistence, there is nothing across which Change can occur. Without persistence, the bidirectional identity of change-permission and T ∧ S ∧ Φ has no third leg to stand on. The classical theist, sensing the opening, sometimes adopts a momentarist line: God’s relation to the world is a sequence of momentary creative acts, each instant freshly produced, with no need for persistence running through them. Pratītyasamutpāda (dependent origination) supplies the metaphysical scaffolding. Sider supplies the contemporary analytic version. The objection deserves a careful answer.

The Yogācāra school of Mahāyāna Buddhism develops kṣaṇabhaṅga most rigorously: every existent (bhāva) is a kṣaṇa (momentary instant), arising and ceasing in the same instant, replaced by a successor that is causally continuous but numerically distinct. Vasubandhu (4th–5th c.) and Dharmakīrti (7th c.) elaborate the doctrine; the ālaya-vijñāna (storehouse consciousness) and the bhāva-santāna (continuum of being) provide the explanatory work that ordinary persistence would otherwise do. Nāgārjuna’s Mūlamadhyamakakārikā presses the analysis further: nothing has svabhāva (intrinsic nature), and śūnyatā (emptiness) is the proper analysis of every apparent persistent.

In analytic philosophy, the structurally similar position is stage theory: Sider (Four-Dimensionalism)²⁰⁸ develops the most rigorous version, on which ordinary objects are momentary stages bound by genidentity (Lewin’s term, taken up by Carnap and Reichenbach, for cross-time identity-by-causal-continuity rather than strict numerical identity) or counterpart relations. Hawley (How Things Persist)²⁰⁹ defends a closely related view. Lewis (On the Plurality of Worlds;²¹⁰ Philosophical Papers II)²¹¹ supplies the perdurantist framework. Quine’s stage-theoretic readings of ordinary objects in Word and Object (²¹²) and Russell’s logical-construction program are ancestral. Heller (The Ontology of Physical Objects)²¹³ and Hudson (A Materialist Metaphysics of the Human Person)²¹⁴ develop the position in different directions.

The target of the present reply is the structural claim: that the world is a sequence of momentary stages bound by external relations only, with no persistence operating across them. The philosophical opponents of stage theory, endurantists like van Inwagen (Material Beings),²¹⁵ Wiggins (Sameness and Substance),²¹⁶ Lowe (The Possibility of Metaphysics),²¹⁷ who hold that ordinary objects are wholly present at each moment of their existence rather than composed of momentary stages, have written extensively against perdurance. The present reply is narrower than their full case: it locates the structural commitment that any momentarism must make to be the position it claims to be, and shows that the commitment instantiates the structural roles the position is defined as denying.

Three forms of the move develop the position at three depths: the strict-momentarism form, the causal-continuity form, and the counterpart-theoretic form. The reply argues that each fails for the same structural reason at increasing levels of refinement.

First. Strict momentarism (the cleanest form of the doctrine) claims that each instant is a complete fact unto itself, with no relations to any other instant. Each kṣaṇa arises and ceases in the same instant, and what we call succession is a cognitive imposition on what is in fact a sequence of unrelated complete facts.

Strict momentarism cannot describe its own claim. To say that there are many kṣaṇas, in succession, is to relate them, and the relation of succession is exactly what is being denied. The position requires either that the kṣaṇas be related or ordered, in which case a cross-stage structural fact is admitted, or that they be wholly unrelated, in which case the theory cannot recover succession. Wholly unrelated plurality does not literally collapse to one fact, but it cannot constitute a sequence without an ordering relation. The strict-momentarist version of the doctrine collapses on its own description: either the position is false because succession obtains, or the position is empty because nothing in it can be a succession of anything.

Second. The causal-continuity form weakens the claim: each moment is distinct, but moments are bound by causal relations. The bhāva-santāna (the continuum of being) is a causal stream, not a persistence condition; the appearance of a continuing self or object is the cognitive trace of the causal continuity, not the deliverance of an underlying substance.

A causal relation between distinct kṣaṇas requires a structural fact connecting them. That supplies directed dependence and cross-stage bearing, but it does not by itself establish numerical identity or persistence of one bearer. The momentarist may accept causal continuity while denying strict persistence; the structural account must therefore distinguish the thin Φ-role from an endurantist persistence type. Bhāva-santāna names what Φ names: the structural condition that makes A’s effect on B a real relation rather than a coincidence of pattern. If the doctrine is read as eliminating Φ, then the causal continuity it appeals to has nothing to run on and no force; if the doctrine is read as preserving Φ under a different name, then it has not eliminated persistence, only relabeled it. The Yogācāra commentators are aware of this tension; the elaborate apparatus of ālaya-vijñāna (storehouse consciousness as the carrier of karmic seeds across instants) is precisely the structural persistence the strict doctrine forswears, smuggled back in to do the explanatory work the strict doctrine cannot do without. The relabeling is honest about the explanatory need; what it does not show is that the structural role has been eliminated.

Third. The counterpart-theoretic form goes deepest: stages are momentary, but the genidentity relation (or the counterpart relation, or the appropriate cross-time similarity-plus-causation relation) ties them into the unified entities ordinary discourse names. The claim that this stage is the same person as that earlier stage is true under genidentity even though strict numerical identity does not obtain. Sider develops the position with full rigor; the technical apparatus is sophisticated and the resulting metaphysics is internally coherent.

The genidentity or counterpart relation is one proposed surrogate for persistence, not strict persistence under another vocabulary. For genidentity to obtain between this stage and that earlier stage, there must be a structural fact (not internal to either stage, since each is momentary, but extending across them) that makes the relation hold. Whatever fact it is, it is a structural fact spanning instants, and that is what Φ consists in. Sider’s own stage theory is more precisely distinguished from perdurance: stages are not temporal parts of a four-dimensional whole but the primary objects, with persistence analyzed as a counterpart relation among numerically distinct stages (Four-Dimensionalism).²¹⁸ On stage theory, the persistence structure is the counterpart-genidentity relation itself, which may realize the thin Φ-role as cross-stage continuity. Stage theory nevertheless denies that the later stage is numerically the same entity as the earlier stage; the relation supports survival-talk without restoring strict identity. On perdurance, the persistence structure is the four-dimensional whole of which stages are temporal parts. Both posit a structural fact spanning instants; they differ on whether that fact is a relation among stages or a whole of which they are parts. The question of which is more fundamental is not settled here and need not be: either description preserves Φ. The reply is not that perdurance or stage theory is incoherent (neither is) but that neither is a refutation of Φ: each is Φ in a different decomposition. The endurantist holds that a thing wholly present at each moment persists; the perdurantist holds that a thing has temporal parts at each moment and the four-dimensional whole persists. Both posit a structural fact spanning instants. Where they differ is on the metaphysics of the parts, not on whether persistence (in the structural sense §2.6 names) obtains.

A kindred move is the bundle theory variant: the world is a bundle of tropes or property-instances, with no underlying substance. Each instant is a co-presence of tropes; the apparent persistence of an object is the resemblance and causal connection between successive bundles. The reply is structurally identical: either the bundles are unrelated (no succession, no persistence, no anything) or related (the relations are the structural facts §2.6 names). The bundle theorist who gives accounts of cross-time identity at all gives an account of structural persistence under another name.

Another kindred is the radical Heraclitean: nothing persists; everything is in flux; the river is never the same river twice. The reply is the same. To say that the river is in flux is to say that something stands in the relation of being-the-changing-pattern across the flux: the bed, the volume, the channel, or the whole spatiotemporal manifold of the river-event. Without structural persistence of some kind, ‘flux’ has no subject and the claim has no content. Heraclitus needs more structure than his slogan announces, and the more careful presocratic readings (Kahn, The Art and Thought of Heraclitus)²¹⁹ acknowledge as much.

What converges. Three forms and none lets the structural elimination stand. Strict momentarism collapses on its own description: succession is the relation it denies. Causal-continuity smuggles Φ back in under the name bhāva-santāna or its analogs; the explanatory need is met only by reintroducing exactly what the strict doctrine forswears. The counterpart-theoretic form is internally coherent but is not a refutation of Φ; it is Φ in a different decomposition, with the counterpart-genidentity relation (or, on perdurance, the four-dimensional whole) as the persistence analog. The endurantist and the perdurantist agree on the structural fact; they disagree on its metaphysics of parts. The question of which description is more fundamental (the four-dimensional whole, the momentary stages in counterpart relations, or the wholly-present enduring object) is a question within the structural framework and answerable within it. It does not bear on whether Φ obtains. Sider’s stage theory is a fundamentality proposal within a structurally committed metaphysics: it specifies how the persistence structure is best analyzed, not whether it exists. The meta-position held here is that fundamentality questions of this kind are downstream of the instantiation question: once Φ is established as a necessary structural condition, further questions about its realization (by temporal parts, by counterpart relations, or by simple endurance) remain open to metaphysical investigation without threatening the structural conclusion.

Sibling counter-constructions: emergent-time positions, on which time itself is derivative rather than fundamental, are §4.2.3. The Shoemaker objection, on which time can obtain without change, is §4.2.2. The temporary-intrinsics objection, on which apparent change is the difference of intrinsic properties at different times rather than change of a single intrinsic, is §4.2.7.

The deepest form of the momentarist insight is one the structural account preserves rather than denies: there is something importantly correct about the observation that nothing in the world is fixed, that every object is in some sense an event, that what we call a thing is the persistent pattern of its becoming. The structural account names that pattern Φ (the persistence condition) and locates it as one of three structural roles whose joint instantiation is what change consists in. The momentarist who pushes the insight to the elimination of Φ has gone too far: persistence is not a substance behind the pattern but is the pattern itself, the structural fact that there is a pattern at all rather than a sequence of unrelated fragments.

4.2.9 Existential Inertia

Two traditions agree that things persist and disagree about why. The classical theist holds that no contingent thing keeps itself in being for an instant: existence must be conserved, moment to moment, by a being that holds it there, on pain of annihilation. The existential-inertia theorist denies the dependence: a thing that exists tends to remain in existence on its own, absent a sufficient destroyer, and needs no concurrent sustainer to persist. This book sides with the inertialist’s conclusion and rejects the inertialist’s ground. Nature is not held in being by anything outside it; there is no concurrent divine sustainer to invoke. But neither does it persist by brute default. It persists because Φ (the retention of identity across an interval) is what persistence consists in. The conservation debate is conducted over a gap the structural account closes from below.

The conservationist case is Thomist in origin. Aquinas holds that God’s causality is required not only to bring things into being but to keep them there: creatures stand to God as the air stands to the sun’s illumination, ceasing the instant the influx ceases (ST;²²⁰ De Potentia).²²¹ Freddoso (“God’s General Concurrence with Secondary Causes,”;²²² and the 1994 sequel)²²³ reconstructs the medieval taxonomy, mere conservationism (God conserves things in being but creatures produce their own effects), concurrentism (each created effect is produced by God and creature together), and occasionalism (God is the only real cause, creatures mere occasions for his acting), and defends the concurrentist reading on which divine sustaining is required for every created effect. Feser (Aquinas;²²⁴ Five Proofs of the Existence of God)²²⁵ presses the demand through act and potency: a thing’s existing at each moment is a potency that must be actualized by something already actual, terminating in Pure Act. Against the demand stands the existential-inertia thesis. Beaudoin (“The World’s Continuance: Divine Conservation or Existential Inertia?”)²²⁶ and, in its most developed form, Schmid (“Existential Inertia and the Aristotelian Proof,”)²²⁷ and Schmid and Linford (Existential Inertia and Classical Theistic Proofs)²²⁸ argue that persistence is the default state of concrete objects and that the conservationist owes a positive argument the tradition has not supplied. Kvanvig and McCann (“Divine Conservation and the Persistence of the World,”)²²⁹ hold the conservationist line with the sharpest form of the counter-challenge: what grounds a thing’s continuing to exist from one moment to the next, they ask, if not a sustaining cause? The structural account answers that question without adopting either side’s location for the ground.

First. The conservationist argument runs on a hidden premise: that continued existence is a kind of change, a fresh actualization at each moment, a coming-to-be repeated. Feser’s act/potency version makes the premise explicit: existence-now is a potency, and a potency requires a concurrent actualizer. Grant the framework and the conclusion follows; the demand is only as strong as the premise that persistence is a transition. It is not. A transition is a passage from one state to another. Mere continued existence need not be a fresh transition, although a persisting thing may undergo many qualitative changes while remaining the same thing. The conservationist must therefore argue that existing at a later time is a continuously derived effect, not simply infer that persistence as such is the absence of all change. The act/potency machinery earns its concurrent cause where there is becoming to account for; applied to mere continuation it manufactures a becoming that is not there and then requires a cause to sustain the manufactured becoming. Continuation is not continuous becoming. Nothing crosses from potency to act in a thing’s simply remaining what it is, and so nothing is required to drive the crossing. The sun-and-illumination analogy (ST)²³⁰ imports exactly this picture (a continuous effect requiring a continuous cause) and holds only if continued being is an effect of the relevant kind. Illumination is a process; existence is not. (The simultaneous-causation machinery the Thomist offers to make the sustaining concurrent rather than prior is taken up at §5.1.3; the point here is upstream of it: there is no sustaining-effect to make concurrent.)

This defeats the act/potency form of the demand, but not every form of it. The concurrentist need not say that persistence is a becoming at all. Freddoso’s continuous-creation reading (;²³¹)²³² treats created existence as a continuously-derived effect: dependent at every instant on a sustaining cause, yet with no transition, nothing crossing from potency to act, no becoming to drive. So construed, the demand never leans on the transition premise and survives its denial, but it survives as a demand for a ground, in virtue of what the thing exists now, not for a cause of change, and a grounding demand is met below rather than here. Two replies converge on it. The ground the concurrentist places in an external sustainer is supplied one level down by Φ (third move): once continued existence has a structural analysis, a continuously-deriving sustainer is not thereby shown redundant. Excluding the sustainer requires an additional argument that the constitutive account exhausts the relevant explanatory demand. And independently, as Chapter 6 will argue (§6.1), no candidate is available to do that work: a being answering to the classical specification (atemporal, aspatial, immaterial) is ∅ under that description and could derive nothing, so even a sound continuous-derivation demand would have nothing to name as the deriver. The transition form fails here; the continuous-effect form is answered downstream, not waved past.

Second. Conceding that persistence is not a transition does not end the matter, and the most precise opponent presses past it. Schmid’s question is not “what changes a persisting thing?” but “in virtue of what does it persist rather than not?”, a demand for a ground, not for a cause of change. Schmid and Linford (²³³) are right that the demand cannot be waved off: a thing’s continuing to exist is a fact about it, and facts about what is have grounds. The existential-inertia theorist who answers “it just does: persistence is the default” has named the explanandum (the fact to be explained) and called it an explanation. The structural account does not take that exit. Φ provides a constitutive analysis of persistence: persistence consists in the relevant retention or continuity of identity across an interval (§2.6). That answers what persistence is. It does not, merely by definition, answer the conservationist’s distinct explanatory question of why this particular thing continues rather than ceases; treating the constitutive analysis as a sufficient sustaining explanation would beg that question. This is deeper than any contingent disposition the object carries (deeper than a tendency-to-persist that might be present or absent, masked or triggered) because Φ is the condition under which the object is a determinate something at all, at either moment. The conservationist is right that persistence is grounded and not brute; the inertialist is right that the ground is not God. Both put the ground outside the persisting thing: the theist in a sustaining cause, the inertialist’s critics in a missing one. It is one level down, in the structural condition any persisting thing already instantiates by being a thing across time. This is the book’s inversion performed at the site of conservation: the feature the tradition reaches for (a sustaining ground beneath contingent existence) is real, and is not what the tradition takes it to be.

Third. Schmid’s modal-collapse instinct supplies the sharpest reply available to the opponent. If persistence is grounded in Φ, and Φ is necessary (the book argues exactly this) then is the persistence of every particular thing necessary too? If the ground cannot fail, and the ground is what holds the thing in being, then the thing cannot fail to persist, and the contingency of the world collapses into necessity. The worry is the existential-inertia analogue of the modal-collapse argument Schmid presses against divine simplicity, and it is the one objection a structural ground must answer or forfeit the field.

It does not go through, and the reason is a distinction the book has already drawn. The necessity of Φ is the necessity of the condition (that whatever persists does so by retaining identity across an interval), not the necessity of any particular persister’s retaining it. The level-distinction is fixed at §2.8.4: that the conditions of contingency are necessary does not make the contingent facts they condition necessary; the framework within which things come and go cannot itself come or go, while the things within it still do. L9, the non-propagation of constitutive grounding (§4.4.9), is the grounding-theoretic statement of that same distinction rather than a second defense: a constitutive ground hands its modal status to the condition it grounds, never to the particular actualizations that satisfy it. L9 is earned at §4.4.9 for the necessity of change: there □(T∧S∧Φ) and □◇C are not ground and grounded but one fact under two readings, so the necessary possibility of change carries its necessity by identity, not by transmission across a grounding relation. What the conditions ground constitutively, not productively are the particular changes, with the necessity propagating to none of them (□◇C, not □C). The persistence case runs that constitutive half of the schema once more, with one shift worth marking: at §4.4.9 the constitutive ground bears on each particular change, whereas here it bears on persistence under a constitutive condition, leaving the contingent de re fact (that any particular thing persists) outside the reach of the necessity.

Φ grounds persistence constitutively, not productively: it does not produce each thing’s continuation as an effect produces its output, which is the only way necessity would propagate Spinoza-fashion; it is what continuing-to-be consists in, the identity-retention condition any persister instantiates simply by being a thing across the interval.

So the □ attaches to the conditional, read de dicto (a claim about the proposition, not the thing), □(if x persists, x persists by Φ): necessarily, whatever persists does so by retaining identity across an interval. It does not attach to the de re □(x persists): no particular thing’s continuing is thereby made necessary. Φ’s being necessary no more makes this oak’s persistence necessary than the necessity of the conditions of value makes any particular thing valuable of necessity. The oak persists by Φ while it persists, and ceases when its identity-retention fails. Modal collapse is blocked not at two joints but at one stated twice: the level-distinction of §2.8.4 and L9 of §4.4.9 are the same non-propagation seen from the side of levels and from the side of grounding.

The nearest neighbors. Three positions also locate persistence in what the thing is, and the account must not blur into any of them. Dispositional essentialism grounds persistence in a first-order power to persist: a disposition the object carries, one that could be present or absent, masked or triggered. Φ is not such a power: a disposition is a further feature of an already-individuated thing, whereas Φ is the condition under which there is an individuated thing at all, at either end of the interval, to bear features.

Beaudoin’s naturalist inertia (²³⁴) reaches the right conclusion (no sustainer) by treating persistence as a brute default; it differs from the structural account exactly as the brute inertialist of the second move does, by leaving ungrounded what the structural account grounds.

And Spinozist necessitarianism grounds persistence in the necessity of substance, paying for it with the collapse just blocked: it runs the ground’s necessity straight into the persister, making every continuation necessary. The structural account takes the necessity of the ground together with the level-distinction, so the ground is necessary while the persister is not. The difference from each neighbor is one difference told three ways: Φ is a condition, not a first-order feature; the ground is structural, not brute; and the necessity in play is the necessity of that condition, not of whatever satisfies it.

The result feeds back into the case against the classical sustainer. The conservationist demand is supposed to deliver a being that does the sustaining. Once persistence is grounded structurally, the demand has no work left for such a being to do, and Chapter 6 will close the gap from the other side (§6.1), arguing that a being answering to the classical specification (atemporal, aspatial, immaterial) is ∅ under that description and could sustain nothing even if sustaining were required. The conservation argument needs both halves to stand, and both fail. The fuller engagements with the being the conservationist names are at §5.2.1 (the simplicity that is supposed to make it the unconditioned ground), §5.1.3 (the simultaneous causation supposed to make its sustaining concurrent), and §5.1.1 (the foundation argument that reasons toward it). The narrower result those depend on is that continued existence does not, by itself, hand the theist the explanandum the argument needs.

What holds. The existential-inertia debate is run between a conservationism that grounds persistence in a sustaining God and an inertialism that grounds it in nothing at all. The same move answers both. Persistence is not a transition, so it needs no concurrent actualizer (against the conservationist); but it is not brute either, so the inertialist’s default is not the last word (against the strongest inertia critics). Persistence is grounded in Φ, the retention of identity across an interval, which is deeper than any disposition and which does not, by its own necessity, necessitate what it grounds (§2.8.4; L9, §4.4.9). The conclusion the inertialist wants is secured (Nature persists without a divine sustainer) on a ground the inertialist cannot supply. The local result is that Φ states the constitutive condition under which persistence counts as persistence and that its necessity does not necessitate any particular persister. Whether a further sustaining explanation is required remains a separate grounding dispute; this section blocks the inference from mere continuation to continuous re-creation but does not settle every conservationist formulation by definition alone.

4.3 Physics and Mind

This cluster is different. These objections do not try to escape the structure; they grant it and ask whether it is rich enough. Can mere structure accommodate conscious experience, intrinsic nature beneath relations, or the uncanny fit between mathematics and the world?

Because these challenges come from inside the picture, they cannot be met by the earlier reply that a rival merely relocates structure. The diagnostic shifts. Where §§4.1–4.2 test accounts that claim to avoid the structure, this cluster tests whether the account can answer genuine questions from within.

4.3.1 The Hard Problem of Consciousness

Chalmers (1996) named it the hard problem and gave it its contemporary form. The easy problems (attention, memory, integration, report, discrimination) submit to mechanistic explanation by showing which brain processes perform which cognitive work. The hard problem is different in kind. Even granting a complete mechanistic account of every neural process, the question remains: why is any of this accompanied by experience at all? Jackson (1982) sharpened the point with Mary the color scientist. Nagel (1974) sharpened it with bats. Levine (1983) gave the residue a technical name: the explanatory gap. Kripke (1980, Lecture III) had earlier pressed the modal version: if mind-brain identification is necessary and zombies are conceivable, identification fails. Chalmers’s The Conscious Mind (1996) and The Character of Consciousness (2010) consolidated these lines into the contemporary case, with Block (1995), Strawson (1994), and Goff (2017) extending the position in different directions. Against them stand two deflationary traditions: the higher-order representational accounts (Rosenthal 2005; Carruthers 2000), on which qualitative character reduces to structured self-monitoring; and the eliminativist/illusionist wing (Dennett 1991; Frankish 2016), on which the qualitative residue is itself a misleading artifact of introspection. The present reply sits closer to illusionism than to higher-order theory, but the reason is specific: what illusionism calls a false seeming, the structural account describes positively, as what self-modeling consists in when encountered from inside.

The hard problem bears on the argument’s most ambitious claim. The later structural and identity arguments of §§2.4–2.8 hold that T, S, and Φ exhaust the structural roles required by change-realizing reality. Section 2.1 supplies the modal framework and Necessary Possibility theorem; it does not by itself establish structural exhaustion. If experience is genuinely non-structural (if qualia are irreducible qualitative features that no structural description captures) the structural exhaustion claim fails. The Maximally Great Being is not thereby required, but a non-structural ingredient is, and that ingredient opens exactly the space the classical tradition has always wanted to repopulate. The section’s defensive claim below (that any qualitative ground must still instantiate T, S, and Φ) is argued in full here; its more ambitious positive claim — that phenomenal interiority just is structural self-modeling encountered from within — is only asserted here and is earned later, at §7.8 and §8.3. A reader who rejects that later identification should treat this section’s closure as provisional on it.

A clarification before the moves. The question is not whether the concept of experience differs from the concept of any particular neural process: it obviously does, just as the concept of water differed from the concept of H₂O for most of human history. The question is whether the conceptual difference tracks a genuine ontological division, whether the qualitative adds a fact to the world beyond what any structural account could specify. Three forms of the move press that ontological claim at three resolutions (the explanatory gap, the conceivability argument, and the knowledge argument), and the space they cover is the space in which the hard problem could succeed. Each fails for a different reason.

First. The explanatory-gap move holds that every structural description of the brain leaves the qualitative character of experience unexplained. Full mechanistic accounts of information processing, attention, memory, and report can be given, and at the end of the account the question remains why any of this should be accompanied by experience at all. The gap is not merely epistemic; it tracks a genuine ontological divide.

Explanatory gaps are facts about concepts, not automatically facts about entities. The gap between thermodynamics and mechanics was real before statistical mechanics closed it. The gap between chemistry and physics was real before quantum mechanics closed it. Each gap seemed to require an extra ingredient; each was eventually closed by deepening the structural account rather than by appealing to something outside structure. An incomplete vocabulary is evidence that the structural account needs deepening, not that the ontology lies elsewhere. The gap as Chalmers describes it cannot by itself defeat the structural argument, because whatever the ground of experience is, it must instantiate T, S, and Φ. A qualitative ground lacking causally structured succession cannot produce an experience with temporal character. One lacking relational extension cannot produce spatial character. One lacking causal persistence cannot connect one moment to the next as a unified stream. The defensive claim is therefore secure: even if irreducible qualitative features exist, they are features of a structured medium rather than a way around the structural requirements. The more ambitious claim (that phenomenal interiority is identical to structural self-modeling encountered from within) requires a positive account of self-modeling, not merely the failure of dualism. The gap is evidence of structure’s depth, or of a further feature structure must accommodate; it is not evidence of a structureless domain.

Second. The conceivability argument goes further. A philosophical zombie can be conceived: a being physically and functionally identical to a conscious human but lacking any inner experience. If zombies are conceivable, they are possible. If possible, consciousness is not identical to any structural or functional fact. The difference between a conscious being and a zombie is precisely what structural description leaves out.

Zombie conceivability does not automatically generate metaphysical possibility. Kripke’s (1980) framework requires that what is conceived be genuinely possible, not merely pictured. A zombie is pictured as a functional duplicate lacking experience. But when one specifies what it is for the zombie to be a functional duplicate (to process information, to integrate across senses, to produce reports, to model its environment, to adapt its behavior to changing conditions) one has at least specified the complete structural profile of self-modeling. The anti-physicalist insists that the phenomenal fact could still be absent. The structural reply should not pretend that this insistence has been refuted by definition. Its narrower point is that the alleged absence, if coherent, would show that self-modeling has an interior feature not yet captured by the external description; it would not show that experience floats free of T, S, and Φ. The zombie names an imagined separation between structural profile and phenomenal interiority. Whether the separation is ultimately coherent is the hard-problem dispute; what it cannot supply is a ground outside structure.

The deeper move presses the inference itself, and here Chalmers does not rest on a naive step. He grants, with Kripke, that conceivability generally underwrites epistemic and not metaphysical possibility: watery stuff without H₂O is conceivable yet metaphysically impossible, because a hidden essence divides how the kind is picked out from what it is. His claim is that consciousness is the lone exception: a phenomenal concept has no essence hidden behind its mode of presentation, the appearance being the reality, so for experience conceivability and metaphysical possibility cannot come apart, and the conceived zombie reaches genuine possibility.

The disanalogy fails, and it fails on what a phenomenal concept is. A phenomenal concept is not a window onto a non-structural quality but a reconstruction the system runs on its own states: the stored, re-deployable modeling of a configuration it has occupied, the account §8.3 gives of what Mary acquires. So understood, it is bounded as every concept is: it reaches only what the conceiver’s modeling has mapped, never the modal structure entire. The wedge Chalmers closes between presentation and physical essence reopens at another seam: between the field a conceiver has mapped and the structure that field only partially maps. Conceiving a zombie is failing to locate, within one’s own field, the link between structural profile and interiority; and a gap in the map is epistemic possibility (what the field has not ruled out), not metaphysical possibility, which is a permission of the structure itself. The zombie is epistemically possible and modally idle. This does not convict the argument of formal fallacy; it turns on identifying phenomenal concepts with field-internal reconstructions, a claim §7.8 and §8.3 earn rather than one established here. But it sharpens the reply beyond relocation: the inference from conceiving to possibility is only as strong as the assumption that phenomenal concepts outrun the conceiver’s field to reach the structure itself (the very claim a structural account of those concepts rejects), so it cannot independently establish what it needs.

The reply can be tightened into a dilemma the conceivability argument does not pass through cleanly. Either the zombie’s conceivability is independent of consciousness’s non-structural status, or it is not. Taken as independent, a datum any inquirer could read off without first settling the metaphysics, the step from conceiving to metaphysical possibility is just the move Kripke discredited (watery stuff without H₂O is conceivable and impossible), and treating the conceived zombie as evidence of the conclusion affirms the consequent. Taken as not independent, the zombie rates as ideally conceivable only because one already treats interiority as something no structural fact constitutes, so the premise has assumed the non-identity it was meant to establish. Chalmers escapes the first horn only by the disanalogy, the claim that consciousness alone carries no essence behind its presentation; but that is the second horn, since the disanalogy is precisely what installs the prior treatment of interiority as non-structural. The one exit is the structure-and-dynamics claim pressed at §8.3 §A2, that physical description yields only structure while experience is neither, which would make the zombie ideally conceivable on independent grounds. The present account denies exactly that, holding interiority to be the inside-aspect of structural self-modeling, so the exit is closed and the inference is left either affirming its consequent or presupposing its conclusion. This does not upgrade the earlier verdict to a charge of bare formal fallacy: the point is conditional, that absent the structure-and-dynamics support the inference has no independent legs, and whether that support stands is the dispute §8.3 carries.

The disanalogy rests, at bottom, on a blank. That a phenomenal concept has no hidden essence is what lets the zombie look ideally conceivable, for if nothing about what experience is obstructs the subtraction, the structure can be kept while the interior is removed. But “no hidden essence” specifies nothing; it is the refusal to say what a conscious state is, recast as an insight, and the modal argument is built on that silence. Filled, the silence closes the argument. To be an experience at all is to be undergone from a standpoint, and a standpoint is not structureless: it is somewhere and not nowhere (relational extension, S), it observes across a now that has a before and an after (succession, T), and it is the standpoint of something that persists as the same standpoint through what it undergoes (a persisting bearer, Φ). These are not accompaniments of phenomenal character but what having it consists in, hence conditions on there being any interior to hold fixed while the physics is subtracted. Consciousness therefore presupposes T, S, and Φ rather than conferring them, and what a thing presupposes is prior to it; a candidate fundamental cannot draw its own character from conditions more basic than itself. So consciousness cannot be the structureless bedrock the conceivability argument needs it to be, and the hidden essence Chalmers declares absent is just this: that T, S, and Φ are essential to consciousness, the standpoint’s own constitution standing behind its felt presentation as a chemical structure stands behind water’s. Consciousness is on this count relevantly like water rather than the exception the disanalogy requires: it has a hidden essence, so subtracting its interior while the structure remains is conceivable only in the way watery stuff without its microstructure is conceivable, and reaches epistemic possibility alone.

The disanalogy backfires at a second turn. Set aside whether water’s essence is exactly H₂O; whatever its microstructure, water is water only as something that occupies a place (S), lasts through time (T), and is made of persisting substance (Φ). What is essential to water, at the level on which the modal argument turns, is the same triad essential to consciousness. The kind Chalmers offers as the contrast, a thing with a hidden essence set against a supposedly essence-free quale, shares consciousness’s essence at the floor: both are instances of spatial, temporal, substantial constitution, differing in the vocabulary their presentation wears and not in whether they answer to T, S, and Φ. Read as transparency, then, “no hidden essence” does not lift experience clear of structure but hands it over as structure, and the zombie, a complete structural standpoint with its interior removed, describes one thing as at once having and lacking what it is. This is the explanatory-order point, and the one skeleton beneath the mental and the physical alike, that §4.3.5 presses against idealism, registered here at the root from the side of the phenomenal concept. Fundamentality fixes an order of explanation and leaves untouched what the fundamental must be like to play its part, so naming consciousness fundamental cannot exempt it from the conditions any standpoint meets in order to be one.

Even granting the inference for the sake of argument, the zombie cannot escape a cost. Neither cost that follows is levied from outside; both are the shape of the construction, since a complete physical duplicate whose behavior is identical already fixes that the phenomenal makes no physical difference, and holding the physics fixed while the interior goes missing already locates the difference in a law that could have been otherwise. Because it is a physical duplicate, it withdraws its hand from the stove exactly as we do, reports that it withdrew because of the pain, and avoids the burner thereafter: all driven by the nociceptive-motor arc, with the feeling, by hypothesis, doing none of the work. The zombie-defender must therefore hold that in us too the feeling is not what moves the hand; the arc is, and the hurt is a passenger that could vanish with no behavioral trace. The same follows for the report: the duplicate sincerely says “it hurt,” for purely structural reasons, so our own conviction that the hurt caused the withdrawal is generated by the arc rather than by the hurt, and would be produced identically by a being with no hurt at all. The position is coherent, but its price is steep: a thoroughgoing epiphenomenalism on which the felt character of experience explains neither our behavior nor our judgments about that behavior, including the judgment that there is a hard problem at all. The structural identification does not pay this price: if the feeling is the arc encountered from inside, then “withdrew because it hurt” and “withdrew because of the nociceptive arc” are one explanation under two descriptions, and the felt character does causal work as the structure it is. This does not refute the conceivability of the zombie; it relocates the burden. The separability the zombie pictures is purchasable only by accepting that feeling is causally idle, and the idleness is harder to credit than the separability is to imagine. And the idleness does not stay contained on the side of behavior. The conceivability intuition itself — the judgment that the interior could be subtracted while the structure remains — is a phenomenal judgment, produced in us by the very arc that produces it in the duplicate; on the defender’s own commitments it would be made, word for word, by a being in whom there is nothing to track. Chalmers names this tension himself and gives it its most sustained treatment as the paradox of phenomenal judgment: a zombie’s utterances about its own experience would be caused, and would strike it as justified, in exactly the way ours are (Chalmers 1996, ch. 5). An intuition certain to occur whether or not its object is present cannot be evidence that the object is there, so the argument’s motivating datum is undercut by its own conclusion. To keep the zombie the defender must treat the sense that consciousness outruns structure as a report of phenomenal fact while conceding the report would be identical were there no such fact. The identification is not so placed — if the feeling is the arc from within, the judgment “it hurt” is caused by the feeling it reports, and phenomenal judgments are about the phenomenal because they are the phenomenal registering itself. The defender may answer, as Chalmers himself does, that consciousness is known by direct acquaintance rather than by such tracking (1996, ch. 5); but an acquaintance whose deliverances are, by the same closure, produced independently of what they acquaint is one more primitive posited to keep the first, and the ledger of primitives only lengthens.

A further pressure follows on the modal side. Hold the physical facts fixed across worlds and ask what would have to differ for one of them to be a zombie world: not a physical law, by stipulation, but the law linking the physical to the phenomenal. The zombie-defender must therefore hold that the psychophysical bridging laws are separable from the physical laws: that the same physical structure could be governed by different bridging principles, or none. That is a real commitment, and it is not free: it requires that the phenomenal be a further fundamental feature, additional to the structural facts and linked to them only contingently, in the manner of Chalmers’s naturalistic dualism. The structural identification denies exactly this. If phenomenal interiority is not a further feature bridged to structure but the inside-aspect of the structure itself, then there are no separable psychophysical laws for a zombie world to vary. The link is not a contingent law laid over the physics but a description of what the physics already is from within, and it inherits whatever necessity the structure has. On that reading the zombie world is not ‘the same physics governed by a different bridging law’; it is the same structure described as both having and lacking its own interior, which is no coherent world at all. The argument does not establish this by itself: it turns on whether interiority is a further feature or an inside-aspect, the dispute the rest of the chapter and §6 carry. What it shows is the shape of the cost: the separability the zombie requires is purchasable only by treating consciousness as an additional fundamental ingredient, and once that ingredient is refused on the independent grounds the structural account gives, the bridging law it would vary is no longer there to be varied. The two costs compound rather than sit side by side. The bridging law is contingent by hypothesis and, by the epiphenomenal result just reached, links a feature that does no physical work. So the defender posits two fundamentals at once — an extra property and an extra law — neither of which earns its place by explaining anything, both retained only to honor a modal intuition the preceding paragraph has left without evidential standing. That the law may be nomologically necessary is no relief: the necessity is one more brute posit, not a discount on the two already made. Weighed as the best-fit judgment this book runs everywhere else, the verdict is lopsided — more fundamental posits, less explained, on a shakier base — and the identification carries none of the freight: no separable law, no added property, and phenomenal judgment reconnected to its object. It prevails not by proving the zombie impossible but by being, on the ordinary abductive standard, the markedly cheaper and more explanatory theory.

Third. Jackson’s knowledge argument presses from a different direction. Mary, raised in a black-and-white room, knows every physical fact about color vision. When she leaves the room and sees red for the first time, she learns something new: what it is like to see red. If complete physical knowledge left something out, physical description is not exhaustive. The something left out is the qualitative character of experience.

Mary learns something new because she enters a new causal relation, not because she acquires access to a non-structural domain. Prior to leaving the room, Mary has structural descriptions of what red-seeing is; she does not have the structural state of red-seeing. The transition from description to instantiation is a real cognitive event: a new configuration of her self-modeling apparatus that includes the actual state rather than merely its description. What Mary learns is a structural fact: what it is for her to instantiate the specific configuration that red-seeing consists in. Description and instantiation are always distinct relations to the same fact. No description collapses into instantiation without ceasing to be a description. Mary’s transition confirms this, and the transition is fully within the structural account.

Jackson himself later withdrew the knowledge argument. In “Mind and Illusion” (2003), he concluded that a complete physical description would after all fix all the qualitative facts, and that the original argument’s appeal to what Mary learns upon leaving the room trades on a representational confusion rather than tracking a non-physical domain. The retraction travels less widely than the original thought experiment, and readers weighing the argument are best served by considering both.

A kindred deflation drops the hard-problem apparatus for higher-order theories of consciousness. Rosenthal (2005) argues a mental state is conscious when accompanied by a higher-order representation of it, and qualitative character reduces to the structure of self-monitoring. The reply is the same in thinner vocabulary. Higher-order representation requires relational distinction between the representing state and the represented state (S), temporal succession through which the representing obtains (T), and a persistent subject across which the relation is maintained (Φ). Higher-order theory has not explained qualia without structure; it has described qualia as a form of structural self-relation. Dennett’s heterophenomenological method (Consciousness Explained, 1991) goes further (treating ‘from inside’ language as a pattern of intentional behavior rather than a report from a Cartesian theater), but third-person analysis of the intentional stance itself presupposes T, S, and Φ as conditions on any behavioral pattern, including the pattern of seeming to have qualia.

The Russellian alternative, that structure has an intrinsic phenomenal inside, is §4.3.3. Wigner’s applicability challenge, which presses a different mode of non-structural residue through mathematics rather than consciousness, is §4.3.6.

What converges. The three forms do not establish a way around structure. The explanatory-gap move shows that the external structural vocabulary may be incomplete from the first-person side; it does not show that the first-person side lacks temporal character, differentiation, or persistence. The zombie argument presses the imagined separation between structural profile and phenomenal interiority as far as it will go; even if that separation remains dialectically live, it is a separation within a structured medium, not an escape from one. Mary’s transition is the move from description to instantiation: a real cognitive event within the structural account, not access to a non-structural domain; Jackson himself later withdrew the argument on related grounds.

The positive thesis can now be stated without overclaiming. Experience is best understood, on the present account, as structural self-modeling encountered from inside rather than described from outside. That claim is not secured merely by defeating zombies; it is the constructive account §7 must earn. The conceivability step itself (whether picturing a zombie reaches a genuine possibility or only a gap in the conceiver’s own modal field) is pressed further at §8.3 §A2, where the doubly-field-internal and structural-residue diagnostics carry in full the engagement the present section runs in compact form. Here the section both presses the inference, marking the conceived zombie as a gap in the conceiver’s field rather than a permission of the structure, and shows that even a zombie granted as possible purchases its separation only at the price of epiphenomenalism and supplies no ground outside structure. The consistency check on the other side is supplied at §4.3.2: the deflationary engagement with eliminativism and illusionism there operates at the level of conditions on representation, not at the level of what is represented, and so does not undercut the present identification of interiority with the inside-aspect of structural self-modeling. The result needed for Chapter 4 is narrower and is fully secured: the hard problem does not re-open the classical-theist space or require a structureless ingredient. Descartes formed the cogito (Latin for “I think”, the seed of cogito ergo sum) inside the very self-reflection whose existence was in doubt; what he reached was the certainty that something was doing the doubting. Chalmers opened The Conscious Mind with the observation that consciousness is the one fact about the world of which we are most certain and the one we understand least. Both halves of that remark are true. The certainty is the instantiation; the not-understanding is the gap between instantiation and description. The gap, on this account, names the depth the structure has when the structure is what is doing the looking, rather than an ontological elsewhere. Chalmers’s most sustained response to structural and functional accounts (the organizational invariance principle of The Conscious Mind §3) accepts that functional organization is necessary for phenomenal experience but presses that the same organizational configuration could in principle produce no experience or a different experience, leaving the bridging relationship unexplained. The structural reply is that organizational invariance names a depth constraint rather than a separate domain: explaining why this configuration produces this qualitative character rather than another is the work of the inside-view account of structural self-modeling that §7.8 develops as closure-plus-throughput. Chalmers’s challenge does not establish a non-structural residue; it identifies the specific structural relationship whose inside-view character remains to be specified, and that specification is Chapter 7’s task, not a concession to dualism.

The sensation has interiority, and on the present account, that interiority is what the architecture looks like from within itself, not a further layer behind it.

4.3.2 Eliminativism and Illusionism

Eliminative materialism, defended by Patricia and Paul Churchland (1981, 1989), centrally targets the propositional-attitude categories of folk psychology, especially belief and desire, as theoretical posits an adequate neuroscience may replace. It should not be treated as automatically eliminating qualia or the conscious subject in every version; those stronger denials belong to distinct eliminativist or illusionist positions. The propositional attitudes, on the Churchlands’ line, are theoretical posits of an inadequate framework, to be replaced by mature neuroscience as phlogiston was replaced by oxidation. Frankish’s (2016) illusionism, anticipated by Dennett (“Quining Qualia,” 1988; Consciousness Explained, 1991), presses a related line: phenomenal consciousness is an illusion. There seems to be a what-it-is-like-ness to experience (Nagel’s 1974 term for the felt qualitative character of having an experience), but the seeming is itself a representational state that does not require any phenomenal character to be what it represents.

Both positions are aimed at the hard problem (§4.3.1). They dissolve it by denying that what the hard problem is about exists.

The argument does not turn on the existence of phenomenal consciousness. It turns on the conditions for any process, including the process of formulating eliminativism, processing neural states, generating illusory representations of consciousness. Eliminativism is a position. Positions have content. Content is distinguishable from other content. The distinction obtains across moments, across substrates, across thinkers. T, S, and Φ are the conditions on which any content (including the content of eliminativism) is what it is.

A targeted observation. Eliminativism does not in fact succeed in eliminating consciousness; it eliminates the folk-psychological vocabulary for consciousness in favor of a neuroscientific vocabulary. The neuroscientific vocabulary still distinguishes one neural state from another, one organism’s processing from another’s, one moment’s representation from a later moment’s. What was phenomenal consciousness becomes neural information processing. The vocabulary changes; the conditions do not.

Illusionism is in a slightly different position. The illusionist concedes that something seems to be the case and contends that the seeming is mistaken. But the seeming itself is a state: a representation, a process, an event. Whatever its ultimate nature, it occurs in time, distinguishes itself from the world it represents, and persists long enough to be analyzed. The structural conditions are not features of the represented; they are features of the representing. Illusionism reduces what is represented; it does not reduce the conditions of representation.

A note on consistency with §4.3.1. The deflationary stance just stated may read in tension with the positive thesis of §4.3.1 that experience is structural self-modeling encountered from inside. The two sections operate at different argumentative levels. Interiority is identified at §4.3.1 as what self-modeling is when encountered from within: a structural-identification claim about the explanandum eliminativism and illusionism deny. This section engages those positions at the level of the conditions on any representational activity, including their own: any content, including the eliminativist’s denial and the illusionist’s diagnosis, instantiates T, S, and Φ. The defensive claim that structural conditions are conditions on representing rather than on what is represented does not commit the framework to denying that there is anything to represent; what is represented in the case at issue is the inside-aspect of structural self-modeling that §4.3.1 identifies positively and Chapter 8 develops. The two sections divide labor: the earlier one names what is being denied and identifies it positively; this section marks the conditions any denial of it must instantiate.

The meta-problem move (Chalmers, “The Meta-Problem of Consciousness,” 2018; cf. Frankish, Illusionism as a Theory of Consciousness, 2016) presses a methodological reframe: instead of asking why phenomenal consciousness exists, explain why reflective people say the hard problem is hard. The meta-problem strategy is that explaining the impression of a hard problem (why introspective mechanisms represent cognitive processes as having irreducible phenomenal character they do not possess) dissolves the first-order question without answering it on the hard problem’s own terms. The structural reply: the meta-problem move is itself a structural exercise. Explaining why people report a hard problem requires an account of how introspective mechanisms represent their own outputs, how that representation diverges from the process it represents, and how the sense of seeming arises as a feature of sophisticated self-monitoring. That account requires T (the temporal succession of the introspective act and the report it generates), S (the relational distinction between the introspecting state and the state it represents), and Φ (the persistence of the self-modeling system across the introspective act).

What is established. The hard-problem dispute is, on the present argument, downstream of the structural conditions. Whether phenomenal consciousness is real, illusory, or eliminable is a question for philosophy of mind. T, S, and Φ are required either way.

4.3.3 Russellian Monism

Physics gives increasingly precise causal and mathematical profiles of physical entities — an electron’s charge, spin, rest mass, and interaction amplitudes — within the domains covered by our theories and experiments. It does not yet provide an exhaustive profile in every possible context. It does not specify what the entity is. There could be something it is like to be an electron, or something it is not like to be one; physics has nothing to say either way, and nothing about physics suggests it ever will. The observation is not mysticism but a remark about what physics, by its own methodological lights, sets out to answer.

Russell (The Analysis of Matter, 1927) gave the observation its modern form. Eddington (The Nature of the Physical World, 1928) pressed it in the same decade under the heading the physical world is a shadow world. Feigl (1958) developed a ‘two-attribute’ version before the label ‘Russellian Monism’ existed. Chalmers (1996, 2015) revived the position as a genuine competitor to dualism and eliminativism. Goff (Consciousness and Fundamental Reality, 2017) developed constitutive panpsychism as the most systematic contemporary form. Strawson (“Realistic Monism,” 2006) pressed the panpsychist inflection harder: ‘real physicalism,’ on his argument, requires panpsychism, because structural description cannot on its own deliver concreta. Alter and Nagasawa (2015) collected the contemporary debate; Stoljar (Ignorance and Imagination, 2006) developed a neutral monist variant; Pereboom (Consciousness and the Prospects of Physicalism, 2011) pressed a proto-phenomenalist version that tries to bracket the panpsychist commitment. Against the Russellian stand the structural realists (Ladyman and Ross Every Thing Must Go, 2007; French The Structure of the World, 2014), who press the opposite conclusion: not that structure requires intrinsic natures, but that the demand for intrinsic natures is confused, and structure is all there is. The present reply sits with the structural realists on the first point (the demand for quiddities presupposes a separability the argument specifically forecloses) and adds what the bare structural realist position does not specify: which structural roles, and why they are necessary.

The Russellian pressure is precise. The argument that runs from the Cogito-S5 Pipeline of §2.1 through the structural and identity arguments of §§2.4–2.8 holds that T, S, and Φ (the three structural conditions, borne by Time, Space, and Substance) exhaust the structural roles necessary existence requires. Russell grants structure, grants the necessity of the roles, grants that whatever exists instantiates them, and asks what the relational structure is made of. The quiddities (the intrinsic natures filling the structural positions) are the residue the argument leaves out. If the question is well-formed, the structural exhaustion claim is incomplete; the structural roles do not exhaust what a complete ontology must specify, and the space for a non-structural ground reopens exactly where the argument thought it had closed.

A clarification before the moves. The Russellian is not a dualist. Intrinsic natures and structural roles are not two separate kinds of fact co-existing in one world. They are, on the view, one world described twice: once from outside (the structural description physics gives) and once from inside (the intrinsic natures, which include or consist in phenomenal or proto-phenomenal properties). What turns on the view is whether ‘what structure does’ and ‘what structure is’ are genuinely separable specifications of one fact, or whether specifying what structure does already fixes what structure is. Three forms of the move press that separability at three resolutions: the bare structural-residue move, Strawson’s panpsychist version, and Goff’s constitutive version. The space they cover is the space the Russellian position occupies. Each collapses a separation the argument forecloses.

First. The structural-residue move holds that physics gives us the causal-structural profile of reality but tells us only what entities do, not what they are. The intrinsic nature of the relata is underdetermined by the structural description, and a complete ontology requires both: structure and the intrinsic natures that fill it.

The separability assumption fails at the level of structural role definitions. Time is not merely a description of a relation between states; it is the specification of what causally structured succession consists in, including what it is for a state to be a state and what it is for succession to be causally directed. Space is not merely a description of relations between positions; it is the specification of what relational non-identity consists in, including what it is for distinct positions to be distinct. Substance is not merely a description of persistence; it is the specification of what it is for causal capacity to be maintained across transitions. The present account intends these as constitutive specifications of relata and relations together. But the Russellian challenge is precisely that a complete causal-role profile may leave its categorical or intrinsic realizer underdetermined. Repeating the role definition does not settle that dispute. The account defends role-exhaustion as the book’s position while acknowledging that excluding quiddities requires an additional argument against multiple realization of one complete structural profile. The question ‘what plays the role’ presupposes that role and player are separable items: a separability the constitutive reading denies.

Second. Strawson’s panpsychist version pushes harder. The intrinsic natures cannot be nothing: a structural skeleton with no filling is no structure at all. Something must play the role of what-the-relata-are. Either that something is itself structural (in which case the regress continues) or it is non-structural. Panpsychism names the non-structural option: the intrinsic natures are phenomenal or proto-phenomenal. Qualia ground the structure because something must, and only qualia are non-structural in the right way.

The regress Strawson generates dissolves when the starting assumption is dropped. The demand for a non-structural ground rests on the picture of structure as a skeleton requiring external filling: a picture the argument rejects at its foundation. A structural role is not a formal shell to which content must be added. It is a constitutive specification of what its relata are. If quiddities are genuinely distinct, they support at least a thin difference-role. It does not follow from non-identity alone that they occupy a spatial manifold or instantiate Spatiality’s loaded type. The stronger conclusion requires the independent role-to-type bridge rather than a definition of difference as Space. The panpsychist move, on this diagnosis, describes the structure at a different altitude rather than escaping it.

Third. Goff’s constitutive panpsychism develops the position furthest. Macro-consciousness is constituted by the phenomenal properties of micro-level entities. The combination problem (how micro-phenomenal properties combine into unified macro-experiences) is real but tractable if the micro-properties are the right kind. The qualitative character of our experience is not additional to the physical substrate; it is what the physical substrate is, in itself, at the micro level. The present account describes the outside; panpsychism describes the inside.

Constitutive panpsychism requires combination across micro-entities, and combination is itself a structural fact. For micro-phenomenal properties to combine into macro-experience, they must stand in relations of integration, temporal persistence across the combining process, and differentiation from alternatives. These are T, S, and Φ operating at the micro level. The combination problem is not merely a technical obstacle Goff’s view must solve; it is the point at which the view’s structural commitments become visible. Whatever relates the micro-phenomenal properties to each other (whatever integrates them, sustains their unity, and distinguishes the combined state from its constituents) is a structural relation, and the micro-phenomenal properties themselves stand in structural relations at their own level. The panpsychist inside, on this reading, is the structure encountered at a smaller grain rather than something beyond it.

Name the motivation directly. Goff’s case for placing experience at the micro level is, at bottom, an argument against strong emergence: to derive an experiential interior from a wholly non-experiential exterior would be a brute, unintelligible leap, and constitutive panpsychism is built to avoid it by making the interior fundamental rather than emergent. The present account owes a verdict on that dilemma, because ‘structure has an inside’ can sound like the very derivation Goff forbids: an experiential inside conjured from a non-experiential outside. It is not, and the reason is the one already in hand. Goff’s dilemma (strong emergence or fundamental micro-experience) assumes that the structural description is, in itself, non-experiential: that ‘outside’ and ‘inside’ name two stocks of fact, one of which must be either reduced to or brutely produced from the other. That assumption is the separability the argument forecloses. The inside is not derived from the outside; it is the same structural fact encountered from within (§4.3.1). Where there is no non-experiential base standing apart from an experiential product, there is nothing for emergence (strong or weak) to bridge. The role-not-quiddity move does not take a side in Goff’s dilemma; it dissolves the dilemma, by denying the non-experiential base from which the objectionable derivation was supposed to begin.

A kindred move drops panpsychism for neutral monism: the intrinsic base is neither phenomenal nor physical but a neutral third thing of which both are aspects (Stoljar 2006; earlier, James 1904). The reply follows the same path. Whatever neutral stuff fills the structural positions must stand in relations of distinction, succession, and persistence at its own level. Neutral monism relocates the structural roles beneath the mental/physical distinction rather than eliminating them.

A more radical position in the neighborhood is eliminative ontic structural realism. Moderate OSR, the Ladyman and Ross view assumed above, and the position French systematizes in The Structure of the World (2014), holds that structure is ontologically primary and that objects are thin nodes defined by the roles they play. This is a natural ally of the argument: objects are specified by structural roles, not by hidden intrinsic natures. Eliminative OSR presses further. There are no objects at all. Only relations exist; the apparent ontology of relata is a reification structure-talk encourages and strict metaphysics must abandon. Parts of Ladyman and Ross (2007) are read this way in the rhetorical register; French’s own formulation is more cautious. If eliminative OSR is right, Substance’s role as causal capacity maintained across transitions has nothing to attach to. There is no persisting anything: only a pattern whose relata were never there.

The eliminative reading is unstable in a way moderate OSR is not. Relations without relata are not relations in any serviceable sense; they are the grammatical form of relations whose relata the metaphysics has declined to specify. Whatever the eliminativist calls the pattern must have locations where the pattern takes one form rather than another, and loci of differentiation just are what Substance names at minimal grain. The argument does not require objects to be thick particulars with intrinsic natures (moderate OSR already agrees they are not). It requires what any coherent structural description requires: that the structural positions be distinct from one another (Space), that they persist through transitions (Substance), and that the transitions succeed in a causally directed order (Time). Eliminative OSR avoids this conclusion only by dropping enough structural vocabulary that nothing remains to describe a world with, at which point the position has become a refusal of ontology rather than an alternative to it. Moderate OSR, which keeps the vocabulary, coincides with the present position at a different altitude.

The hard problem that motivates the Russellian response (that qualia seem irreducible to structural description) is §4.3.1. The applicability challenge, which presses a non-structural residue through mathematics rather than consciousness, is §4.3.6.

What converges. Each of the three forms describes the structure at a different altitude rather than escaping it. The structural-residue move presupposes a separability the role specifications foreclose: role and player are not separable items, because the roles constitutively specify what relata and relations are together. Strawson’s panpsychism requires quiddities to stand in relations of non-identity among themselves: a domain that already instantiates the thin difference-role at its own level, relocating the structure rather than escaping it. Goff’s combination problem is where the view’s structural commitments become visible: whatever integrates, sustains, and differentiates micro-phenomenal properties is T, S, and Φ operating at smaller grain.

State the position flatly. The book is not Russellian: it denies what every Russellian affirms, that the structural roles leave a residue, an intrinsic nature distinct from the role, still to be specified. Nor is it anti-Russellian in the eliminative structural realist’s key, which dismisses the ‘what it is’ question as confused. It is the third option. The Russellian question is well-formed and its datum (that physics gives what things do and not what they are) is real; what fails is only the inference from that datum to a non-structural quiddity. The book keeps the Russellian insight and discards the Russellian ontology.

The Russellian’s deepest insight is real: structure has an inside. The form of that inside is established at §4.3.1 at the level of cognition: experience is what self-modeling consists in encountered from within. What §4.3.3 extends is the diagnosis downward. The inside does not precede structure. It is what the structure consists in at whatever grain the question is asked. The present account denies that an electron must have a fixed quiddity beyond its complete structural profile; current physics does not by itself decide that metaphysical question. What it has is a structural role and whatever role-playing consists in, described at the limit of what description reaches. Positing an intrinsic nature behind the roles requires cognitive access, which requires the roles. The quiddity is accessed through the architecture it claims to underlie.

Russell’s challenge is among the deepest the literature has produced. What it reveals, on the present account, is the depth within structure rather than a layer beneath it.

4.3.4 Cosmopsychism

Goff (2017, 2019) and Shani (2015) develop cosmopsychism as the inverse of constitutive panpsychism (the view that consciousness is fundamental at the micro-level and macroconsciousness arises by combination of microconscious parts): the cosmos is fundamentally a single conscious subject, and finite minds are derivative parts or aspects of that cosmic consciousness. The view sits closer to Spinoza and to Advaita Vedānta (the Indian non-dualist tradition holding that ultimate reality is one undifferentiated consciousness) than to standard Western theism, but it shares the central commitment that a single fundamental subject grounds the structural roles (T, S, and Φ, the three structural Conditions the book’s argument derives, borne by Time, Space, and Substance). Goff’s “priority cosmopsychism” attaches the view to Schaffer’s (2010) priority monism: the cosmos is the one fundamental concrete object, and finite things (including finite minds) are derivative.

The position is formulated to escape the combination problem that troubles constitutive panpsychism. Where constitutive panpsychism builds macroconsciousness from microconscious parts and has to explain how subject-summing works, cosmopsychism starts with the unified cosmic subject and faces the inverse task (decombination) of explaining how finite minds derive from a cosmic mind. Chalmers (2016) and Miller (2017) have pressed the decombination problem as no easier than its mirror: a unified subject does not obviously yield discrete subjects with their own perspectival closures. Goff’s reply invokes “subsumption”: finite subjects are aspects of the cosmic subject in the way a region of a visual field is an aspect of the field. The proposal is genuinely under development, and its prospects are open.

The challenge to the present argument has two layers. The first is the priority claim: cosmopsychism asserts that the cosmic subject is metaphysically prior to T, S, and Φ, that the structural conditions are features of the subject’s mode of existence rather than independent. The second is the unity claim: that consciousness, properly understood, is fundamentally one and that the appearance of structural plurality is derivative.

The reply addresses both layers.

On the priority claim: a cosmic subject capable of grounding finite minds must stand in a determinate relation to them. That relation requires difference, directed dependence, and real bearing at role-resolution. Cosmopsychists need not grant that decomposition literally runs over time; upgrading the directed grounding relation to temporal articulation requires the separate role-to-type argument. Even granted in full, the unified phenomenal field reading (Shani’s cosmopsychism, perennialist non-dual consciousness, Advaita’s cit as undifferentiated awareness), the priority reversal goes through not at the mode of cosmic cognition (which may indeed be a non-discursive unified field, engaged as a configuration at §2.6.8 and as a subject-type at §7.8) but at the relation the cosmic subject must stand in to what it grounds. If the cosmic subject lacked T, the cosmic-to-finite decomposition could not be temporally articulated; if it lacked S, the finite minds grounded would not be differentiated; if it lacked Φ, no grounding relation could persist long enough to ground anything. The structural conditions are not features of the cosmic subject’s mode of existence; they are conditions on there being a cosmic subject of the kind cosmopsychism requires. The priority is the wrong way around.

A prior challenge the reply can afford to leave open is whether a composite ‘cosmic subject’ exists to bear the priority at all. Mereological nihilism denies that any such composite is there to be fundamental (the pressure §4.3.11 takes up directly), and the structural reply is consistent with either verdict, since the structural conditions are prior to the cosmic subject whether or not the composition obtains.

On the unity claim: the structural argument does not deny that there is a single fundamental ground. It denies that the ground is a subject. A subject (whether finite or cosmic) is a configuration of self-modeling under structural constraints (§7.8 and Chapter 8). The structural conditions are what such a configuration requires. They are not what such a configuration is. Cosmopsychism’s unity is preserved at the level of structure; what the argument removes is the further claim that the structure is itself a subject.

The decombination problem, on the present account, is no longer a problem the cosmopsychist must solve in its original form. Decombination is the question how one consciousness gives rise to many. If consciousness is not the fundamental ground, the question takes a different shape: the structural conditions support multiple configurations of self-modeling, and finite subjects are among them. Cosmic-subjecthood is one such configuration the structure can support, not the source of the others.

This also answers the parsimony argument Goff presses in the cosmopsychist’s favor. Cosmopsychism advertises itself as simpler than its rivals: leaner than dualism, since it posits no second fundamental kind, and more honest than functionalism, since it incurs no strong-emergence debt for the experiential. Granted on both counts; but the advantage is won against dualism and brute physicalism, and it does not transfer here. The structural account posits a single fundamental category and no strong emergence (experience is the inside-aspect of structure, §§4.3.1–4.3.3, not a second kind bridged to the first nor a product conjured from a non-experiential base) so it matches cosmopsychism’s parsimony scorecard line for line. Where the two part, the structural account posits less at the ground: cosmopsychism must seat a subject there, and a subject is, by the account of §7.8 and Chapter 8, an articulated configuration of self-modeling under structural constraints, the more demanding posit, not the simpler one. The priority reversal is therefore also the parsimony verdict: grounding in the structural conditions, from which subjecthood is derived, is leaner than grounding in a cosmic subject that already presupposes them.

What converges. The convergence with the structural account remains significant. Cosmopsychism and the present account agree that there is a single fundamental ground and that finite minds are configurations within it. The disagreement concerns whether the ground is a subject or a structure. The argument of §7.8 and Chapter 8 establishes that subjects, as configurations of self-modeling under structural constraints, presuppose the structural conditions; the structural conditions do not presuppose subjects. Cosmopsychism’s cosmic subject is then a particular configuration the structure can support rather than the structure’s ground.

4.3.5 Idealism

In 1710 George Berkeley argued that to be is to be perceived. In 1781 Immanuel Kant argued that the conditions we find in experience (succession, extension, substance) are contributed by the cognitive subject rather than found in mind-independent reality. In 1900 Edmund Husserl began a phenomenological program that would relocate the world’s structure to the consciousness that constitutes it. In 2019 Bernardo Kastrup argued that the most parsimonious metaphysics is one on which all reality is a single field of cosmic consciousness, individual minds its dissociated alters. Four positions, three centuries apart, share one move: they relocate the structural conditions to mind. If the move succeeds, the necessity established in this book is a necessity about cognition, not about reality — it shows what any experience requires, not what there is.

This section makes one claim against the whole family, and makes it once. Whatever a metaphysics takes to be fundamental (divine perception, the transcendental subject, the constituting ego, a field of cosmic consciousness), if it is structured enough to do the work the position needs it to do, it instantiates T, S, and Φ (Temporality, Spatiality, and Physicality: the three structural Conditions, borne by Time, Space, and Substance). Idealism does not remove the conditions from reality. It relocates them to whatever it puts at the bottom. The relocation is the concession the argument needs, offered under another name.

Why the relocation is forced. The idealist inference runs from a premise about priority to a conclusion about reality: because mind is where the conditions are met, in perception, in intuition, in constitution, the conditions are features of mind rather than of what there is. The inference fails at a single joint. To be a mind of the relevant kind is already to host content, and to host content is to sustain distinguishable items (S) in an ordered succession (T) borne by something that persists through it (Φ). Strip away all distinction, succession, and persistence, and what remains isn’t a minimal consciousness — it is no candidate for hosting a world at all. So the conditions the idealist locates in mind are conditions on there being a mind to locate them in. That the mind is fundamental does not touch the point: fundamentality is a claim about explanatory order, while structure is a claim about what the fundamental thing has to be like to play its role. The two are compatible, and idealism needs them run together — it needs ‘consciousness is fundamental’ to deliver ‘the conditions are mind-imposed,’ where the first leaves the second exactly where it stood.

Whether there is a contentless or objectless awareness that hosts no world at all is a question about the character of experience, taken up where experience is treated directly; it does not bear here, because the fundamental idealism posits is precisely one that must yield a structured world of appearance. The strongest real instance of that further question is acosmic non-dualism (Advaita Vedānta’s nirguna Brahman,²³⁵ and the contemplative literature’s ‘pure consciousness event’),²³⁶ which holds that the world of appearance is not, at the deepest level, real at all: exactly the premise this section’s own inference needs and does not itself defend. Naming the case this precisely is owed even where closing it is not. A reader who wants the acosmic claim engaged on its own terms should look for it at §8.14.5, where the non-dual traditions are engaged directly and both the two-truths and objectless-awareness forms of the claim are met, not assume this section has already closed it.

A boundary worth marking before the four forms. Not every observation about perceiver-relativity is an idealist premise, and the difference is worth having explicit before Berkeley’s own argument is engaged. Locke’s classical distinction between primary qualities (shape, extension, motion, structural facts about a mind-independent object) and secondary qualities (color, heat, wetness, taken as powers an object has to produce a certain effect in a perceiver) grants that some felt qualities are perceiver-relative while leaving a mind-independent bedrock fully intact — a view fully compatible with physicalism, committing nothing to idealism. Berkeley’s actual step is a further and heavier one: primary qualities are just as perceiver-relative as secondary ones (a coin looks elliptical from an angle and circular from above; an object’s size varies with viewing distance), so the same reasoning that demotes color and heat to the mind demotes shape and extension as well, and no privileged mind-independent remainder is left standing anywhere to stop the collapse. Only that further move reaches idealism; granting that felt qualities are perceiver-dependent, on its own, is Locke, not Berkeley, and does not yet concede anything this section needs to answer.

Berkeleyan idealism. Berkeley (A Treatise Concerning the Principles of Human Knowledge;²³⁷ Three Dialogues)²³⁸ holds that all reality is mental, what seems mind-independent being sustained by God’s perception. Grant it. God must then perceive, and to perceive is to hold a perceived-this apart from a perceived-that, borne across the standing of the perceiver: a T/S/Φ event at the level of the divine mind. The conditions hold inside Berkeley’s own ontology; the idealist has renamed what bears them, not removed them. The framework-invariance argument at §6.1 develops this, showing that the inversion of classical theism holds under an idealist picture as readily as under a physical one.

Transcendental idealism. Kant (Critique of Pure Reason)²³⁹ gives the family its most disciplined form. Space and time are forms of intuition, substance and causation categories of the understanding; the conditions are necessary for any possible experience precisely because the subject contributes them, so they disclose the form of experience rather than the character of things in themselves. This is the version worth slowing down for, because it does not deny the conditions — it grants their necessity and reads it as a fact about us. The reply, developed at §2.2.3 (where Stroud’s substitution objection²⁴⁰ and Langton’s Kantian Humility²⁴¹ are worked through), turns on what is left on the far side of the veil. The book argues that a purported thing in itself with no ordering, differentiation, or bearing is indistinguishable from ∅ (§2.7). Kantian humility can resist the inference: absence of predicates available to us need not be absence of noumenal character. The ∅ fork is therefore the book’s substantive metaphysical challenge to the noumenon, not a result Kantianism concedes merely by restricting our knowledge. The parting from Kant is over the status of that result. His verdict is a limit on knowledge: the thing in itself is real but cognitively sealed off. The claim here is not epistemic but metaphysical — a fundamental with no temporal, spatial, or substantial character is not unknowable but nothing, and the sealed-off remainder Kant preserves is, pressed for content, ∅.

The retreat to imagination does not reopen the escape. To imagine a form is to instantiate the field within which it is the form it is: the act opens a region of placement (distinguishable loci, a relational order, distinctions that hold across the act), and the imagined form takes a position in that region by being imagined at all. Merely conceptual names S relocated to the cognitive register, at a finer grain, rather than S’s absence. Two cautions keep the point honest. What the imagined region instantiates is structural S (distinguishable loci in a traversable order, §2.4.7), not a metric manifold, so no imagined Euclidean space is smuggled in. And the claim runs opposite to Kant’s: the imagined form is structured because the imagining is itself a T/S/Φ event on the substrate side (§2.7), not because extension is the necessary form of any intuition — the line that would concede the very transcendental idealism the noumenon-collapse at §2.2.3 is built to block.

Phenomenology. Husserl’s transcendental phenomenology (Logical Investigations;²⁴² Ideas I)²⁴³ generalizes the Kantian move: the world is constituted in consciousness through intentional acts, and the conditions of constitution are the conditions of any world-for-a-subject. Heidegger (Being and Time)²⁴⁴ recasts it existentially, Merleau-Ponty (Phenomenology of Perception)²⁴⁵ embodies it. Grant the whole program: constitution is still a temporal, spatial, substantial process, and the constituting ego is still a being of T, S, and Φ. Husserl’s own genetic phenomenology (passive synthesis, the time-consciousness of the Lectures on the Phenomenology of Internal Time-Consciousness) rediscovers the conditions inside consciousness, which is the argument’s claim located one level down rather than refuted. The reduction does not escape the conditions; it arrives at them. The engagement runs in full at §2.2.3.

Analytic idealism. The contemporary form is the one the earlier chapters do not already carry, and it is the strongest, so the section spends its own argument here. Bernardo Kastrup (Why Materialism Is Baloney;²⁴⁶ The Idea of the World;²⁴⁷ Analytic Idealism in a Nutshell)²⁴⁸ holds that consciousness is fundamental and matter is its appearance from within a particular cognitive perspective. Unlike Berkeley he needs no personal God: the cosmic field is impersonal, neither omniscient nor creator. Unlike Kant he makes a positive claim about what reality is, not a boundary around what can be known. The engine is dissociation — individual minds are dissociated alters of one cosmic consciousness, and the outward face of the physical world is how one alter’s mental contents look from without. Adjacent, by a different line, is Philip Goff’s cosmopsychism (Galileo’s Error;²⁴⁹ Consciousness and Fundamental Reality),²⁵⁰ which starts from a single cosmic mind and treats individual minds as its decompositions; David Chalmers (Reality+)²⁵¹ is sympathetic to neighboring views without taking on analytic idealism as such. The position has real defenders and real opponents (Frankish’s illusionism²⁵² is the contrary pole; Stoljar’s type-B physicalism²⁵³ the standing alternative), and it earns a real engagement.

Why the ground is still called mind. One question precedes the structural reply and sharpens it. Every mind the theory can exhibit, each alter it sets out to explain, is personal: it is someone’s, undergone from a point of view, a subject to whom its contents are present. The fundamental field is none of this, being impersonal, subjectless, without a perspective from which anything is had, so the word mind is made to span a discontinuity as wide as the gap it was introduced to close.

Pressed on what entitles the label, the position is caught in a pincer whose jaws close from opposite sides. Make the field impersonal, as Kastrup does, and the word mind thins toward an honorific. A ground stripped of subject, of perspective, and of any unity a self would bring keeps the name by bare experientiality alone, and owes an account of why that is a mind rather than a posit the word merely flatters. Make it personal instead, a genuine subject with a point of view, and two bills arrive at once: the plural alters collapse into the contents of a single subject, which is Coleman’s problem²⁵⁴ and not the plurality the view needs, and a personal ground of all reality stands most of the way to God.

That second cost is the one worth pressing. A personal cosmic mind need not be divine by stipulation, since one might posit a limited cosmic subject, neither omniscient nor creator. But the impoverishments that would keep it short of divinity are hard to sustain for a mind in which every other mind is contained and from which all reality is said to dissociate. What knows all contents looks omniscient, what all appearance issues from looks creative, and the idealist now owes an account of why this universal personal ground stops short of the divine one the view defined itself against.

The jaws are connected, which is what makes the pincer bite: personhood is what rescues the field from the honorific charge and just what delivers it to the divine one, while the retreat toward a blind, instinctive ground that escapes divinity is the retreat that drains the word mind. The band between them, personal enough to be a mind yet impersonal enough not to be God, is a narrow one, and whatever is set in it must still be a distinguishable, persisting, successive ground.

This is the section’s structural point wearing its conceptual face. What makes an alter a mind and not a stretch of the field is that a distinguishable someone (S) persists as that someone (Φ) across a succession of states it undergoes (T); subtract the three and the personal mind is not rendered impersonal but unmade. The contrast with theism is then exact, since the two positions share the word and invert its content: classical theism makes the ultimate mind by making it maximally personal, a knower and willer, while analytic idealism makes the ultimate mind by making it sub-personal, a field that is no one’s. The shared term names contraries, and they meet only where the argument says every ground must, in the conditions that let anything be a knower or an experience at all.

The lineage runs through Schopenhauer, whom Kastrup cites directly (The Idea of the World).²⁵⁵ Schopenhauer (The World as Will and Representation)²⁵⁶ reads the Kantian thing in itself as Will: a blind, impersonal striving, not cognition but sheer drive, individuating into organisms as its appearances. This is the template at its most austere — not cosmic mind but its non-representational twin, a ground neither mental nor material in any phenomenal sense. One observation answers it. The Will strives, and striving is a process. To strive across resistance is to have a before in which the resistance stands and an after of movement against it (T); to strive against something distinct is to individuate, and individuation is non-identity within a manifold (S); to remain the particular Will-form whose striving marks it across its encounters is to persist (Φ). Schopenhauer’s Will is phenomenally undifferentiated (without qualia, that is, without any felt qualitative character, without representation, and without a subject), yet it is structurally constituted in the very act of being the kind of ground that individuates into organisms and holds those individuations through time. It does not escape T, S, and Φ by being blind; it instantiates them in the mode of blind striving rather than conscious representation.

Three forms of the analytic-idealist move show what each commits to.

First, dissociation. Individual minds are dissociated alters of cosmic consciousness; the world’s outwardness is the look of one alter’s contents from without. Kastrup’s own model for this is clinical: dissociative identity disorder, in which one psyche presents as several alters that do not share access. The analogy is exact enough to use, and it is where the reply lands. For alters to be distinct rather than identical to the field, the field must differentiate (S); for the ground to be stable enough for alters to form from it, it must persist (Φ); for dissociation to be the dynamic process the view needs (alters forming, holding, dissolving), it must run in time (T). Even a single case of dissociative identity is a structured process in a structured psyche; nothing weaker produces alters. So the cosmic field is not the absence of T, S, and Φ but a vast instance of them under another name. On this reading Kastrup’s view is a panentheist relabeling (panentheism: all reality is within the cosmic mind, which yet exceeds it) of the structural account rather than its rival.

The dissociation model also faces a pointed objection of its own, and meeting it strengthens the reply. Coleman²⁵⁷ argues that subjects do not sum or divide: a single consciousness cannot host genuinely separate points of view, so the clinical analogy either describes one subject with partitioned contents, which is not the plurality idealism needs, or helps itself to the very plurality it was meant to explain. The objection is real and it tells against Kastrup’s use of the analogy. It leaves the structural point untouched and in fact sharpens it. On the first horn, partitioned contents within one subject are distinguished from one another (S), held apart across the partition (Φ), and sustained through the dissociative dynamics (T); on the second, genuinely plural subjects require a field able to individuate and bound them, which is those same conditions read at the substrate. Coleman’s is a problem about how a mind divides, and every candidate answer to it is stated in T, S, and Φ.

The sharpest version of Kastrup’s reply targets the level at which the conditions are claimed, not the fact that dissociation is structured. Grant, he can say, that the alter’s world is structured — that is just the appearance. What the argument calls Nature might be exactly that appearance: the world as experienced from within a dissociated alter. Then the necessity of Nature is a necessity of appearance, not of the substrate; T, S, and Φ are features of the alter’s experienced world, which he happily grants, while the substrate (undifferentiated cosmic consciousness prior to any dissociation) is bound by none of them. This is the best objection in the section, and it has a clean answer: dissociation is substrate-side. A substrate that produces an alter possessing T, S, and Φ must, at its own level, be able to differentiate (to have this alter rather than that), to hold the differentiation across the stretch in which the alter keeps a separate perspective (Φ), and to have the producing be an event rather than a timeless fact (T). Those are not features of the alter’s appearance; they are conditions on the substrate’s power to generate an alter at all. The appearance/substrate distinction is therefore not structured-versus-unstructured but two levels of structure: the alter’s experienced conditions and the substrate’s productive ones. Kastrup’s substrate is undifferentiated in the phenomenal sense (no qualia, no point of view), and the argument does not contest that; it is not undifferentiated in the structural sense dissociation requires, and that is all the argument needs.

Second, cosmopsychism. Goff reaches nearly the same place from the other direction: bottom-up panpsychism faces the combination problem (how do micro-experiences combine into a macro-experience?), so start instead from one cosmic experience and treat individual minds as its decompositions. Cosmopsychism sits closer to the structural account than either physicalism or bottom-up panpsychism, but its modal scope decides its fate. If the cosmic consciousness is contingent (able to fail to exist, or to exist without structure), then absolute nothingness is possible, a possibility foreclosed at §2.8. If it is necessary, it must meet the structural conditions, and the cosmopsychist now owes the very modal argument the structural account supplies. And the combination problem returns inverted: a cosmic mind that decomposes into alters decomposes along lines of distinguishability (S), persistence (Φ), and temporal differentiation (T), so the Triconditional Cross-Entailment (§2.6.7), that T, S, and Φ require one another, applies to whatever the cosmic mind is. Decomposition confirms the conditions at higher resolution; it does not step around them.

Third, information. Chalmers’s later work (Reality+)²⁵⁸ and the simulation literature²⁵⁹ press a related thought: perhaps the world is a simulation and what is fundamental is information or computation, not matter. Taken from the information-theoretic side, this is engaged at §4.1.5; what this section adds is that information is itself a structural notion. It requires distinguishable states (S), transitions between them (T), and a substrate that carries them (Φ). A simulated world exhibiting T, S, and Φ is a world exhibiting T, S, and Φ, and the simulator’s own substrate is structured in turn. The regress ends at the conditions, not before them.

What the position actually buys. Analytic idealism’s deepest attraction is that it dissolves the hard problem of consciousness: if experience is fundamental, the question how non-experiential matter generates experience cannot arise, there being no non-experiential matter to do the generating. That dissolution is genuine — taking experience as primitive is no more illicit than any account’s taking its own ground as primitive. But the hard problem doesn’t retire here — it trades, and the trade is asymmetric in kind. Physicalism owes a cross-category derivation, the experiential from the non-experiential, which is the hard problem proper; idealism owes only an intra-category one, the structuring of what is already experiential. That asymmetry is the one place the comparison stops mirroring, and it falls to idealism: unlike a parsimony tally, which runs both ways, this debt is lighter on the idealist side.

How much lighter is worth pressing, because the intra-category debt is not small. Idealism must say why the fundamental consciousness is differentiated and structured at all, dissociated into alters and articulated into a lawful world, rather than an undifferentiated field. Set beside the hard problem this looks like the easier question, an internal shaping rather than a category crossing, and it is easier only in category, not in force. The move from an undifferentiated substrate to a structured world of alters is a move from no T, S, or Φ to all three; it is the appearance of articulated structure from a ground said to have none, which is a structural analogue of the hard problem run one level down. The analogy should not be overstated: both terms stay within the experiential, so this is not the phenomenal-generation gap itself, and to that extent idealism’s dissolution of it is genuine. What recurs is the gap’s form, not its content: an unexplained passage from the homogeneous to the differentiated, and that formal debt is discharged only in T, S, and Φ. Idealism has not dissolved the problem of how structure arises; it has renamed the terrain on which that problem is posed, from matter-and-experience to substrate-and-differentiation. And its own answer, when it gives one, must credit the substrate with the power to differentiate, to persist, and to unfold in time — with exactly T, S, and Φ. The generation question is won; the differentiation question is inherited; and the differentiation question is answered only by the conditions whose fundamentality the position set out to deny. What experience positively is, the identification with self-modeling structure, is Chapter 8’s work, resting on the substrate-neutral ground supplied at §7.8; the point here is only that the differentiation debt is structural.

The same lineage makes a second appeal, to parsimony: idealism posits only experience, where physicalism posits experience and non-experiential matter and still owes the passage between them. The general reply is at §5.1.4 — comparative parsimony cannot adjudicate between structure and an alternative to it in either direction, because the count presupposes the very conditions (distinguishable posits set against a common measure) it would use to decide. Here the count that makes a single field look simpler omits the differentiation the field must undergo to appear as a world of alters, and that differentiation just is T, S, and Φ at the level of the cosmic mind. Entered on the ledger it is no saving: the apparent economy is bought back exactly where the work is done. Parsimony cannot select idealism over the structural account, because what it would economize on is what idealism must keep.

One objection cuts across all four forms. The reply to every idealism has the same shape, that whatever is made fundamental turns out to be structured, and that very uniformity invites the charge that the argument is unfalsifiable: no matter what the idealist names, the section will declare it structured enough, so the thesis rules nothing out and says nothing. The charge would land if the conditions were imposed by stipulation. They are not, and the section accepts the burden the charge identifies. There is a way for idealism to win: specify a fundamental that hosts a world of appearance while instantiating none of T, S, or Φ (no distinction among its contents, no order to their succession, nothing that persists to bear them), and the argument fails. The claim is not that such a posit is forbidden but that every attempt to specify it collapses into ∅, the state that on examination is no state at all (§2.7, §2.8). The idealist is owed that opening and has it; what the idealist lacks is a filling for it. Conceded, then, is just what the symmetry objection concedes against the cosmological tradition elsewhere in the book (§5.1.2): the reply earns no victory by fiat. It earns it only because the unstructured fundamental, pressed for content, keeps turning out to be nothing.

The dispute is downstream of the conclusion. Step back from the four forms and the shape of the whole engagement comes into view. Idealism and physicalism are two redescriptions of one skeleton. Subtract the words mental and physical from each, and what remains on both sides is the same thing: distinguishable items held in an ordered succession by something that persists — T, S, and Φ carrying the explanatory load. Their dispute is over which vocabulary the skeleton wears. The book’s positive conclusion is a claim about the skeleton, not the vocabulary, and it is therefore settled before the mental-versus-physical dispute is even joined. This is why the argument need not defeat idealism to reach its end. It need not choose between idealism and physicalism to establish its positive conclusion (that conclusion holds whatever the ground’s metaphysical character turns out to be, as established at §2.2.3); and it need not defeat idealism to overturn classical theism (the inversion holds under idealism as readily as under physicalism, as established at §6.1). What it requires of idealism is only what idealism, on examination, always grants: that whatever is fundamental be structured.

What converges. Idealism in its four forms (Berkeleyan, transcendental, phenomenological, analytic) shares with the structural account a single refusal: it will not derive consciousness from antecedent non-conscious matter. The four part company over what is fundamental (divine perception for Berkeley, the conditions of possible experience for Kant, the constituting ego for Husserl, cosmic consciousness for Kastrup), and the book’s reply is one and the same across them, distributed over §2.2.3, §6.1, and this section: whatever is put at the bottom, if it is structured enough to do the work the position needs, it instantiates T, S, and Φ. The conditions are conditions on anything that hosts process, and a mind is one kind of host. So the structural conditions are mind-compatible without being mind-derived: consistent with consciousness being widespread and substrate-independent (the substrate-neutral ground is supplied at §7.8, life as multiply realizable closure-plus-throughput, and the consciousness identification that rides on it is carried in Chapter 8), and consistent with its being fundamental in a sense the argument need not adjudicate. What the argument is not consistent with is the one claim idealism would need and never establishes: that what is fundamental is unstructured, that it lacks T, S, and Φ. That claim, examined, identifies nothing.

4.3.6 The Unreasonable Effectiveness of Mathematics

Riemann’s 1854 habilitation lecture generalized geometry beyond Euclidean space and introduced the conceptual machinery later developed into Riemannian curvature. Roughly sixty years later, Einstein, working with Grossmann and others, used differential geometry to formulate general relativity and describe effects including the bending of light. Sophus Lie’s 1873 transformations answer to no force or particle known at the time; sixty years later, the symmetry groups he catalogued become the structural language in which the strong and weak interactions are written. David Hilbert worked out the theory of infinite-dimensional function spaces in the 1900s as part of a program in pure functional analysis; by the late 1920s Hilbert-space methods became a central formulation of quantum mechanics, though later algebraic, path-integral, and other formulations show that no single notation should be called uniquely indispensable. The pattern is not one anecdote or two. It is, as Wigner said, unreasonable.

Wigner (The Unreasonable Effectiveness of Mathematics in the Natural Sciences, 1960) gave the observation its classical statement and left the explanation open. Hamming (1980) restated the puzzle in a working mathematician’s voice and left it as unresolved as Wigner had. Steiner (The Applicability of Mathematics as a Philosophical Problem, 1998) pressed the quantitative version: specific anticipations, each precise, cumulatively demanding an explanation a naturalist cannot easily supply. Plantinga (Where the Conflict Really Lies, 2011) developed the theistic interpretation rigorously: if reality and mathematics share a common rational origin, the fit is intelligible; the designer who authored the world also authored the minds that describe it. Tegmark (Our Mathematical Universe, 2008, 2014) pressed the radical inversion: the universe is a mathematical structure, and applicability is identity. Against the theistic reading stand naturalistic explanations of varying depth: Pincock (Mathematics and Scientific Representation, 2012) developed a ‘mapping account’ on which mathematics applies when the abstract structure models the concrete; Bueno and French (Applying Mathematics, 2018) pushed a partial-structures version; Colyvan (The Indispensability of Mathematics, 2001) defended the Quine–Putnam line as a foundation for mathematical realism without theistic auxiliaries. The present reply sits closest to the structural realists (Ladyman and Ross 2007) in treating applicability as a fact about shared structure rather than about correspondence between two independent domains, and specifies what the bare position does not: which structure is shared. It is equally not Tegmark’s Mathematical Universe Hypothesis: applicability is here grounded in the instantiation of the T∧S∧Φ architecture, not in Level-IV plenitude or an identity of physical reality with abstracta, while §4.2.4 leaves robust Platonism and uninstantiated abstract structure as live positions rather than identifying them with ∅. The agreement of mathematics and physics is one instantiated structure recognized twice, not reality being mathematical; the full structural-realism-versus-MUH non-equivalence is developed at §4.3.7.

The challenge bears directly on the Triconditional Identity of §2.6.10 and on the Necessity of Nature of §2.8.5. The applicability theorist replies that the precise correspondence between a priori mathematical structures and a posteriori physical reality is a positive fact about the world that the structural roles do not explain, and that the best explanation appeals to an authoring mind whose rational character underwrites both. If the explanatory ask is well-founded, either Nature is not self-sufficient (against §2.8.5) or the structural roles do not exhaust what a complete account must specify (against §2.6.10).

A clarification before the moves. The puzzle is not merely that mathematics is useful. Utility is a weak notion that any formal tool might satisfy. The puzzle is that specific mathematical structures, derived for internal reasons decades or centuries before any physical application, turn out to match the physical world with high precision: Riemannian geometry (smoothly curved manifolds generalizing Euclidean surfaces to arbitrary curvature), Lie groups (continuous symmetry groups indexed by smooth transformations), and Hilbert spaces (infinite-dimensional state spaces equipped with an inner product), each arriving years before the physics that needed them. Three forms of the move press that specificity at three resolutions (the general observation, Steiner’s anticipation argument, and the precision claim), and the space they cover is the space the applicability theorist occupies. The replies test whether any form can preserve the external/internal gap it requires to get started.

First. The unreasonable-effectiveness move holds that mathematicians develop structures for internal reasons (elegance, generality, formal consistency) and those structures turn out to describe physical reality. The fit is not accidental; it is evidence physical reality and mathematical structure share a common origin. A common origin is most naturally a rational mind.

Mathematics is not a script the universe follows from outside. It is a description of the structure of the medium itself, produced by cognitive faculties that are themselves configurations of that same medium. The mathematician’s reasoning proceeds under the same constraints that govern the physical world, because both mathematician and world instantiate T, S, and Φ. There is no gap between the describer and the described requiring a bridge. What looks like pre-established harmony between two domains is a single necessary architecture recognized from two locations within itself.

Second. Steiner’s anticipation argument presses from another angle. Pure mathematics, developed with no physical application in view, repeatedly anticipates the deep structure of physical reality. Riemannian geometry existed sixty years before general relativity needed it. Group theory existed decades before particle physics needed it. Anticipation is the point: mathematics is not merely consistent with physics; it arrives first. Arriving first requires explanation.

The anticipation dissolves when one asks what the mathematician was doing. Riemann developed his generalized geometry by following the logical consequences of letting curvature vary freely, by asking what relational structures are possible given Space’s constraints at a certain level of abstraction. Lie developed his groups by asking what transformations preserve structural identity across a manifold: the same question Space asks at higher generality. The mathematician is not transcending the structural roles to anticipate them from outside. The mathematician is mapping T, S, and Φ from inside, following the logical consequences of what the roles necessarily permit. When physics later needs those structures, nothing has been pre-tuned. Both the mathematical investigation and the physical phenomena were already constrained by the same necessary architecture. The anticipation, on this reading, is not external fit but two independent descriptions meeting at their point of shared necessity.

Third. The specific-precision move appeals to high resolution. It is not merely that mathematics is applicable: it is that the specific mathematical structures match the specific physical phenomena with astonishing exactness. Hilbert spaces are exactly what quantum mechanics requires. SU(3) (the special unitary group of 3×3 matrices, parameterizing the color-charge symmetry of quarks) is exactly the symmetry group of the strong force. Riemannian geometry with Lorentzian signature is exactly what general relativity needs. The precision is too exact to be coincidence. It suggests a designer who chose both.

Specific precision is what structural necessity looks like at high resolution, though the examples should not be overstated as if each formalism were uniquely forced independently of theory. Lie groups capture the transformation-structure physics repeatedly needs when asking what preserves structural identity across a manifold. Riemannian geometry becomes the right language for general relativity once extension, curvature, and causal distribution have to be represented together. Hilbert spaces provide the standard formalism for quantum mechanics when states are represented so as to preserve superposition, transition amplitudes, and unitary evolution. The specific structures are not tuned to fit the physics from outside; they are mathematical vocabularies that repeatedly articulate the same structural roles at different levels of the architecture. The precision, on this reading, is the signature of shared structure recognized from two directions rather than a sign of external authorship.

A word is owed on the probabilistic register Plantinga and Steiner share. Plantinga’s inference is abductive (the fit is more probable given a common rational origin than given naturalism), and Steiner’s is cumulative: each precise anticipation is individually unlikely, the conjunction far more so. Both ledgers presuppose that the fit is a contingent datum whose probability theism raises and naturalism leaves low. The structural reading does not compete on that ledger; it contests how much contingency the ledger has to score. The shared structural architecture can explain why mathematics is broadly applicable, but T, S, and Φ are too coarse to necessitate the particular success of SU(3), Hilbert-space quantum theory, or Lorentzian geometry. Those high-resolution fits depend on contingent physical laws, theory selection, and the richness of mathematics. The applicability puzzle is therefore reduced at the general structural level, not dissolved as a single necessity. The independence Steiner’s cumulative case needs (each anticipation a separate low-probability event whose improbabilities multiply) is the first casualty: the cases do not multiply because they are not independent coincidences but the same architecture surfacing again. The comparison is narrowed rather than conceded: what remains contingent is the particular physics, not a further mathematics–world harmony standing over it, and a residue of that shape is not a datum theism explains better than naturalism.

A kindred move drops Plantinga’s theistic framing for mathematical Platonism: the Quine–Putnam indispensability arguments (Quine 1948; Putnam 1971) treat mathematical truths as abstract facts physics cannot dispense with, and the explanatory puzzle reduces to correspondence between Platonic abstracta and concrete physics. The Platonic challenge is engaged at §4.2.4; the version here assumes its result: no mathematical view by itself supplies a counterexample to the structural conditions in concrete change-realizing worlds, and the structures mathematicians apply to the world are structural roles described under mathematical vocabulary.

The information-theoretic ontology that treats reality as fundamentally informational is §4.1.5; the mathematical-universe hypothesis, that reality is itself a mathematical structure, is §4.3.7. The hard problem of consciousness, which presses a different mode of non-structural residue, is §4.3.1. The contemporary-physics challenges (Bell non-locality, Everett, and the quantum-field-theoretic dissolution of the particle concept) are §4.3.8.

What converges. Each of the three forms collapses the external fit it required to get started. The unreasonable-effectiveness move treats mathematics as a script the universe follows from outside, when it is a description of the medium itself produced by cognitive faculties that are themselves configurations of that medium. Steiner’s anticipation dissolves once the mathematician is located inside the architecture: Riemann and Lie were mapping T, S, and Φ from within by following the logical consequences of what the roles necessarily permit, not anticipating them from outside. The specific-precision cases show that mathematical formalisms can capture physical structure at high resolution. The structural roles help explain why such mapping is possible, but they do not uniquely derive those formalisms or exclude alternative mathematical representations. External authorship is therefore not forced, yet neither is the detailed fit exhausted by the thin triad alone.

Wigner delivered The Unreasonable Effectiveness of Mathematics in the Natural Sciences: a physicist asking why mathematicians’ work kept arriving in his hands before he knew he needed it. He offered no explanation. He named the fit unreasonable and let the word do its work. On the structural account, the word does less work than it was originally asked to carry. Riemann developed his geometry under the constraint of a mathematical practice conducted in Space and Time by a persisting Substance: Riemann himself, alive, thinking. Lie developed his groups by asking what transformations preserve structural identity across a manifold: the question Space itself asks at higher generality. Hilbert built his spaces out of relations between states, which is what Space builds at every level. The applicability the theist reads as the signature of external authorship is, on the present account, the signature of shared necessity. The ‘unreasonable effectiveness’ recedes once math is recognized as the internal grammar of Time, Space, and Substance. The mathematician who marvels at applicability is marveling that structural descriptions describe.

What answers Wigner’s unreasonable, on the structural account, is that the word named a surprise the architecture does not in fact support: the world is what the architecture looks like when a part of the architecture describes the whole.

Put in the vocabulary of Modal Pressure (§2.6.10): mathematics is what Modal Pressure looks like when attended to in the notation built to track inference, and physics is what Modal Pressure looks like when attended to in the notation built to track state transition. Both notations are tracking the same running structure. Their agreement is not pre-established harmony between two domains. It is one pressure, recognized twice.

4.3.7 Mathematical Universe Hypothesis

Tegmark (2008, 2014) defends the Mathematical Universe Hypothesis: physical reality is a mathematical structure. Not modeled by a mathematical structure, identical to one. Every mathematical structure exists in the same sense; the one we inhabit is the one whose internal observers detect themselves as inhabiting it. The hypothesis is meant to be the most parsimonious possible ontology: nothing exists but mathematical structures, and the question why our particular structure exists has no answer because every structure does.

The MUH is the clearest contemporary alternative ground to the present argument’s T-S-Φ (the three structural conditions, borne by Time, Space, and Substance). It deserves direct engagement.

A mathematical structure, on the standard account Tegmark relies on, is an isomorphism class of relations on a set. As pure mathematics, it may instantiate abstract relations without temporal succession, spatial extension, or causal persistence. The present argument concedes that much.

The issue begins only when Tegmark identifies a structure with physical reality and invokes internal observers. A structure that realizes observerhood, state evolution, and actual physical processes must supply more than an abstract isomorphism class: it must realize succession, differentiation, and bearing at role-resolution, and its physical realization must meet the further type-level burden developed in the book. The reply is therefore conditional on the MUH’s identity thesis. Pure mathematical structure is not refuted; a mathematical structure proposed as the complete physical realizer does not escape the structural roles.

Read in the present argument’s terms, the MUH looks less like an alternative to T, S, and Φ than a vocabulary for them: the mathematical-structure ontology specifying in pure mathematics what the present argument specifies as necessary conditions. But an equivalence asserted from one side is not yet an engagement. Tegmark’s thesis is an identity claim (physical reality is a mathematical structure), not the weaker claim that mathematics supplies one vocabulary among others for a Nature specified independently; and he would resist the redescription as flattening a substantive metaphysical thesis into a notational variant. The thesis must therefore be met on its own terms. Taken there, it does not escape the conditions. Any mathematical structure identical with an observer-containing physical reality must have differentiated positions or states and support an internal ordering adequate for observer-processes. Invariance under isomorphism is a criterion of structural identity, not temporal persistence or Substance. The Φ-role requires a further account of how determinate capacities or continuities are realized inside the structure. Tegmark may still call this a redescription rather than a discovery, but the resistance cannot consist in denying that his structure has distinguishable elements, persists, and carries succession, for a structure without these could not be the one we inhabit. The identity thesis does not bypass the structural conditions; it presupposes them.

A second pressure. Tegmark argues that every mathematical structure exists. The claim is comparative: the structure we inhabit is one of an infinite plenitude. The plenitude is not posited idly: its motivation is anthropic. The apparent fine-tuning of our universe’s constants needs neither a designer nor brute luck if every mathematical structure is equally real, since observers necessarily find themselves in one of the life-permitting members of the ensemble. This is the selection-within-an-ensemble strategy the fine-tuning argument meets directly (§5.1.5), and that verdict carries over here with one addition proper to the MUH. Even granting the ensemble, a structure can house observers who detect themselves only if it has distinguishable elements, persistence, and internal succession, so anthropic selection does not select away from T, S, and Φ; it ranges only over structures that already satisfy them. The plenitude universalizes the structural conditions rather than relieving them. And it faces a difficulty in its own right. Our reasoning about the plenitude is an act within a structured world, but that act-side fact does not determine the object-side ontology of the other structures. The sharper objection is whether Tegmark has supplied a clear, non-equivocal sense of physical existence applying equally to observerless and observer-containing mathematical structures. Anthropically selected structures must realize the roles required for observers; the plenitude as a whole need not.

What converges. What the MUH offers is a mathematical formalization of the necessary ground, not an alternative to it. Its parsimony is real (nothing exists but mathematical structure), but the structure it cannot do without is the structure the present argument names. The argument’s conclusion that T, S, and Φ are necessary is, translated into Tegmark’s vocabulary, the claim that mathematical structure is necessary; within his own hypothesis, T, S, and Φ are the conditions any structure must meet to be the one we inhabit. Read either way, they are not what the hypothesis leaves behind; they are what it cannot be stated without.

4.3.8 Non-Locality, Branching, and the Particle Concept

In 1964, John Bell wrote down an inequality any local hidden-variables theory of quantum mechanics must satisfy. Eighteen years later, Alain Aspect’s team confirmed experimentally that the inequality is violated in the actual world. Bell’s theorem, together with the experiments, rules out local hidden-variable accounts satisfying the theorem’s auxiliary assumptions, including the relevant independence conditions. It does not by itself select one metaphysical interpretation of quantum mechanics. That was one door. Everett’s 1957 proposal had already opened another: the quantum state never collapses but branches, with each measurement partitioning reality into decoherent branches, each of which contains a copy of the measuring apparatus and the measurer. And in the relativistic extension of quantum mechanics, as Doreen Fraser and David Baker pressed across the 2000s, the particle concept itself loses its frame-invariant referent. Three results. Three places where a piece of the classical metaphysical picture has had to give ground to the mathematics.

The contemporary literature on these three loci is well-developed. On Bell non-locality: Maudlin’s Quantum Non-Locality and Relativity (3rd ed. 2011) gives the most sustained metaphysical analysis; Wallace’s The Emergent Multiverse (2012) develops the Everettian alternative frame. On Everettian branching: Wallace 2012 stands as the systematic contemporary defense; Saunders, Barrett, Kent and Wallace’s Many Worlds? (2010) is the standard anthology; Albert (Quantum Mechanics and Experience, 1992) and Maudlin (Philosophy of Physics: Quantum Theory, 2019) press the major objections. On QFT: Fraser (2008) and Baker (2009) give the philosophical case against Fock-space particle ontology; Wallace (‘In Defense of Naiveté,’ 2006) and Ruetsche (Interpreting Quantum Theories, 2011) develop field-interpretive alternatives. The structural reply does not enter these debates as a rival position. It accepts the most demanding reading of each result and asks what the structural roles look like once that reading is incorporated.

The challenges so far in §4.3 have come from philosophy of mind and philosophy of mathematics. This final cluster comes from physics itself. Three questions press the account from contemporary physics: one from Bell’s result on correlations, one from Everett’s branching, one from QFT’s treatment of the particle concept. Bell non-locality appears to threaten Space’s relational non-identity: distant regions are correlated in ways no local structural description admits. Everett multiplies worlds in a way that appears to threaten Substance’s persistence and Time’s unique directed succession. QFT appears to remove the thing Substance was meant to name: particles, as well-defined countable objects, do not survive the relativistic formalism. If any of these is taken at face value, the Necessity of Nature of §2.8.5, supported by the Triconditional Identity of §2.6.10 (the result that Time, Space, and Substance, the bearers of the three structural Conditions Temporality, Spatiality, and Physicality, must be co-instantiated for any concrete existent), faces a counterexample from physics: the kind of counterexample a metaphysical argument has to address directly.

First. Bell non-locality holds that quantum correlations between spatially separated outcomes outrun local hidden-variable accounts. Bell (1964) proved a limit on correlations between spatially separated measurement outcomes that any local hidden-variables theory must respect. Experimental violations of that limit, Aspect et al. (1982), followed by the loophole-free experiments of the 2010s (Hensen et al. 2015; Giustina et al. 2015), establish that quantum correlations outrun local common causes. Maudlin (Quantum Non-Locality and Relativity, 3rd ed. 2011) has pressed the philosophical consequence with the most clarity: the world, on the best reading of the evidence, contains non-local dependencies between spatially separated events. If Space guarantees relational non-identity (if distinct positions are ontologically distinct loci) Bell’s result seems to show that distinctness is weaker than the argument requires. The outcomes display correlations incompatible with local hidden-variable factorization, while the no-signalling structure prevents their use for superluminal communication. Describing this as instantaneous causal influence would add an interpretation the experiments alone do not establish.

Non-locality refines Space’s specification without dissolving it. Bell’s inequality is a constraint on what kinds of correlations are possible between spacelike-separated outcomes; it is not a denial that spacelike-separated outcomes are distinct. The correlation presupposes the distinction: Alice’s measurement and Bob’s measurement are two measurements, at two locations, with two outcomes, and the correlation Bell quantifies is the relation between those two. What Bell rules out is a specific metaphysical picture: local hidden-variable determination. What survives is a richer structural picture in which non-identity is preserved but the relations binding non-identical positions exceed signal-bounded causation. Maudlin’s own preferred reading (non-locality as a fundamental relational feature of the world) is Space specified at higher resolution, not Space refuted. The argument has never required that all spatial relations be mediated sub-luminally. It has required that distinct positions be distinct. Non-locality leaves that requirement intact and tells us what else is true at its level. A sharper version of the worry does not stop at the correlations. If the entanglement-geometry program in quantum gravity is right — if spatial separation is itself emergent from entanglement structure, as the ER=EPR conjecture and allied approaches propose — then Spatiality is pressed one level below distinctness: the reply above secures that the outcomes are two, not that the spatial gap between them is fundamental. The structural answer runs parallel to the emergent-time handling at the QFT locus below. The S-role never required that spatial extension be fundamental, only that real differentiation be realized; a proposal on which space emerges from entanglement must still ground an order of real differences, and the role then attaches to whatever grounds the emergence — entanglement structure included, which is itself a structure of distinct, differently related relata. Whether any emergent-spacetime program succeeds is a live physical question, an alignment matter for the framework rather than something this survey derives; what no such program can deliver is differentiation from no differences at all.

Second. Everettian branching holds that the quantum state never collapses but forks into decoherent branches at every measurement. The Everett interpretation (Everett 1957; Wallace, The Emergent Multiverse, 2012) holds that the quantum state never collapses. Measurement interactions branch the universal wave function into decoherent branches, each containing a copy of the measurement and a copy of the observer. On Wallace’s mature development, branching is continuous and ubiquitous, the branches are real, and the classical world is one branch’s-eye view of a structure larger than any single branch. Time’s unique directed succession appears to split. Substance’s persistence appears to fork. The Triune architecture, specified at the level of a single temporal-spatial-substantial flow, appears to meet a physics in which that flow is one trajectory among indefinitely many.

The Everettian picture, at its clearest, does not dissolve Time and Substance. It multiplies them. Each branch instantiates the structural roles. Branching itself requires what any differentiating operation requires: distinct outcomes (Space), distinct directions of succession within each branch (Time), and the maintenance of causal capacity within each branch (Substance). On a realist Everettian reading, the universal wave function and its decoherent branch structure are taken as physically real. Whether branches are exact ontological loci or emergent, approximate patterns is part of the interpretation and should not be settled by the structural vocabulary alone. The argument’s claim that existence requires T, S, and Φ (Temporality, Spatiality, and Physicality) is satisfied within every branch. That there are many branches speaks to the scale of the structural commitment, not its nature. The Everettian owes the argument one answer: does the universal wave function, considered as the whole of reality, itself instantiate the structural architecture the argument requires? For T and S, it does. Unitary evolution and decoherence supply an ordered dynamical structure, while decoherent branch-patterns become distinguishable for practical purposes. Whether the universal state supplies the Φ-role as a persisting physical entity or only as mathematical state structure depends on the ontology adopted. Everett multiplies the domain across which the architecture is instantiated without replacing the architecture itself.

A qualification on dialectical scope is owed before proceeding. The preceding reply works on Wallace’s (2012) mature Everettianism, granting the branching structure and showing that it instantiates T, S, and Φ. Maudlin’s own position is not Everettian; he is among the interpretation’s most sustained critics. His principal objections concern branch individuation (the formalism provides no physical criterion for individuating branches, leaving the probability calculus without a derivable foundation), the Born-rule problem (recovering the Born-rule probabilities from branching ratios requires a notion of branch weight that cannot be derived from the wavefunction alone without circularity), and the absence of primitive ontology (Everettianism specifies a formalism without saying what physically exists in spacetime — the universal wavefunction is not itself local, and no catalogue of local beables is provided). These objections press whether Everettianism should be granted in the first place, not whether, granted, it preserves the structural roles. The structural argument is therefore not adjudicating between Maudlin and Wallace. What it establishes is scope-neutral: whatever interpretation of quantum mechanics is adopted, Everettian branching, Bohmian mechanics (whose particle trajectories are paradigm local beables instantiating S, Φ, and T directly), GRW collapse theories (whose flash or matter-density ontologies similarly supply local Φ-realizers), or algebraic approaches, the structural roles must be satisfied by whatever the interpretation countenances as its primitive ontology. Maudlin’s local-beable insistence is, from this vantage, a demand that the Φ-role be explicitly realized by something in spacetime; the structural argument agrees, and adds that every serious candidate for that something satisfies the role.

Third. The QFT particle-concept problem holds that the relativistic formalism dispenses with frame-invariant particles. Quantum field theory in its relativistic form dispenses with the classical particle concept. Fraser (“The Fate of ‘Particles’ in Quantum Field Theories with Interactions,” 2008) and Baker (“Against Field Interpretations of Quantum Field Theory,” 2009) develop the philosophical consequences. In interacting relativistic QFT there is no Fock-space particle description invariant across frames, no particle-number operator that commutes with the full Hamiltonian, no countable thing whose identity persists across the field’s evolution. Particles, as well-defined individuals, do not survive the mathematics. What survives, depending on interpretation, is a structure of fields, algebraic observables, excitations, state-transition constraints, or some more austere QFT ontology. If Substance was meant to name particle-persistents, QFT has removed its referent.

Substance was never specified as particle-persistence. Substance is defined at §1.4.3 as the maintenance of causal capacity across transitions, leaving open what realizes the maintenance. Particles were one candidate: the nineteenth-century candidate, inherited by the early twentieth. Fields are another. The reply does not simply trade particles for fields as if field ontology were uncontested. Whether the preferred interpretation speaks in terms of fields, local algebras, excitations, gauge-invariant structures, or state-transition constraints, the structural point is the same: the formalism preserves determinate capacities across transitions. There are states or observables that can differ, rules or constraints governing their transformations, and stable relations across which those transformations are specified. That is the role Substance plays here. QFT refines the realizer of Substance (perhaps beyond both particles and naïve fields) rather than eliminating the role.

Maudlin’s primitive-ontology insistence bears directly here: an unspecified disjunction of realizers (fields, algebraic states, excitations, gauge-invariant structures) is not yet an ontology, and a role-claim ungrounded in a specified ontology is, on his methodological view, physics-as-decoration. The structural argument’s response is not to choose among the candidates but to explain why the choice can be deferred at this level without losing force. The Φ-role specifies what any acceptable realizer must do: make a determinate physical capacity available to the ordered regime in which a real difference can occur. When the ontology is fundamentally dynamical, that is the maintenance of causal capacity across transitions. When a proposal treats time as emergent, it need not posit a fundamental clock-time; it must instead show how its fundamental constraints or states ground an emergent order in which determinate capacities can be carried and made effective. A proposal that denies any such capacity-and-order story has not yet supplied a realizer of change, rather than having supplied a counterexample to it.

This is a division of roles, not a collapse of Φ into T. T supplies the order of before and after; Φ supplies what can remain determinate and make a difference through that order. The present survey does not decide whether any timeless quantum-gravity program successfully supplies the required emergence account. That is a live physical question, and the framework’s use of the role here is abductive rather than a result it certifies in advance. Committing to one realizer would narrow the claim without strengthening it; remaining at the role level preserves the comparison across the live interpretive options. Where Maudlin’s demand is directed at physical theory (QFT owes a primitive ontology) the argument agrees, and defers to that ongoing debate on its own terms.

What converges. Each of the three questions finds the physics specifying the structural roles at higher resolution than the classical picture anticipated. Bell’s inequality is a constraint on what kinds of correlations are possible between spacelike-separated outcomes, not a denial that the outcomes are distinct; the correlation presupposes the distinction, and non-locality refines Space’s non-identity rather than dissolving it. The Everett multiverse multiplies the domain across which the architecture is instantiated without replacing the architecture: every branch instantiates T, S, and Φ, and the universal wave function supplies the ordered, differentiated structure at its own level. QFT refines the realizer of Substance beyond the classical particle concept without eliminating the role: the maintenance of causal capacity across transitions survives the loss of the particle concept. None of the three tells us the roles are dispensable. Physics, pressed to its contemporary frontier, leaves the structural roles intact: each result refines rather than dissolves a Time in which things happen, a Space in which things are distinct, and a Substance whose causal capacity persists across the happening.

Fraser and Baker have pressed the philosophical consequences of interacting quantum field theory that the physics literature more often takes up as technical rather than metaphysical questions. What contemporary physics finds at its frontiers is, on the present reading, the structural architecture itself rather than the classical picture specialists grew up pressing against.

4.3.9 Structural Realism

In a 1989 paper in Dialectica titled “Structural Realism: The Best of Both Worlds?”, John Worrall proposed that what survives across theory change in physics is not the entities science postulates but the structural relations encoded in its mathematics. The Fresnel equations survived the abandonment of the luminiferous ether; what was preserved was the structure, not the substance. Worrall’s epistemic structural realism (we can know the structure but not the natures of the relata) became, over the next two decades, the most discussed form of scientific realism. James Ladyman pressed the position further. In “What is Structural Realism?” (1998) and then with Don Ross in Every Thing Must Go (2007), structural realism turned ontic: there are no relata with intrinsic natures awaiting our knowledge of them; reality is structures all the way down. Steven French’s The Structure of the World (2014) gave the position its most systematic book-length defense. Michael Esfeld and Vincent Lam developed a moderate version on which objects are retained as nodes within the structure but stripped of independent intrinsic natures (Esfeld 2004; Esfeld and Lam 2008). Anjan Chakravartty’s A Metaphysics for Scientific Realism (2007) defended a semirealist alternative: structures and detectable causal properties are real, with dispositional natures grounding the causal properties. The position now occupies a continuum: epistemic structural realism (Worrall) at one end; moderate ontic structural realism (Esfeld, Lam) in the middle; eliminativist ontic structural realism (Ladyman, Ross, French) at the other. The challenge to the present account is that the trio T, S, Φ (the three structural Conditions, borne by Time, Space, and Substance) may be a parochial selection from a much richer relational fabric the position describes, and the argument’s necessity claim of §2.8.5 (Necessity of Nature), supported by the Modal Identity and Triconditional machinery of §§2.6–2.7, may be the wrong claim about the wrong feature.

The challenge sits closer to the present account than to its standard rivals, and the reply is correspondingly cooperative. The present argument does not posit T, S, and Φ as objects with intrinsic natures; it identifies role-level conditions of genuine change-realizers. OSR’s claim that the world is fundamentally relational is compatible with directed order, relational differentiation, and real bearing. Whether a given physical relational fabric also realizes temporal succession, spatial extension, and causal persistence is a further type-level question, not something settled merely by calling every relation Time, Space, or Substance. The disagreement concerns both the modal status of the roles and the step from roles to their realizers. OSR is typically silent on whether the relational structure it describes is necessary or contingent; the present argument fills in the modal claim. Three forms of the position sharpen what it can commit to and what each commitment costs.

First. The eliminativist form (Ladyman and Ross 2007; French 2014) eliminates objects entirely. There are no things; there are only patterns of relation. The challenge: any pattern of relations could constitute a world, and the trio T, S, Φ is a parochial choice from the space of relational fabrics. The reply runs through §§2.6–2.7: any pattern of relations capable of supporting a configuration (any pattern offered as a realizer of change) must include the three roles at the appropriate resolution. Bare obtaining directly supports determinacy or difference, but does not by itself entail temporal extension or causal persistence. Those stronger types require the separate change-realizer arguments. French’s careful development in The Structure of the World makes the point sharper, not weaker. The relations he identifies as fundamental (group-theoretic symmetries, gauge structures, permutation invariance) are relations among distinguishable items or positions (S). In familiar dynamical formulations, they are also articulated through an order of states or events (T) and through constraints or capacities that hold determinate physical difference in place (Φ). But a thoroughgoing atemporal-structure reading need not concede that the formal parameter in an equation is fundamental time. It owes a different account: how the atemporal structure grounds an emergent order in which real differences can occur, and how determinate capacities are available within that order. Merely redescribing the base as structure does not discharge those obligations; neither does the present argument certify that every timeless structural proposal meets them.

The division matters. T is the order of before and after; Φ is the determinate capacity or bearing condition that is exercised in that order. They are mutually required for real change but not the same role. The eliminativist version therefore does not by itself undermine the role-level claim. It reformulates ontology in relational terms and can resist intrinsic-nature-bearing objects. But it remains a live challenge at the type level: an eliminative structuralist may deny that relations require persistent causal bearers in the book’s richer sense. The present account can reply that a relational fabric proposed as physical reality or as a realizer of change must meet the relevant type arguments; it does not treat eliminative OSR as defeated merely because relations are named. T, S, and Φ are roles; the configurations satisfying the roles need not be intrinsic-nature-bearing objects in the sense OSR rejects.

Second. The moderate form (Esfeld 2004; Esfeld and Lam 2008) retains objects as nodes within the structure but denies them independent intrinsic natures. The position is straightforwardly consistent with the present account: configurations are nodes within the structural fabric, and T, S, and Φ are the relational roles the fabric must include for any node to obtain. Esfeld’s central claim (that nodes are exhausted by their relational profiles) itself requires the relational profiles to be temporally extended, relationally distinct, and persistent across the relations they enter into. The moderate OSR theorist and the present account converge: both deny that fundamental ontology requires intrinsic-nature-bearing objects; both affirm that fundamental ontology requires relational structure with the features the present account names. The disagreement, if there is one, is about whether the relational structure described by current physics (group symmetries, gauge invariance) exhausts what is fundamental, or whether the structural roles T, S, Φ are necessary at a level prior to physics. The present argument takes the second view and offers §2.6 for the Modal Identity result and the mutual-requirement / Triconditional machinery, and §2.7 for the collapse of the unstructured alternative: the three roles mutually require one another and cannot be reduced to a more primitive base.

Ladyman’s Principle of Naturalistic Closure gives the modal claim its sharpest methodological challenge: no metaphysical hypothesis should be accepted unless it serves to unify well-confirmed hypotheses from more than one specific science. The necessity of T, S, and Φ satisfies this constraint in an unusual but genuine way. The three structural roles are not a hypothesis derived from one science and then applied to others; they are the conditions on which any scientific inquiry operates. Temporal succession is presupposed by every state-evolution formalism in physics; it is equally presupposed by the experimental design structures in cognitive science, economics, and biology. Relational extension is constitutive of field theory, topology, and network science alike. Causal persistence is what conservation laws, developmental biology, and institutional economics each track, in different vocabularies, at different scales. The claim that T, S, Φ are necessary is not a physics claim that chemistry must accept on physics’s authority; it is a claim about conditions that each science independently instantiates whenever it describes anything at all. In Ladyman’s own terms, this is a unification claim (identifying the structural joints across all sciences, not merely within one). And it is warranted not by rationalist intuition but by the demonstration in §§2.6–2.7, specifically §2.6.7 (Triconditional Cross-Entailment), where the physics-facing case is made that any nomological structure requires T, S, and Φ as co-instantiated, that the three roles mutually require one another and are jointly presupposed by any coherent structural description. The Principle of Naturalistic Closure rules out free-floating neoscholastic machinery; the present claim is not free-floating but confirmed at every juncture where a science describes structure, which is to say at every juncture where science operates.

Third. The semirealist form (Chakravartty 2007) retains both structure and detectable causal properties, treating dispositional natures as the ground of causal-property detectability. The position is the most ontologically committed of the OSR family and the most consistent with the present account. Chakravartty’s detectable causal properties are configurations of T, S, and Φ at the level at which physics meets phenomena: dispositions are temporally extended (manifesting over time), relationally specified (manifesting in relation to other dispositions or stimulus conditions), and causally persistent (carrying their manifestation profile across the relations). The present account does not adjudicate between Chakravartty’s dispositionalist commitments and the more austere ontologies of moderate OSR; what it requires is what all three positions concede when fully developed: the relational fabric instantiates the structural roles.

A kindred move worth marking. The structuralist tradition in philosophy of mathematics, Resnik (Mathematics as a Science of Patterns, 1997), Shapiro (Philosophy of Mathematics: Structure and Ontology, 1997), runs a parallel argument in the mathematical case: mathematical objects have only relational properties, and the structures themselves are what is real. The reply parallels the OSR reply: a mathematical structure contains distinguishable positions or relational differences, directly supporting the S-role. Stability across mathematical instances is not automatically physical persistence, and the temporal course of our inquiry is an act-side feature rather than a property of the abstract structure. Mathematical structuralism therefore supports the thin relational point while leaving T- and Φ-realization open. The argument’s modal claim is consistent with both.

Two adjacent positions live elsewhere. The Tegmarkian “mathematical universe” hypothesis (Tegmark 2008, 2014) runs a parallel ontological-monism move with mathematics rather than relational structure as the fundament; it is engaged at §4.3.7. The information-theoretic ontology that treats reality as fundamentally informational is engaged at §4.1.5.

What converges. The present account converges with each of three forms of OSR (eliminativist, moderate, semirealist) at increasing levels of ontological commitment. Eliminativist OSR rejects intrinsic-nature-bearing objects, which the present account does not require. Moderate OSR retains nodes as structurally exhausted, which the present account already affirms. Semirealist OSR retains dispositional natures, which the present account is consistent with without requiring. The disagreement, where there is one, is about whether the trio T, S, Φ is necessary at a level prior to the relational fabric current physics describes, or merely one description among others of that fabric. The case for the prior-necessity reading is given at §§2.6–2.7, with §2.6.7 (Triconditional Cross-Entailment) providing the physics-facing demonstration that any nomological structure requires T, S, and Φ as co-instantiated, and OSR theorists who reject it owe a counter-argument, and what the literature has produced (group-theoretic symmetries, gauge invariance, permutation structure) does not yet supply one. What it supplies is rich confirmation that the relational fabric, however described, includes the structural roles.

Structural realism was proposed to reconcile the no-miracles argument for scientific realism with the pessimistic meta-induction from theory change. Two decades after Worrall, Ladyman and Ross gave the position its fullest ontic form. Two more decades later, French and Esfeld and Chakravartty have populated the space between epistemic and eliminativist with positions that disagree about objects but converge on structure. The present argument takes the convergence seriously. Whatever the relational fabric of fundamental physics turns out to be, T, S, and Φ are the conditions on its being a relational fabric at all.

4.3.10 Bundle and Trope Theory

Bundle theory, developed in different forms by Hume, Russell, and more recently by Paul (2017), holds that objects are bundles of properties without underlying substrates. Trope theory, developed by Williams (1953), Campbell (1990), and Maurin (2002), holds that properties are particularized (this redness here, that hardness there) rather than universals shared across instances. Both positions challenge the substance role of Φ (causal persistence across temporal change, the third structural Condition alongside T and S, borne by Substance just as T and S are borne by Time and Space): if there are no underlying substrates, what does causal persistence inhere in? If properties are particularized tropes, what bundles them into the persisting configuration the structural account requires?

The reply distinguishes the metaphysics of property-bearing from the structural role of Φ. The substance role is the role of causal persistence across temporal change: whatever obtains in such a way that the same configuration is causally connected from one moment to the next. Bundle theory and trope theory are accounts of how property-bearing works at a fine grain; they are not denials that something persists. A bundle that persists across time is, for the present argument, an instance of Φ. A collection of compresent tropes (tropes co-occurring at the same location) that maintains its compresence across temporal succession is, for the present argument, an instance of Φ. The present argument is not committed to the substrate metaphysics that bundle and trope theories deny; it is committed to the structural role those metaphysics still have to fill, however they fill it.

Two finer questions the bundle and trope theorist press do not change this. The first is Bradley’s old regress: if the tropes are bound into one bundle by a relation, and that relation is itself a trope, it requires its own binder, and the bundle never closes. The regress is a real pressure on bundle theory’s internal economy. Renaming compresence as S holding across Φ does not by itself solve it; the bundle theorist still owes an account of what unifies the tropes, whether primitive compresence, a non-relational tie, or another device. The structural account can remain neutral on that account while requiring any successful proposal to realize differentiation and continuity. Whatever the bundle theorist settles on (a primitive compresence relation, a non-relational tie, a brute fact of togetherness) the structural roles are what that settlement must instantiate, not a rival to it. The second is whether the bundle endures, wholly present at each moment, or perdures through distinct temporal parts. Φ is neutral between them: it asks only that the configuration be causally connected from one moment to the next, a demand the endurantist meets through identity and the perdurantist through genidentity among temporal parts. The persistence-metaphysics dispute (pressed in its perdurantist form at §4.2.8) settles how Φ is realized, not whether it is.

What is established. The reply matches the reply to ontic structural realism (§4.3.9) in shape. The present argument identifies the roles any configuration must instantiate; metaphysical accounts of how the roles are filled are downstream of the necessity claim. Bundle theory, trope theory, substance theory, and process theory are competing answers to a question that arises only after §2.6.8 has specified the Φ / Physicality role and §2.7 has shown how that role belongs to the Conditions-as-Nature structure.

4.3.11 Mereological Nihilism

Van Inwagen’s Material Beings (1990) defends the most precisely formulated version of mereological nihilism the analytic tradition has produced. There are no chairs, no tables, no planets, no stones: only simples arranged chair-wise, table-wise, planet-wise, stone-wise. The single exception is living organisms, whose self-maintaining causal activity Van Inwagen takes to ground a unity mere arrangements lack. The position shrinks the ontology where Lewis’s plurality (§4.4.1) multiplies it, and it does so with the same systematic care. Sider’s “Against Parthood” (2013) presses one step further: even the organism exception is rejected, and what remains is only simples in arrangements.

Mereological nihilism (also called compositional nihilism or compositional eliminativism) has a precise lineage. Classical mereology (Leśniewski 1916; Tarski 1929) supplies the formal vocabulary of parthood, and Lewis’s Parts of Classes (1991) systematizes the contemporary defense of unrestricted composition the nihilist most directly opposes. Van Inwagen (Material Beings, 1990) develops the organism-exception version with the most systematic care. Merricks’s Objects and Persons (2001) extends Van Inwagen’s exception to persons and presses the overdetermination argument against composite objects more broadly. Sider’s “Against Parthood” (2013, Oxford Studies in Metaphysics vol. 8) develops the strictest nihilist version: no composites at any level, organisms included. Dorr’s “What We Disagree About When We Disagree About Ontology” (2005) defends compositional nihilism through paraphrase: nihilist and non-nihilist sentences can be made truth-conditionally equivalent. Cameron’s “Truthmakers and Ontological Commitment” (2008) develops a kindred deflationary line. Against the nihilists: Markosian’s “Brutal Composition” (1998) treats composition as a brute fact admitting no further criterion; Schaffer’s “On What Grounds What” (2009) and the monist literature (Schaffer 2010) defend the cosmos as the fundamental composite from which parts derive; classical mereology (Goodman 1951; Simons’s Parts, 1987) defends unrestricted composition. The structural account sits closest to the moderate-OSR line of §4.3.9 in being neutral on composition while specific about what any arrangement must instantiate.

The nihilist pressure bears on Φ specifically and on the §2.7 architecture more generally. If composite objects do not exist, then the persistent configurations the argument has been describing (patterns of features held together as one substantial bearer across temporal succession) have no bearer to inhere in. T and S might survive as relational facts among simples, but Φ, the persistence of structured pattern, would require a pattern that is something over and above its constituents. The nihilist position, in its strict form, dissolves the Substance role into arrangements of underlying simples and denies that the arrangement composes anything further.

A clarification before the moves. The nihilist is not denying that the world contains the arrangements ordinarily called “chairs” and “tables.” Van Inwagen and Sider both grant the arrangement-facts; what they deny is that the arrangements compose further objects. “Simples arranged chair-wise” describes a real configuration on every nihilist account; what it does not describe, on the strict version, is a chair as an additional ontological item. The question turns on whether the structural argument requires composite objects beyond the arrangements, or whether the arrangements themselves carry the structural roles. Three forms of the move press that question at three resolutions: the basic eliminativist version, Van Inwagen’s organism-exception version, and Sider’s strict version. Each reads the structural argument as requiring composite objects and replies by denying them. Each misreads what the argument requires.

First. The basic eliminativist move denies composition wholesale. There are no composite objects (chairs, tables, planets, organisms), only simples in various arrangements. The structural argument, in describing the persistence of configurations across time, has been describing composites the nihilist rules out. Without composites, no Φ; without Φ, the Triconditional fails.

The argument does not need composites in the sense the nihilist denies. What it needs is that the arrangements themselves instantiate the structural roles. A pattern of simples arranged chair-wise that persists across a temporal interval instantiates Φ at the level of the arrangement, whether or not the arrangement composes a further object called “the chair.” Persistence is a feature of the pattern, not a posit beyond it. The nihilist who grants that the arrangement obtains, and obtains across time, has granted the structural role; what they have refused is the additional ontological commitment to a composite as a further item. The argument is neutral on that further item. T, S, and Φ specify what arrangements consist in, not what arrangements compose. Compositional nihilism leaves the structural roles where they are.

Second. Van Inwagen’s organism exception holds that living things compose but artifacts and inanimate aggregates do not. The criterion is unity: the activities of an organism’s parts are coordinated in service of the organism’s life in a way that the parts of a chair are not coordinated in service of the chair. Living things therefore qualify as composites; chairs and tables do not. The structural argument, on this reading, gets purchase only where Van Inwagen’s exception holds.

The criterion Van Inwagen invokes is itself Φ at the appropriate resolution. Causal persistence of a configuration across temporal change, where the configuration’s persistence is constitutively maintained by its internal causal activity: that is what Substance names in the structural argument, and it is what Van Inwagen identifies as the mark of organic unity. The argument generalizes the criterion rather than restricting it. Any arrangement whose obtaining is constitutively maintained across time by its own causal activity instantiates Φ; the question is whether the maintenance is real, not whether the arrangement is biological. Crystalline lattices maintain themselves through electromagnetic binding; weather systems maintain themselves through fluid dynamics; the argument applies wherever maintenance obtains. Van Inwagen draws the line where the maintenance is most conspicuous; the argument follows the maintenance wherever it goes.

Third. Sider’s strict version denies even Van Inwagen’s exception. There are no composites at any level, organisms included; only simples in various arrangements, including the arrangements ordinarily described as living. The arrangement-facts obtain (Sider does not deny that simples are arranged organism-wise), but no further composite organism exists. The structural argument, on this version, must run entirely on arrangement-facts, without the resource of any composite bearer.

The recasting is the right one for the strict nihilist, and the reply is correspondingly cleaner. Arrangement-facts that obtain across time, with patterns maintained by causal activity among their constituents, instantiate the structural roles whether or not they compose further objects. The argument operates at the level of what arrangements consist in. Sider’s strict version concedes the arrangement-facts; it denies the further composite. T, S, and Φ are specifications of what arrangement-facts must be like to be the kind of facts they are: temporally persistent, relationally distinct, causally maintained. The argument’s necessity claim runs at the arrangement-fact level. Whether the additional composite is also there is a separate ontological question the structural argument leaves open. Sider’s nihilism, taken at its strictest, is compatible with the structural argument: it constrains the inventory of composites without contesting the structure of arrangements. Sider’s nihilism is only one of two commitments he brings to bear here; the other (his joint-carving structuralism, on which the live question is whether T, S, and Φ belong to the fundamental ideology rather than sit above a deeper structure) is the subject of §4.4.7, where that question is engaged on its own terms.

A kindred move runs through priority monism (Schaffer 2010). Where the nihilist denies composites by eliminating them, the monist denies composites as fundamental by absorbing them into the cosmos: only the whole cosmos fundamentally exists, and the apparent composite objects are derivative aspects of it. The reply runs in the same direction. Whether the cosmos is fundamental and its parts derivative, or the simples are fundamental and the apparent composites eliminated, the structural argument operates at the level of the arrangements that obtain (at whatever ontological level the arrangements are located) and on the structural conditions any obtaining arrangement must instantiate. The composition question is orthogonal to the structural-role question. The same applies to the bundle/trope alternatives canvassed in §4.3.10: those answers, like nihilism, address what the inventory contains rather than what the inventory’s items must be like to obtain.

What converges. Each of the three forms misreads what the argument requires. The basic eliminativist move denies composites the argument does not need; the structural roles attach to the arrangements, not to the composites the arrangements may or may not constitute. Van Inwagen’s organism-exception invokes a unity criterion that just is Φ at the level of organic maintenance; the argument generalizes the criterion to any arrangement constitutively maintained by causal activity. Sider’s strict version concedes the arrangement-facts while denying the composites, and the arrangement-facts are exactly where the structural argument operates. Nihilism, at every resolution, leaves the structural roles where they are.

One clarification of scope is owed, because Van Inwagen’s signature pressure is not compositional but modal-epistemic. His deeper question (how one knows that T, S, and Φ are necessary rather than merely pervasive, given how far that claim sits from the conceptual and near-by possibilities the modal faculty handles reliably) is not answered by the compositional reply above, and is not meant to be. That pressure is upstream of this section, and the book’s answer to it is methodological rather than intuitive: the necessity verdict is claimed as the output of the derivation of §§2.6–2.7, not of a modal intuition about distant scenarios, which is precisely the non-intuitive line Van Inwagen’s own modal skepticism leaves open. This section discharges the compositional half of the Van Inwagen inheritance; the modal-epistemic half is engaged on its own terms at §4.1.13, and paid where the necessity is derived, not where composition is discussed.

What Van Inwagen saw clearly, and Sider sharpened, was that the modern compositional tradition had multiplied entities far past anything the world’s structure demanded. The classical-mereology pedigree (from Leśniewski’s plenitudinous parthood through Lewis’s Parts of Classes to contemporary unrestricted-composition defenses) produced an inventory the nihilist was right to refuse on grounds of austerity. What the nihilist tradition has not always seen as clearly is that the world’s structure does not consist in that inventory either. It consists in what arrangements (composite or not) must be like to obtain. The composition debate is real; it runs on a level the structural argument leaves open. Wherever simples are arranged in patterns that obtain across time, are distinct from other patterns, and are maintained by causal activity, T, S, and Φ are there. The argument is for the structure of what obtains, not for the inventory of what composes.

4.3.12 Haecceitism

Haecceitism holds that the facts about which individual is which are not exhausted by the qualitative and relational facts: two situations can agree on every qualitative and relational property and still differ over which individual is which. Adams’s Primitive Thisness and Primitive Identity (1979) gives the position its canonical analytic statement: a thisness is the non-qualitative property of being this very individual, grounding numerical identity prior to and independently of any qualitative or relational characterization. The motivating case is older: Black’s “The Identity of Indiscernibles” (1952) describes a symmetric universe of two qualitatively indiscernible iron spheres, two and not one, with nothing qualitative to tell them apart. Lewis’s On the Plurality of Worlds (1986) is the locus of the contemporary debate, rejecting non-qualitative differences between worlds in favor of a counterpart-theoretic treatment, on which an individual figures in other possible situations not literally but through its qualitatively closest counterparts there, so that qualitative character fixes all there is to fix. The subsequent literature (Skow 2008) separates a deflationary form, which grants haecceitistic possibilities without positing primitive thisnesses, from the robust Adamsian form, which grounds them in genuine non-qualitative properties. What unites the family is the refusal to reduce individuation to qualitative or relational structure.

The position bears on the spatial Condition directly. Positional distinctness is grounded in relational extension at §2.4.7, which states the load-bearing Constitutive Premise: genuine non-identity at the positional level just is relational non-identity, distinctness constituted by a traversable interval, not by intrinsic difference. Haecceitism is the most direct denial available. If a primitive thisness grounds the distinctness of two items independently of any relation between them, then distinctness need not be relational, and the premise on which the derivation of Spatiality turns is false. The relationalist pressure on that same premise, Dasgupta’s Individuals: An Essay in Revisionary Metaphysics (2009), which §2.4.7 already engages, pushes from below, holding individuals to be less fundamental than structure; haecceitism presses from above, holding thisnesses to be more fundamental than any structure they enter into.

Grant the position at full strength. Haecceitism may well be the correct account of individuation as such. Black’s symmetric universe is a real difficulty for any view that grounds all numerical distinctness in qualitative difference, and the obvious relational repair (distinguishing the spheres by their distances from things) cannot break a symmetry in which every relational description one sphere satisfies is satisfied by the other. The concession should be larger still. The consideration sometimes pressed against the haecceitist (that even a primitive thisness must be picked out as this rather than that, and so already trades on a contrast) does not by itself settle the matter, because how a thisness must be referred to is a fact about representation, not about what the thisness is. A haecceitist who locates the individuating work in the order of being while assigning the contrast to the order of description has not thereby been refuted. The reply does not run through that bridge.

What the structural argument requires is narrower than the thesis the haecceitist denies. The derivation does not need every numerical distinctness to be relational; it needs the distinctness that change crosses to be relational. Spatiality is the Condition movement presupposes, and movement is displacement across an interval: a crossing. The distinctness at issue in §2.4.7 is therefore positional distinctness specifically: the non-identity a displacement traverses. Primitive thisness, whatever else it grounds, does not by itself supply an interval to cross. A thisness delivers a label (this individual, not that one), and a label is not a distance. Two items distinct in thisness alone, with no relational extension between them, are non-identical in just the way red is non-identical to blue: distinct, with nothing between them that anything could move across. That is exactly the line §2.4.7 draws between a difference of contents or names and a distinctness of positions. The haecceitist’s primitive grounds that the items are two; it does not make their twoness positional.

The flagship case confirms the premise rather than threatening it. Black’s two spheres are positionally distinct, and they are positionally distinct because a relational interval separates them: a distance something could cross, the very feature the Constitutive Premise names. What the case withholds is a qualitative ground of distinctness; it does not withhold the interval. The distance in Black’s original case suffices for positional distinctness while leaving the spheres qualitatively indiscernible. It does not prove that distance is necessary for numerical plurality: a robust haecceitist may posit co-located indiscernibles. Black’s universe is therefore compatible with relational individuation of position, but it does not by itself refute primitive thisness. The haecceitist is right that the distinctness is not qualitative, and right that no description breaks the symmetry; what follows is that the distinctness is relational-extensional, which is what Spatiality is.

A defender might press the harder case: two indiscernibles with no interval between them, co-located, or differing only in which is which across possible worlds. If such distinctness is genuine, it is distinctness without a crossable interval, and the premise fails. The case divides, and neither half reaches the spatial claim. Where the two are co-located with no interval between them, there is nothing a displacement could traverse from the one to the other; whatever makes them two, it is not a positional distinctness, and the argument never held that primitive thisness fails to ground individuation of a non-positional kind. Where the distinctness is the trans-world question of which individual is which, it concerns identity across worlds, not the distinctness of positions within Reality, and the spatial claim is about the latter. Neither half exhibits a positional distinctness across which nothing could move: the one thing the Constitutive Premise rules out.

What converges. The robust Adamsian thisness, the deflationary haecceitistic possibility, and Black’s symmetric universe all press one point: that distinctness is not qualitative. The structural argument grants it. What none of them supplies is a positional distinctness without a crossable interval, and that, not the reduction of all individuation to relation, is what the Constitutive Premise asserts. Where a haecceitist case carries an interval, it instantiates the relational extension Spatiality names; where it carries none, it is individuation of another kind, which the argument leaves untouched. Haecceitism constrains what grounds identity; it does not displace what grounds position. Like bundle theory and mereological nihilism (§4.3.10, §4.3.11), it answers a question about the inventory of individuals, not the question of what an interval must be for something to cross it.

4.4 Modal Variants and Extensions

This cluster questions the modal machinery the argument rides on: necessity, possibility, and possible worlds. If the semantics of possibility differed, the canonical lift might not follow. The reply begins with a distinction. The argument’s operator ◇ means real possibility: membership in one fixed global domain W, the complete range of ways Reality permits (§§1.2.1, 2.1). A theory may instead model local reachability, epistemic openness, or world-relative accessibility. Those are legitimate relative modalities, but variation in them does not by itself vary membership in W. The engagements below ask whether a rival has offered a different account of real possibility or only a different relation inside or alongside the same global domain.

The objections of §4.3 press whether the present account leaves something out: experience, intrinsic nature, mathematical applicability, mind itself. The objections of this section press the modal apparatus the argument runs on. If the modal logic is wrong (if necessity, possibility, and counterfactual reasoning do not behave as the argument assumes) then □(T∧S∧Φ) does not follow even if T, S, and Φ (Temporality, Spatiality, and Physicality, the three structural Conditions the book’s argument derives, borne by Time, Space, and Substance) are exhaustive.

The space of moves is well-charted. Lewis (1986) develops concrete modal realism, on which possible worlds are concrete entities of the same kind as the actual world; the actual world is just the world we inhabit. Williamson (2013) develops necessitism, on which everything that exists exists necessarily and the only contingency is qualitative rather than ontological. Counterpart theory (Lewis 1968) replaces transworld identity with similarity relations and changes what “the same necessary being” can mean across worlds. The Barcan formulae (named for Ruth Barcan Marcus’s 1946 derivations linking quantifiers and modal operators) press whether quantification and modality commute, with consequences for whether necessary existents must exist in every possible world. Stalnaker (1976) presses the moderate realist line, on which possible worlds are abstract maximal properties or ways the world could have been. Kripke (1980) gives the standard picture against which the alternatives press their cases. Plantinga (1974b) develops a robust modal realism for theological purposes. The literature is dense, technical, and inconclusive on most of the central questions.

The argument’s strategy across the section is to distinguish three questions that modal theories often run together: the global domain of real possibility, the semantics used to represent it, and the ontology assigned to worlds or domains. The canonical proof depends on the first, not on Lewisian concreteness, transworld identity, Barcan formulae, or a particular local accessibility relation. Structural roles enter only where a candidate world or ground is offered as a realizer of genuine change. A thin abstract representation need not instantiate temporal succession, spatial extension, or causal persistence; a theory that offers it as physical reality, a ground of experience, or a complete change-realizer inherits the richer burden. The sections below therefore test each framework at the resolution it claims.

4.4.1 Modal Realism

David Lewis began with counterpart theory in the 1960s, added the analysis of counterfactuals in the early 1970s, and over the following decades built the most extravagant ontology any analytic philosopher has ever defended. When you say “I could have taken the other job,” Lewis holds, you are saying truly that somewhere, in a universe spatiotemporally isolated from this one, a person who resembles you in the relevant respects took that job. The universe is concrete: just as concrete as this one. It contains stars and tables and chairs. You will never causally interact with it. But it exists, full stop, in exactly the sense this one does. Modal language is not abstract after all. It is a report on a plurality, and the plurality is the ontology.

Lewis developed Modal Realism across two decades, in Counterfactuals (1973) and On the Plurality of Worlds (1986), one of the most systematic books of late analytic metaphysics. The position remains his alone: its orbit has structured a generation of work on modality without persuading others to adopt the concrete plurality unchanged. Divers (2002) and Nolan (2002) develop Lewisian variants with modified commitments. Bricker (2020) gives the standard contemporary taxonomy of how Lewisian worlds individuate. Against the concrete variant stand the abstract or ersatz realists: Plantinga (1974b) defends worlds as maximal states of affairs; Stalnaker (1976, 1984) treats worlds as maximal properties; Adams (1974, 1981) as maximal consistent sets of propositions; Kripke (1980) operates with primitively stipulated worlds rather than discoverable ones. Van Inwagen (1986) presses the central internal critique: the plurality is ontologically exorbitant without being explanatorily adequate. Rosen (1990) develops modal fictionalism as an explicit alternative: the plurality as useful fiction, not reality. Armstrong (1989) pushes combinatorialism: possibilities are recombinations of actual constituents, no alternative universes required.

The present reply sits closest to Armstrong’s combinatorialism in rejecting the concrete plurality, and adds what combinatorialism alone does not specify: which structural roles any recombination must preserve. Why they cannot be recombined apart is supplied not here but by the cross-entailment among the three roles (§2.4.8, §2.6): they are not distinct existences for recombination to separate.

Modal Realism presses directly against the necessity claim of the argument. What §2.8.5 establishes is □(T∧S∧Φ): in every world within the fixed global domain of real possibility, the structural Conditions obtain. S5 is earned from the fixed-domain reading of real possibility (§2.1), not assumed as an axiom. Transition Closure is a narrower result about the onset of Change under structural accessibility; it is not an independent derivation of proposition-wise S5. If possible worlds are concrete universes, the question is what every concrete universe in that global domain contains. Lewis’s plurality is the most permissive ontology of possibility ever defended: if any consistent-sounding description individuates a concrete world, worlds without structure might be among them. The argument’s necessity claim is false if even one concrete world lacks T, S, or Φ. (ET) (the Exhaustion-of-Existence claim, that T, S, and Φ exhaust the forms of real existence) is false on the same condition.

A clarification before the moves. The dispute is not with Lewis’s counterfactual semantics or with the usefulness of possible-world talk. The question is whether the concreteness criterion Lewis himself requires (the feature that distinguishes Lewisian realism from ersatz alternatives) permits radically sparse worlds to qualify as concrete. The move comes at several resolutions: the bare sparse-world move and its sharpened recombination form, the isolation-thesis variant, and the primitivist fallback. The space they canvass is the space Lewisian realism can occupy. The replies argue that each collapses concreteness back into structure before the plurality is assembled.

First. The sparse-world move holds that among the plurality there is a world w′ containing a single unchanging state: no succession, no spatial extension, no causal activity. Such a world is conceivable, and the move needs conceivability of a coherent description to entail concrete existence of the corresponding world. w′ contains no T, no S, no Φ, and it concretely exists.

For Lewis, worlds are maximal sums of spatiotemporally interrelated things, with worldmates unified by the relevant natural external relations. This gives Lewisian worlds a spatiotemporal criterion where such relations obtain, but it does not by itself prove that every world contains both temporal succession and spatial extension. A sparse or changeless world remains a live test of the book’s independent type-level arguments. The present argument need not claim that every abstract representation of a world has those types; it claims that a concrete world proposed as a real change-realizer cannot lack them. A world with no temporal succession has no internal change that could individuate one world-state from another, which means no structural fact distinguishes “world w′” from “no world.” A world with no causal persistence has no bearer across which content could be sustained. A world lacking all three is not a sparsely furnished concrete world. It is not a concrete world at all. Lewis’s own concreteness criterion forecloses the sparse-world move before the plurality is even assembled.

The sharpened form: recombination. The sparse-world move has a sharper version that does not empty the world at all: Lewis’s principle of recombination. The principle permits extensive recombination of distinct existences, subject to its own constraints; it does not license every arbitrary description as a world. Recombination does not need the empty world; it needs only a selective deletion of what the Humean treats as distinct. Take a spatiotemporally extended world and deny necessary connections between its distinct local states; the resulting Humean mosaic retains Time and Space while rejecting a necessary productive connection between distinct existences. The issue is whether the mosaic can retain the structural work the present account assigns to Φ without thereby retaining Φ under another description.

The reply must not answer it by saying the mosaic “has no bearer,” since whether a world requires a persisting bearer is precisely what the Humean denies; to assert it here would beg the question.

The reply turns instead on what the mosaic already grants. A mosaic is a distribution of local states across a spatiotemporal manifold: Lewis (1986b, pp. ix–xvi), introducing Humean supervenience in Philosophical Papers, Volume II, exempts the spatiotemporal relations from it and takes them as primitive, so the manifold is conceded, not constructed. The manifold grants differentiation and spatiotemporal arrangement. It does not follow merely from exclusion or Best-System regularity that a productive cross-interval causal operation exists: the Humean denies exactly that inference. The sharper question is whether the mosaic realizes Φ’s broader bearing or continuity role without accepting productive necessary connections. A regular mosaic may remain a substantive rival at the type level rather than collapsing into ∅.

Second. The isolation thesis presses from a different angle. Lewis’s worlds are spatiotemporally isolated from one another: no causal or spatiotemporal connections between them. A world in the plurality need not be internally rich; it needs only to be isolated from other worlds. A single unchanging state, isolated from all other such states, qualifies as a world by isolation alone.

Lewisian worlds are isolated in the sense that no spatiotemporal relations connect their inhabitants; isolation is not an additional between-world relation requiring a common surrounding manifold. The thesis therefore does not establish any particular world’s internal richness. Whether a sparse candidate qualifies as a world turns on Lewis’s worldmate criterion and the candidate’s contents, not on reifying non-connection as a structural relation. The isolation thesis is a structural criterion applied to structured worlds, not an alternative to structure.

Third. The primitivist version goes boldest. Being a world is a primitive fact, irreducible to any further ontology. Worldhood is fixed by the modal role ‘world’ plays: as the primitive truthmaker for modal claims. This is a coherent rival semantic proposal, but it carries a cost: it declines a constitutive account of what makes a world concrete or a genuine realizer of change.

The present argument cannot close primitivism by retorsion alone. At most, it can show that any primitive world introduced as a concrete universe or as a realizer of change inherits the relevant role- and type-level burdens. A primitivist who treats worldhood as primitive while refusing those burdens has replaced structural explanation with an additional primitive. That is a theoretical cost, not a contradiction.

A further Lewisian pressure targets the Modal Identity directly rather than the concreteness of worlds. Lewis’s counterpart theory denies strict trans-world identity. What you would have been had you taken the other job is your counterpart in another world: qualitatively similar, numerically distinct. The Modal Identity of §2.6.6 (□(◇C ↔︎ (T∧S∧Φ)), where ◇C is the possibility of change) appears, under counterpart semantics, ambiguous between a claim about the same T, S, and Φ across worlds (requiring trans-world identity Lewis denies) and a claim about qualitatively similar structural roles in each world (a weaker claim counterpart theory permits but insufficient to ground the necessity the argument needs). Either the Modal Identity assumes identity Lewis rejects, or it retreats to similarity that cannot carry the argumentative weight.

The dilemma dissolves at what the Modal Identity of §2.6.6 actually asserts. The Modal Identity does not claim a numerically identical token of T, S, or Φ recurs across worlds in the way ordinary individuals would have to for strict trans-world identity. It claims any world containing change contains the structural roles change requires: whatever plays Time’s role in w and whatever plays Time’s role in w′ are both instances of a single role-type. Role-type identity is not the kind of trans-world identity counterpart theory forecloses. Counterpart theory denies strict identity of individuals because individuals are, on Lewis’s picture, worldbound particulars. Structural roles are not particulars; they are types any particular must instantiate to be the kind of thing it is. Lewis (On the Plurality of Worlds, §1.5) is explicit that properties are classes of their actual and possible instances: an across-world specification counterpart theory permits for types even as it forbids it for tokens. A structural role meets that specification directly: “whatever plays Time’s role” picks out the class of its actual and possible role-players, which is just what a property is on Lewis’s account. The role-type is therefore not a trans-world individual smuggled past counterpart theory but an abundant property in Lewis’s own sense: identified, as every Lewisian property is, by its actual and possible instances across the plurality. The Modal Identity operates at the level Lewis’s own semantics already allows, and the structural roles fit the property-class regimentation rather than being exempted from it. A second counterpart-theoretic pressure, aimed at the transfer principle rather than at the Modal Identity, runs through the chained-origin cases of Chisholm, Chandler, and N. Salmon; it is engaged at §4.4.11.

A kindred move drops Lewis’s concrete realism for Stalnaker’s (1976) moderate realism: possible worlds are abstract maximal ways the world could be, not concrete alternative universes. The abstract-realist variant is closer to the Platonist challenge of §4.2.4 and is answered there. The reply in the modal-realist vocabulary: abstract maximal ways can be distinguished and compared, directly supporting only a thin difference-role. They need not thereby instantiate temporal succession, spatial extension, or causal persistence. The richer roles apply when a way represents or realizes a concrete change-world, not merely because it is an abstract modal surrogate.

What converges. Each form collapses concreteness back into structure before the plurality is assembled. The sparse-world move founders on Lewis’s own concreteness criterion: a world with no spatiotemporal structure is not a sparsely furnished concrete world but no concrete world at all. Its sharpened recombination form is not closed by definition: the mosaic’s conceded manifold already grants differentiation and arrangement, and whether Best-System regularity can discharge Φ’s bearing role without productive connections is the type-level question the mosaic remains answerable to. The isolation thesis is itself a structural relation: isolation is a spatiotemporal concept, and worlds can only satisfy the isolation criterion if they have spatiotemporal structure to be isolated with. Primitivism about worldhood does not escape the structural account so much as decline it: the refusal replaces structural explanation with an additional primitive, a cost the position carries rather than answers.

A stronger reply: the unity of the frame. The replies so far grant Lewis his plurality and show only that each member of it must instantiate the structure; a bolder reply presses the count itself, and it stands apart from the rest because it reaches further and leans on a premise Lewis is entitled to refuse. A plurality is a plurality only if its members are genuinely many, and genuine two-ness is not bare: for two concrete totalities to be two rather than one totality described twice, there must be a real interval in which their non-identity registers — an interval something could cross (§2.4.7, the Constitutive Premise; §2.4.1, its world-level form). Lewisian worlds are by definition isolated, with no spatiotemporal relation and nothing crossable between them, so the distinctness they require is exactly the kind §2.4.7 declines to grant: a difference with no interval, which is either relational after all — a shared structure, hence one connected order, and isolation is lost — or a bare thisness, “a label without an interval,” which does not make two. Isolation is the absence of a separator, not a mode of separation; “two isolated concrete worlds” asks the cardinal predicate to grip where nothing gives it purchase. The same pressure shows in the very act of pointing: to single out a world as that one is to set it off from this one, and this-and-that is here-and-there in other clothes. “That world” says “not here,” and “not here” asks after a where — either a location standing off from this one, which is a relation and undoes the isolation, or no where at all, in which case “that” has singled out nothing. The plurality cannot be so much as indicated without borrowing the frame it was meant to exceed.

A deeper tightening follows. The primitivist might concede the demonstrative point — that “that world” borrows the here-and-there it was meant to outrun — and insist instead that bare thisness simply is, at the level of being, prior to any pointing or description. The question that move cannot answer is what it is. A world must be a something: it must carry enough determinate content to be a world rather than its own absence. Bare thisness strips every qualitative, relational, and structural feature by design — that is what “bare” means. What remains when all content is gone is ∅, not a minimally furnished concrete world but the null case. “Bare thisness exists” and “nothing exists” are, on every dimension that could distinguish them, the same description. A posit indistinguishable from ∅ has not located bedrock below structure; it has dissolved into the null case — which is the absence of a world, not an instance of one.

Stated precisely, what this reaches is not the numeral “exactly one world” but the unity of the frame. Reality is not the winner of a headcount among possible many; it is a unity-of-structure — the one frame any configuration must instantiate, beyond number rather than one of a kind (§6.2.2). “One,” said of Nature, is not a cardinality but that unity: positions within the one logic of here-and-there can be multiplied without end, but not the logic of here-and-there itself, which is no particular admitting duplication but the frame within which any “two” is sayable. On this reading the plurality is not outweighed but mis-formed — a cardinal count run at the one level where the cardinal predicate finds no grip.

The candor the rest of the book keeps is owed here too. This is not a deductive refutation of a determined primitivist. Lewis can plant precisely on the fault line §2.4.7 marks: he can hold the numerical distinctness of concrete totalities to be primitive — needing no interval and no ground — and the reductio does not move him, because its load-bearing premise is the one he denies. What the reply secures against him is not formal contradiction but a convergent exposure: bare thisness dissolves into ∅, the absence of a world rather than an instance of one, and the cost of maintaining the posit against that dissolution is not confined to the cross-world case: the same bare thisness he plants on can no more make one being one across a change, absent continuous carrying, than it can make two worlds two absent an interval (§2.4.1). Nor can he discount that cost by making the distinctness merely numerical — the bare non-identity of two from three — for that is how abstract objects are told apart, and abstractness is the one price Lewis cannot pay: concreteness is what separates his realism from the ersatz alternatives he wrote against. He wants the numeral’s economy and the world’s concreteness together — a two with a world hung upon it — and they do not combine, since the entire content of concreteness is the here-and-there that a bare numeral forgoes (§2.4.1; §4.2.4). His intervalless plurality is a positive existential posit with no access and no causal trace, carried for a modal-semantic payoff the structural account already supplies from a ground that is independently necessary (§2.8.5). And the terrain the unity reading stands on is not an ad hoc anti-Lewis axiom but the oldest treatment of the one and the many — Plotinus’s source of number that is itself beyond number, mathematical structuralism, Spinoza’s single substance. The plurality remains sayable; what it has lost is its cheapness: distinctness-without-relation is the least defensible of the position’s commitments, and the unity of the frame is what one reaches the moment that commitment is refused.

The cost ledger carries one more line, and it is quantitative. Grant the primitivist his bare distinctness, and the plurality still owes an answer no positive existential posit gets to waive: how much of it is there? Lewis’s own machinery cannot settle the question. Forrest and Armstrong (1984) showed that recombination, applied without restriction to the totality of worlds, generates a world containing duplicates of all the worlds — a world bigger than any world, itself included — so neither the aggregate nor the set of all worlds survives the principle’s own closure. Lewis’s reply (On the Plurality of Worlds, §2.2) restricts recombination by the proviso the principle already carries in its statement (§1.8 of the same work), “size and shape permitting,” a qualification whose content he leaves unspecified; Nolan (1996) argues the restriction is unmotivated on Lewis’s own terms, and the alternative he defends, recombination unbound with more possibilia than any set can hold, concedes the point at issue here: the plurality has no determinate size. Either way the count never closes. A finite tally of concrete totalities would be arbitrary — nothing in the theory could say why that many and no more — and the infinite alternative cannot be stabilized at any cardinality. A posit whose quantity is in principle unanswerable carries exactly the mark the unity reading predicts: number failing at the level of totalities. What looks from inside the theory like a bookkeeping embarrassment is, read structurally, the same finding the interval argument reached from outside — there is no fact of how many, because how many is not a question the frame can answer about itself.

The shape of the mistake can be put more plainly. Each Lewisian world is internally rich — intervals throughout, a full spatiotemporal interior, exactly like the interior of a fiction, which stipulates its own distances freely. What carries no interval is the relation between worlds. That relation is a seam — world-to-world as fiction stands to reality — and every seam we have uncontested access to is interval-free precisely where it is non-concrete: there is no distance from Sherlock Holmes to Hamlet, not an unknown distance but no such fact. The photograph in your hand sits a foot away, while the person it pictures, held now as pictured, is at no distance from you at all; the only way to recover a distance is to switch back from the representation to the living worldline — from the seam to a concrete interior. Lewis is thus a concretist about each world and, without owning it, an ersatzer about the many: he imports the interior’s concreteness across a seam that bears none of concreteness’s marks. His plurality is a set of concrete relata related as if they were non-concrete. And the intuition that makes the picture feel forced — that the worlds are obviously many — leans on distinguishability, which is downstream of the structure and confers no actuality of its own: that two worlds can be told apart no more makes them two concrete existents than telling Holmes from Hamlet makes either one a man. The arrow runs from the Conditions to distinguishability, never back.

On the Plurality of Worlds defends concrete realism against what Lewis called “the incredulous stare”: the reaction he expected from any reader encountering the view for the first time, and which he thought no sufficient response to a working modal semantics. The reaction it provoked was not confusion about the view’s coherence but a genuine ontological concern: one Lewis acknowledged and thought the modal semantics outweighed. What Lewis left behind is not the plurality but a diagnostic: any modal ontology powerful enough to do the work has to specify what makes a world a world, and whatever does the specifying instantiates the structure the argument names. Ersatz worlds individuate by property or proposition; concrete worlds individuate by spatiotemporal aggregation. Either specification is a structural fact. What the plurality purchased in modal expressive power it paid for in ontological commitment, and the ontological commitment, examined for what it requires, confirms the structural conditions the argument derives.

4.4.2 Linguistic Ersatzism

Linguistic ersatzism (from German Ersatz, “substitute”), defended by Plantinga (1974b, 1976), Stalnaker (1976), and Adams (1974), holds that possible worlds are not concrete spacetime manifolds, à la Lewis’s Modal Realism (on which every possible world is an equally concrete, equally real spatiotemporal manifold), but maximally consistent sets of propositions or world-state-descriptions. The actual world is the unique such set whose propositions are all true. Other worlds are abstract objects: surrogates rather than rivals. The position retains the modal apparatus while avoiding Lewis’s ontological extravagance.

The label covers three importantly different positions, and the reply that follows must hold across all of them. Adams (“Theories of Actuality,” 1974) is the version the “maximally consistent set of propositions” gloss fits most exactly: worlds are world-stories, maximal consistent sets of propositions, and the actual world is the set all of whose members are true. Stalnaker (“Possible Worlds,” 1976) is a moderate realism on which worlds are not sets of propositions but abstract ways the world could have been (maximal properties it might have had), and his later Mere Possibilities (2012) refines this into a contingentist account on which there are no robustly existing merely-possible individuals to begin with. Plantinga (1974b, 1976) is the most ontologically robust of the three, and the only theistically committed one: worlds are maximal possible states of affairs, abstract objects that exist necessarily and, on Plantinga’s wider view, have their being in the divine intellect. The differences are real and matter for the metaphysics of abstracta, but they are orthogonal to the point pressed here: propositions, ways, and states of affairs alike are distinguishable, content-bearing items standing in consistency relations, and the reply turns only on what any of them requires to be distinguishable and consistent at all.

The argument’s relation to ersatzism is broadly friendly. Ersatzism does not require Modal Realism’s commitment to a plenitude of concrete worlds; it does not commit the user of modal vocabulary to the existence of donkeys that talk in some other concrete reality. Ersatzism’s worlds are abstract objects, and the question of what abstract objects depend on is a question for the metaphysics of abstracta, addressed in §4.2.4 (Platonism) and adjacent.

Ersatz worlds contain distinguishable propositions, ways, or states of affairs standing in logical relations. This directly secures thin differentiation at the level of the representation. It does not make an abstract surrogate temporally successive, spatially extended, or causally persistent. Those richer burdens belong to the concrete world represented, or to any surrogate proposed as its physical realizer.

A second observation. On Adams’s world-story version, actuality privileges the maximal consistent set all of whose members are true. Stalnaker’s ways and Plantinga’s maximal states of affairs require their own actuality relations rather than this exact set-theoretic formulation. In each case, truth or obtaining is assessed against how things actually are. How things are is Reality: with its T, S, and Φ. The actual world’s privilege is parasitic on Reality being a particular way. Ersatz worlds are alternative description-sets relative to Reality; they are not ontological rivals to it.

What converges. Ersatzism is a successful theory of what we mean by talk of possible worlds. It is not a theory that displaces the present argument’s necessary structure. The two operate at different levels: ersatzism at the level of modal semantics, the present argument at the level of the structural conditions any modal semantics describes.

4.4.3 Combinatorialism

Armstrong (1989, 1997) defends a combinatorial theory of modality. Possibility, on this view, is the combinatorial space generated by recombining the actual world’s fundamental constituents, particulars and universals (the concrete individuals and the repeatable properties they instantiate, in Armstrong’s technical sense), in all possible ways consistent with their adicity (the number of argument-places a universal takes) and type. Possible worlds are not concrete (Lewis) or abstract (ersatzism); they are constructions out of actual elements. The space of possibility is wholly determined by what actually exists.

The combinatorialist’s modality is grounded in the actual. The argument here is consonant with this in its commitment to the actuality of T, S, and Φ. Where the two part ways is on the source of the recombinatorial range. Armstrong takes the actual particulars and universals as primitive and lets recombination do the rest. The present argument identifies T, S, and Φ as necessary features of the actual, which means that recombination cannot generate worlds in which T, S, or Φ fail to obtain, since their failure is not a configuration of the elements but the absence of the conditions for any configuration.

This is a constraint on combinatorialism, not a refutation of it. Combinatorialism is consistent with the argument provided it acknowledges that the recombinations operate within the structural conditions, and that acknowledgement can be extracted from the combinatorial operation itself, without first granting the structural argument’s conclusions. Combinatorial generation requires distinguishable constituents and admissible arrangements, so it directly presupposes a thin difference-role. But the modal construction is not itself a temporal process: possible combinations need not be produced one after another, and the constituents need not persist through an act of recombining. Temporal succession and causal bearing must be established within any concrete change-world represented by a combination, not inferred from the theorist’s operation on the base. The structural conditions are therefore not items in the actualist base awaiting recombination: the difference-role is presupposed by any combinatorial generation, and the richer roles fall due wherever a combination is offered as a concrete change-world rather than a description.

A second observation. Combinatorialism faces a familiar pressure: it cannot easily generate alien properties, properties that no actual entity instantiates. If recombination is restricted to the actual, the modal range is narrower than most modal intuitions allow. Various extensions have been proposed (Armstrong himself, then Sider 2003): typically by admitting alien properties as primitive possibilities not generated by recombination, or by allowing modal primitives beyond the actualist base. The combinatorialist will call these workable refinements, but they are not free, and naming the cost precisely is what the objection requires. The appeal of combinatorialism was exactly its strict actualism (the thesis that only actually existing things exist, with no real merely-possible entities): the promise that the entire space of possibility is fixed by what actually exists, with nothing modal taken as primitive. An alien property admitted as a primitive possibility is, by definition, a possibility not fixed by the actual base, so the extension purchases modal adequacy by conceding the very reduction that made the position attractive. And the resulting view is no longer the actualist economy it advertised but a base of actual elements supplemented by free-floating modal primitives. The dilemma is stable: stay strictly actualist and the modal range is too narrow to capture ordinary modal intuition; widen it and the actualist motivation lapses.

Either horn is consistent with the present point, because T, S, and Φ are not candidate alien properties at all: not properties an entity might or might not instantiate, but conditions on there being entities, properties, or recombinations in the first place. The present argument does not turn on the resolution of this debate. What it offers is a fixed point: whatever combinatorialism includes or excludes, T, S, and Φ are not in the recombinable set. They are what the actual base must already instantiate for there to be anything to recombine.

What converges. Combinatorialism is, like ersatzism, a theory of how modal vocabulary is grounded. The present argument is a theory of what modal vocabulary describes. The two are compatible if the former locates its grounds within the structural conditions the latter identifies.

4.4.4 Modal Fictionalism

Gideon Rosen’s “Modal Fictionalism” (1990), a short paper in Mind, became one of the most discussed positions in the philosophy of modality by proposing a radical middle path: accept the expressive power of Lewis’s plurality, reject its ontological commitment. When you say “I could have taken the other job,” Rosen holds, you are saying truly that according to the fiction of possible worlds, some world contains a counterpart of you who took that job. The fiction does its work without the worlds needing to exist. The plurality becomes a story we tell about modality, not a plurality we discover. Lewis built a universe of universes; Rosen borrowed it, used it, and handed it back.

Rosen’s paper set the terms for two decades of debate. Brock (“Modal Fictionalism: A Response to Rosen,” 1993) pressed the first major objection: fictionalism threatens to reduce modal claims to claims about what the fiction says, which presupposes a non-fictional fact about the fiction. Divers (Possible Worlds, 2002) and Nolan (Topics in the Philosophy of Possible Worlds, 2002) canvassed Lewisian and fictionalist variants side by side. Hale (“Modal Fictionalism: A Simple Dilemma,” 1995) and Noonan (1994) sharpened Brock’s dilemma. Sider (Writing the Book of the World, 2011) moved the tradition toward pragmatic deflation: modal talk as a useful tool whose commitments do not penetrate metaphysics. Woodward (“Is Modal Fictionalism Artificial?,” 2011) defended Rosen’s account against the charge that the fiction is hopelessly artificial; Rosen himself later partially distanced himself from the most extreme version of the view. What the literature preserves is a test case: how far can modal talk be deflated before the deflation undermines the level at which talk itself happens?

Modal fictionalism presses directly against the necessity claim of the argument. The current proof chain separates the relevant work. §2.1 supplies Necessary Possibility / S5, §2.6 identifies the structural roles required by change, §2.6.6 gives the Modal Identity, and §2.8.5 declares the Necessity of Nature: □(T∧S∧Φ), the necessary obtaining of the three structural Conditions borne by Time, Space, and Substance. On the fictionalist picture, that claim is true only insofar as the fiction of possible worlds licenses it. But the fiction is a story, not a feature of reality. If the argument’s necessity claim is a claim about a story, its ontological weight dissolves. The argument has proved something about the fiction we use to talk about modality, not something about reality. The structural-exhaustion claim suffers in parallel: if exhaustion is fictional, no actual alternative was ever excluded.

A clarification before the moves. The dispute is not with the usefulness of fictionalist semantics for counterfactual reasoning. The question is whether the fiction can be told without the structural conditions the fiction is meant to deflate. If telling the story requires T, S, and Φ, the structure is not what the fiction is about. It is what makes the fiction tellable. Three forms of the move press the answer at three resolutions: the bare story-element move, the meta-level refinement, and the no-correspondence move. The space they canvass is the space modal fictionalism can occupy. Each concedes the structural conditions at the level the argument’s claim applies.

First. The story-element move holds that modal claims are true according to the fiction, not true simpliciter (Latin: “absolutely,” without qualification). “Possible world w₀ lacks T, S, and Φ” is a coherent story element; the story is part of the fiction; therefore the fictionalist can say w₀ is possible without committing to its ontology. The argument’s individuation conditions for concrete possible worlds do not apply to story elements.

The relevant world-fiction must be disciplined enough to support the fictionalist’s translation scheme, though fictional discourse can represent contradictions without thereby collapsing. The key question is not whether the act of telling is structured, but whether the fiction licenses the target modal claim under the adopted analysis. The question is whether the story of w₀ (a world lacking T, S, and Φ) can be coherently told. Telling the story requires individuating w₀ as distinct from other story elements (Space), attributing the property “lacks T, S, and Φ” to it (Substance, since attribution is a persistent relation between term and attribute), and unfolding the telling sequentially across moments of consideration as one element follows another (Time). The story about w₀ cannot be told without instantiating the roles w₀ is defined by the absence of. A fiction of a worldless world is a fiction written nowhere.

Second. The meta-level refinement concedes the first reply and still preserves the move. The fictionalist can accept that storytelling instantiates T, S, and Φ and still insist that w₀’s properties are fictional. The story requires structure at the meta-level; it does not require structure within the story. w₀ can lack structure as a story element even if the act of telling the story about w₀ requires structure.

The meta/story distinction blocks any inference from the structure of telling to the structure of the represented world. At most, the actual performance supplies act-side evidence that the actual world is structured. It does not establish that every real possibility instantiates the loaded types. The remaining dispute is therefore about the non-fictional modal resources needed to state which world-fiction is correct, not about storytelling as a cross-world proof. The story can contain any element the fictionalist wants (including structureless worlds), and the argument’s claim about reality holds regardless. The fictionalist’s modal claims reduce to claims about a story, and the reality in which the story is told instantiates the structural roles as the condition of the story being tellable. The deflation runs at the story level; the argument runs at the level that has to be there for stories to be told at all. This is Brock’s dilemma (1993) arriving at the point of deployment rather than in the abstract. Fictionalism can state its modal truths only by presupposing a non-fictional fact about what the fiction says, and the meta-level refinement names that non-fictional fact outright, that reality instantiates T, S, and Φ as the condition of the story’s being told. The refinement does not evade the dilemma; it is the dilemma realized at exactly the level the argument’s claim occupies.

Third. The no-correspondence move presses the boldest version. Modal claims have no truth conditions beyond what the fiction specifies. There is no fact of the matter about whether w₀ is possible beyond what the story says. The argument’s claim that w₀ is incoherent is itself just another story about stories: a meta-fiction whose ontological weight is no greater than the fiction it critiques.

The boldest version is unstable at the meta-level. If modal claims have no truth conditions beyond what the fiction specifies, the fictionalist still owes a non-circular account of the translation’s correctness and of which fiction is privileged. The meta-claim about that semantics is not automatically itself modal, so self-defeat does not follow merely from its formulation. What the move cannot do is extend the no-truth-conditions treatment to that meta-claim itself: the account of the translation’s correctness and of which fiction is privileged must hold non-fictionally, or the position dissolves into a fiction about fictions with nothing to recommend it. The escape is purchased by admitting a level of non-fictional fact. At the level at which the fictionalist’s meta-position is coherent, it requires the structural conditions the argument identifies, each realized through its bearer: a reasoner (Substance) who evaluates fictions (Time) and distinguishes the meta-fiction from the fictions it describes (Space). The bold move does not escape structure. It demonstrates structure as the condition of its own coherence.

A kindred move drops full-blown fictionalism for Sider’s (2011) pragmatic deflationism: modal talk serves practical purposes and carries no deep ontological commitment, but neither does it require a full-scale fiction. The reply follows the same path. Pragmatic deflation evaluates modal statements by their utility for reasoners whose reasoning instantiates the roles; the pragmatic frame is the structure’s tool for its own modal reasoning. Deflation does not escape structure; it describes structure’s modal practice from within.

The Lewisian concrete-worlds alternative is §4.4.1. The modal-collapse worry, that necessary grounds propagate necessity through all of modal structure, is §4.4.9. Williamson’s necessitism is §4.4.6; Sider’s structuralism is §4.4.7. The Whiteheadian process ontology is §4.4.8. The parody objection, that the argument’s modal machinery licenses absurd conclusions, is §4.4.10.

What converges. Each of the three forms concedes the structural conditions at the level the argument’s claim applies. The story-element move fails because the fiction itself must be tellable: telling the story of a worldless world requires individuating it (Space), attributing properties to it (Substance), and unfolding the telling across moments of consideration (Time), a fiction written nowhere. The meta-level refinement saves the move only by conceding that the structural roles operate at the meta-level, which is exactly where the argument’s claim applies. The no-correspondence move preserves its coherence only by admitting a non-fictional level: the account of which fiction is privileged must hold independently of any fiction, and at that level the structural conditions apply.

Rosen’s paper was written at a moment when Lewis’s plurality had begun to look, to many of its readers, like a reductio of the very modal realism it was built to defend. Fictionalism promised the semantics without the ontology: a way to talk about worlds without being committed to them. What the three decades of debate since have shown is that the promise could not be kept at the level the argument is concerned with. The fiction can dispense with the worlds; the storytelling cannot dispense with the storyteller, the time in which the story unfolds, or the space across which its elements are distinguished. Rosen’s instinct (that modal talk does not commit us to a vast plurality) was sound. His conclusion mistook the target. That Rosen himself later drew back from the strongest version of fictionalism is not incidental to the case: he retreated under precisely the pressure these three forms exert, the recognition that the deflation cannot be carried all the way down without dissolving the level at which the deflating is itself done. What modal talk does not commit us to is the plurality. What it commits us to is the structural conditions any plurality, real or fictional, has to be told within.

4.4.5 Modal Expressivism

Modal expressivism, defended by Blackburn (1986, 1993) and developed by Brandom (2008) and Thomasson (2020), denies that modal sentences express propositions about modal facts. Saying necessarily, water is H₂O is, on this view, expressing a stance (a commitment to a certain inferential pattern, a certain way of treating the claim) rather than describing a feature of reality. Modal vocabulary is a tool for organizing thought, not a window onto modal reality. The position parallels expressivism in ethics (Gibbard 1990, 2003) and in mathematics, and it aims to deflate modal ontology without giving up modal practice. We can keep saying necessarily, P while denying that there are modal facts for necessarily, P to be true of.

The argument’s reply turns on what expressing a stance requires. Stances are states. They have content. They are distinguishable from other stances. They are sustained by a thinker who persists from the formation of the stance to its expression. The expressivist’s framework operates on the structural conditions the argument identifies, and the operation is not a marginal feature of expressivism: it is the mechanism by which expressivism replaces descriptive ontology.

The Frege-Geach problem (Geach’s 1965 sharpening of a worry traceable to Frege: how can sentences that merely express stances figure as components in logically valid inferences when embedded in conditionals and other complex contexts?) deserves full engagement. Modal claims figure in valid inferences (modus ponens, from if P then Q and P, conclude Q, and conditional reasoning more broadly) that look like inferences over propositions, not over stances. Schroeder (2008) gives the most sophisticated expressivist response, building stances out of “being for” attitudes (mental commitments toward propositional contents, treated as a basic stance-type with logical structure) whose logical relations are engineered to mirror propositional logic. Gibbard’s quasi-realism (1990, 2003) takes a parallel line: the expressivist earns the right to talk as if modal sentences had truth-conditions by showing that stance-talk supports the same inferential structure realist talk supports.

Both responses are sophisticated. Their actual deployment requires structured agents and inferential practices, but this act-side requirement does not by itself establish that modal vocabulary describes mind-independent modal structure. Stance-composition requires that one stance be distinguishable from another, that combinations of stances yield further stances with determinate inferential properties, that the network of stance-relations is something a thinker can navigate over time. Schroeder’s “being for” attitudes are themselves states with content, sustained by a subject, distinguishable from other states. The structural conditions are required by the resources expressivism uses to do without them.

Two variants of the family require engagement on their own terms, since neither rests its case on the stance-bearing the reply above targets. Brandom (2008) constitutes modal content not by a speaker’s stance but by inferential role: the material inferences a claim licenses within a discursive practice, which modal vocabulary makes explicit. The Brandomian can grant everything said above about stances and still hold that content is fixed by inference, not by any condition on stance-formation. Actual inferring and deontic scorekeeping involve ordered acts, differentiated commitments, and continuing practitioners. That establishes the structural conditions of the discursive practice at act-side resolution. Inferential relations considered abstractly need not themselves be temporal processes or physical bearers, and the expressivist can still deny that modal sentences represent independent modal facts.

Thomasson, in Norms and Necessity (2020), locates modal claims differently again: neither stances nor descriptions but object-language expressions of the semantic rules constitutive of our concepts. Her normativism needs a distinct reply only in its bearer, not in its form. A rule meets the same three conditions even more plainly than a stance does: it is instituted and followed across cases, it discriminates correct from incorrect application and one rule from another, and it requires competent users who persist as the same rule-followers across applications. A rule just is a standing pattern applied over time by a community that endures to apply it, so the bearer changes but the structural demand does not. Whatever bears the modal (stance, inference, or rule) is sustained by the conditions the argument identifies.

A second front opens at the level of scope. Expressivism is most plausible as a localized thesis about a particular fragment of vocabulary: this part of language is expressive, that part is descriptive. The local form requires a fact about which sentences fall on which side, and the local expressivist is content to admit such facts at the metalinguistic level while denying them at the object level. Globalizing expressivism (the thesis that all assertion is stance-expression) generates a different problem. The claim modal sentences express stances rather than facts is itself either a fact about modal sentences (in which case expressivism has been falsified by the position’s own articulation) or a stance about modal sentences (in which case the position lacks the descriptive force its proponents take it to have). Dreier (2004) names the “creeping minimalism” worry from the realist side: as expressivism succeeds in recovering the surface features of realist talk, it becomes harder to tell what the disagreement was about.

The present argument’s intervention is orthogonal to that debate. It does not press the realist demand that modal sentences correspond to modal facts. It presses the prior point that any assertion (descriptive, expressive, or quasi-real) operates on T, S, and Φ. Local expressivism faces this point at its metalinguistic level, where the demarcation between expressive and descriptive vocabulary is drawn. Global expressivism faces it at every level, since the structural conditions are exhibited by stance-formation as decisively as by description.

What is established. Any stance (whether descriptive or expressive, whether about modality or anything else) operates on the structural conditions. T, S, and Φ are not what modal vocabulary describes; they are what any vocabulary requires.

4.4.6 Necessitism

Timothy Williamson published Modal Logic as Metaphysics in 2013, a book whose title names exactly what the project is. His argument (developed across two decades of papers and consolidated in the book) is that modal logic is not a regional formalism for talking about possibility but the central apparatus of metaphysics itself, and that its quantifiers, when interpreted with the seriousness the formalism deserves, range over a constant domain across worlds. Necessarily everything is necessarily something. Whatever there is belongs to the same unrestricted domain at every world, including merely possible objects that are non-concrete at the actual world. Contingency concerns which objects are concrete and which properties they instantiate, rather than variation in the unrestricted domain itself. The Barcan formula and its converse, which together express the commutativity of quantification and modality, are validated. The view is the most ambitious modal-logical position currently defended, and the case for it draws on the simplicity, generality, and explanatory reach of constant-domain QML (quantified modal logic).

Williamson (Modal Logic as Metaphysics, 2013) is the systematic statement; “Bare Possibilia” (1998), “Necessary Existents” (2002), and “Everything” (2003) develop the apparatus. The position descends from Barcan Marcus (1946), who first proved the formula that bears her name in a quantified modal logic and pressed its philosophical consequences. Against necessitism stands a cluster of contingentist responses: Stalnaker (Mere Possibilities, 2012) defends the contingent-domain alternative by arguing that the apparent simplicity of constant-domain quantification is purchased at the cost of a domain populated by “merely possible” entities whose existence-status is obscure; Fine (“Necessity and Non-Existence,” 2005) develops an essence-based account on which Socrates can be necessarily Socrates without necessarily existing; Adams (“Actualism and Thisness,” 1981) defends actualism against the Barcan-formula consequences; Hale (Necessary Beings, 2013) develops a structured-necessity account on which different kinds of necessity have different scopes. Plantinga (1974b) had in effect pressed the objection in advance: such a position misreads modal claims about objects as modal claims about haecceities (individual thisnesses, the property of being this very individual and no other). The contemporary higher-order modal logic literature (Linnebo, Thin Objects, 2018; Linnebo and Rayo, “Hierarchies Ontological and Ideological,” 2012) extends the dialectic into plural-quantification frameworks where the Barcan formula’s status is again disputed.

The present reply sits closer to Stalnaker on the contingentist question and adds what Stalnaker alone does not specify: which contingencies the structural argument requires, and why those contingencies survive Williamson’s framework whether or not constant-domain quantification is endorsed.

The challenge to the structural argument is double-edged. On one hand, necessitism would seem to make the necessary being trivially available: every possible thing exists necessarily, so the necessary being is among them. On the other hand, necessitism eliminates the explanatory work the necessary being is supposed to do. If everything exists necessarily, there is no explanatory contrast between the necessary being and contingent beings, because there are no contingent beings in the relevant sense. The argument’s modal conclusion □(T∧S∧Φ) (the necessary obtaining of the three structural Conditions T, S, and Φ, borne by Time, Space, and Substance) lands trivially in a framework where everything is □ (necessary), which means it lands without distinguishing the structural conditions from anything else. Worse: the argument’s premises require contingency the framework appears to deny. ◇C (the possibility of change) is defended in §2.4 as a substantive possibility-claim grounded in Cartesian phenomenal givenness: change is possible because change is occurring in the very act of doubt. If the framework reduces all contingency to qualitative variation across a fixed domain, the question is whether ◇C remains a genuine modal possibility distinguishable from □C, and whether the contingency of which world is actual (presupposed for the proof’s stretch from □◇C to □(T∧S∧Φ) being non-trivial) survives at all.

A clarification before the moves. The dispute is not with quantified modal logic, the Barcan formula’s formal validity in S5, or the ambition of modal logic to do metaphysical work. The argument shares with Williamson the commitment that modal logic is not a regional formalism but the apparatus through which metaphysical structure is articulated. The dispute is with whether the move from formal validity in constant-domain QML to ontological necessitism preserves the contingencies the structural argument requires for its premises to retain their force. Three forms of the move press that question at three resolutions: the constant-domain compatibility move, the Cartesian-Certainty preservation move, and the actuality-contingency preservation move. The space they canvass is the space necessitism can occupy. Each requires the structural roles its modal apparatus presupposes, and each leaves the argument’s needed contingencies intact at the level the argument needs them.

First. The constant-domain compatibility move presses that the structural argument’s modal conclusion lands trivially in Williamson’s framework. □(T∧S∧Φ) is a necessity claim, and necessitism makes every existence claim a necessity claim. The conclusion is therefore consistent with necessitism but undistinguished by it. If the structural argument’s distinctive contribution is the modal status of T, S, and Φ, and necessitism flattens all modal-status differences at the existence level, the argument has not earned anything Williamson’s framework does not already deliver. The conclusion would be true but not informative.

The reply distinguishes the structural sense in which T, S, and Φ are necessary from the necessitist sense in which everything is necessary. Necessitism’s necessity is constant-domain: the same things exist in every world. The structural argument’s necessity is constitutive: T, S, and Φ are conditions any world must instantiate to be a world. The two are compatible but not equivalent. A necessitist can accept that the things which exist in every world include configurations satisfying T, S, and Φ, and the structural argument adds that the satisfaction of T, S, and Φ is what makes the constant domain a domain of any kind rather than a featureless plurality.

The Barcan formula and its converse, applied to the structural roles, yield exactly what the argument independently derives (every world’s domain instantiates T, S, and Φ), but the formula does not by itself explain why every world’s domain instantiates them. Appendix A’s Domain Semantics and the Barcan Question records that the structural argument quantifies over Conditions rather than individuals, so Barcan and Converse Barcan are orthogonal to the proof’s load-bearing claims at the structural level.

The structural argument proposes a metaphysical explanation Williamson’s formal framework does not itself provide. This should not be attributed to a general truthmaker requirement on Williamson’s part: the narrower point is that constant-domain semantics alone does not establish the book’s constitutive account of T, S, and Φ. The structural account answers: the constitutive role of T, S, and Φ in any configuration is what makes the constant domain a domain, and the quantifier’s range across that domain is the formal shadow of the constitutive fact. The constant-domain framework is exactly the framework the structural argument’s conclusion describes, with the constitutive content filled in.

Second. The Cartesian-Certainty preservation move presses the harder version. The argument’s central premise (◇C, the possibility of change) is defended in §2.4 as a substantive possibility-claim grounded in Cartesian phenomenal givenness: change is occurring in the very act of doubting, and that the doubting is a change cannot itself be doubted without instantiating it. ◇C must therefore mark a genuine modal possibility distinguishable from □C, the necessity of change. On Williamson’s framework, the objector presses, this distinction collapses. Contingency is qualitative: which properties things have varies across worlds, but which things exist does not. If ◇C is read as the claim that change-events exist, necessitism makes that claim trivially necessary. If it is the claim that change-events possibly occur, the modal qualifier ◇ adds nothing beyond the qualitative-contingency level Williamson already permits. Either way, the Cartesian datum (that something might not have been changing) loses its modal weight, and the proof’s first step is undermined.

The collapse does not occur, because the premise’s modal weight does not sit at the level necessitism flattens. ◇C is not a claim about which entities exist in modal space; it is a claim about which configurations the structural roles permit. The Cartesian datum is not ‘a change-event exists’ but ‘the act of thinking now occurring is itself a change’, and the modal scope of the premise is over configurations of the roles, not over the existence of relata. Necessitism leaves the configuration-level intact: even if every entity exists necessarily, the question of which configuration any entity is in remains a qualitative-contingency question, and qualitative contingency is precisely what Williamson preserves. ◇C marks a substantive possibility at the level Williamson’s framework licenses: configurations vary across worlds, the configuration-of-change is one configuration among others, and the Cartesian certainty is that this configuration is instantiated.

Stalnaker (2012) is the natural ally here; the contingentist alternative makes the same observation in a different idiom: even if necessitism is correct about domain-constancy, it is not correct that all modal weight collapses to the qualitative level, because configurations themselves carry modal scope. ◇C is a configuration-level claim and survives the necessitist reframing intact.

The explicit formalization is given at §2.8.4: □F ∧ (◇p ∧ ◇¬p) for F = T∧S∧Φ records the configuration-level contingency the Cartesian datum requires while preserving structural necessity at the F-level. Fine’s (2005) essence/necessity distinction sharpens the point: that change is occurring is a Cartesian datum about the actual configuration; that change-events exist necessarily (if Williamson is right) is a Williamsonian claim about the domain. The two claims are at different levels, and the structural argument’s premise is at the configuration level the necessitist framework leaves untouched.

Third. The actuality-contingency preservation move presses the boldest version. The structural argument’s modal line decomposes into four distinct elements rather than a single stretch. P3 is ◇C, asserted in §2.4 as a Cartesian datum about the actual world: change is occurring in the very act of doubt. The S5 lift from ◇C to □◇C uses the collapse of iterated modalities: once ◇C obtains, ◇C obtains necessarily, since possibility in S5 is itself a necessary feature of modal space. The Modal Identity of §2.6.6 is the biconditional □(◇C ↔︎ T∧S∧Φ): the necessary equivalence between the possibility of change and the co-instantiation of the three structural roles. The theorem-level conclusion □(T∧S∧Φ) follows from □◇C and the biconditional by K-axiom closure under □. The objection’s bite lands not on the validity of the closure but on its informativeness: if which world is actual is not itself contingent (if there are no worlds whose configuration includes change distinguishable from worlds whose configuration does not) the closure runs formally but the conclusion records no metaphysical fact the alternatives could have falsified. The argument needs it to be a substantive fact that we are in a world whose configuration includes change. On Williamson’s framework, the objector presses, the contingency of which world is actual cannot be sustained. Necessitism plus constant-domain quantification reduces the difference between worlds to qualitative variation in property-distribution; the actuality of this configuration becomes a brute haecceitistic fact rather than a contingent one. If which world is actual is not genuinely contingent, the stretch from □◇C to □(T∧S∧Φ) loses its bite: the conclusion follows by necessity from premises that themselves cannot be modally distinguished from their alternatives.

The reply runs at the level the proof actually operates. The actuality-contingency the argument needs is qualitative-contingency, not ontological-contingency, and qualitative-contingency is exactly what Williamson preserves. The proof requires that some configuration is actual and that not every configuration is actual: both are qualitative-contingency claims about how the constant domain is configured, not ontological-contingency claims about what the domain contains. Necessitism endorses the first (every world has a configuration) and is silent on the second (it does not require that every configuration be instantiated). Indeed, necessitism requires qualitative-contingency to be substantive, since otherwise every world is qualitatively identical and the framework has no work to do. The actuality-contingency the structural argument requires is precisely the qualitative-contingency Williamson’s framework treats as the residue of contingency that survives constant-domain quantification.

Adams (1981) is helpful here as a foil: Adams holds that the qualitative-contingency level is itself unstable, but his argument depends on the haecceitist reading Williamson explicitly rejects. Hale’s (2013) structured-necessity account makes the cleaner point: necessities at different levels do different work, and the structural argument’s necessity at the constitutive level is compatible with qualitative-contingency at the configuration level. The stretch from □◇C to □(T∧S∧Φ) is non-trivial because the configuration-of-change is one qualitative possibility among others, and the actual configuration’s instantiation of that possibility is what the proof’s first premise records.

Williamson’s positive defense of S5 in Modal Logic as Metaphysics rests on considerations the structural argument finds hospitable rather than threatening. The present proof likewise derives S5 from fixed-domain semantics: real possibility ranges over one global domain W, so a witness to ◇P remains a witness at every evaluation point. This is distinct from Williamson’s constant-domain thesis about individuals. A fixed domain of possible worlds does not by itself entail a constant domain of entities across those worlds.

The Brouwer axiom (p → □◇p) then follows for the present operator because any actual P is witnessed within W, and every world evaluates possibility over that same W. The Cartesian datum establishes ◇C through actuality; the fixed-domain account, not a prior claim that T, S, and Φ are everywhere instantiated, yields □◇C.

The disagreement is therefore precise. Williamson uses constant-domain QML to motivate ontological necessitism. The structural argument uses a fixed global domain of real possibilities to establish modal invariance, then separately derives the necessary Conditions that any world in that domain realizes. The shared formal result does not settle the ontological reading.

Fine, engaged as dialectical opponent rather than as authority, presses the structural argument from a complementary direction. Fine’s “Essence and Modality” (1994) and “Necessity and Non-Existence” (2005) establish that necessity derives from essence: what makes a modal claim necessary is that it follows from the essence of some entity. Applied to the structural argument: what makes □(T∧S∧Φ) necessary? If it follows from the essence of T, S, or Φ themselves, the argument is circular: T, S, and Φ must already exist and have essences to generate the necessity, which is what the argument is trying to establish. If it follows from the essence of change, Fine will press that conceptual essence generates analytic necessity rather than unrestricted metaphysical necessity. The structural reply is that the argument’s modality is neither essence-based in Fine’s sense nor merely analytic. It is role-necessity: the necessity that anything being a something at all instantiates the structural conditions, grounded not in the essence of T, S, or Φ as entities but in the role-structure that any entity’s being anything at all requires. The book proposes role-necessity as an alternative explanatory basis, but Fine’s framework is flexible enough to ask whether structural roles themselves have essences or are grounded in more basic essences. The challenge therefore remains substantive: the book must defend the priority of role-structure rather than treating it as outside Fine’s apparatus by definition.

A kindred move drops Williamson’s necessitism for Linnebo’s higher-order modal logic, where the Barcan formula is reformulated in plural-quantification terms and the necessitism/contingentism dialectic re-emerges at the level of pluralities rather than individuals. This is the sharpest form of the necessitist challenge, and it cannot be turned aside by the observation that the structural roles are conditions rather than individuals. The first-order replies bank configuration-level contingency as the residue necessitism leaves free, and they can do so because first-order necessitism rigidifies only the domain of individuals. Higher-order and plural necessitism rigidify properties and pluralities as well, and configurations live at exactly that level. Plural logic is not silent where configurations operate; it is a logic of that level. The reply must therefore be earned at the higher order, not asserted to lie beneath it.

It is earned by lifting Williamson’s own distinction to the plural level. Necessitism already separates being from concreteness: whatever exists, exists necessarily, but what is concrete is contingent; a thing has its being in every world and is concrete in only some. Carried up to pluralities, higher-order necessitism rigidifies the being of configurations-as-possibilia, but its contingently-concrete machinery leaves which configurations are concretely actual contingent. That residue is precisely what the proof draws on. The first premise does not record that some plurality exists; it records that the configuration-of-change is the one concretely instantiated. The contingency at work is concreteness-contingency, and concreteness-contingency survives plural necessitism for the same reason it survives the individual case: Williamson’s framework rigidifies being, not concreteness, at whichever order it is run. Linnebo’s apparatus extends the dialectic to the level of pluralities; pressed there, it presses the structural conclusion no harder than at the level of individuals.

The Barcan-formula version of the modal-collapse worry, that any necessity at the existence level propagates to all existential claims, is engaged in its Spinozan-propagation form at §4.4.9, where the constitutive/productive distinction does the work necessitism alone cannot do.

The Lewisian concrete-plurality alternative, which gives modal claims maximal ontological weight rather than minimal, is §4.4.1. The modal-collapse worry, that necessary instantiation flattens all modal distinctions, is §4.4.9. Sider’s structuralism, which deflates rather than embraces necessity, is §4.4.7. The Whiteheadian process alternative is §4.4.8. Modal fictionalism’s attempt to deflate the whole apparatus is §4.4.4.

What converges. Each of the three forms of the necessitist threat operates at a level the structural argument either accommodates or does not depend on. The constant-domain compatibility move is correct that □(T∧S∧Φ) is consistent with necessitism, but it underestimates what the structural argument adds: the constitutive account of what makes the constant domain a domain. The Cartesian-Certainty preservation move presses the strongest threat to the proof’s first premise; ◇C survives because its modal weight sits at the configuration level Williamson preserves, not at the existence level Williamson flattens. The actuality-contingency preservation move presses the boldest version; the contingency the proof requires is qualitative-contingency at the configuration level, which is exactly what necessitism leaves intact, and which necessitism in fact requires if its framework is to do any work at all.

Williamson draws the conclusion that the formal apparatus deserves to be taken at face value: the quantifiers range over a constant domain, the Barcan formula and its converse are validated, and what looked like ontological contingency dissolves into qualitative variation across a fixed plurality. The structural argument shares the formal commitment and adds what Williamson’s framework requires but does not specify: the constitutive role of T, S, and Φ in making the constant domain a domain. What Williamson left behind is not a refutation of the structural argument but a cleaner statement of the level at which the argument’s necessity claim operates: not at the qualitative-contingency level where Williamson preserves modal variation, but at the constitutive level where the variation is hosted. Necessitism is the framework the structural argument’s conclusion is most compatible with at the formal level, and the argument supplies the constitutive reading its formalism leaves underspecified.

4.4.7 Sider’s Structuralism

Theodore Sider published Writing the Book of the World in 2011, a book whose title names exactly what metaphysics ought to do. His argument is that the world has a privileged structure (a set of joint-carving notions) and that the project of metaphysics is to identify which notions those are. Existence quantifiers, mereological parthood, fundamentality, even logical operators: Sider is willing to ask of each whether it carves at the joints or merely tracks our practical interests. Modality is among the notions whose structural status Sider’s methodology invites us to assess. His project should not be reduced to the blanket claim that necessity is merely a relation between structure and language. The sharper rival considered here is a Siderian or Humean extension on which modal ideology is not fundamental and modal truths supervene on non-modal joint-carving structure.

Sider (Writing the Book of the World, 2011) is the systematic statement; “Ontological Realism” (2009) is the methodologically central earlier paper. The position is part of a broader contemporary trend: naturalistic structural realism in philosophy of physics (Ladyman & Ross, Every Thing Must Go, 2007; French, The Structure of the World, 2014, both engaged at §4.3.9), Quine’s structuralism in philosophy of mathematics (1969, the view that mathematical objects are exhausted by their structural roles rather than possessing intrinsic identities), and the structuralist movement in philosophy of science going back to Russell and continuing through Worrall (“Structural Realism: The Best of Both Worlds?”, 1989). Sider’s contribution is to extend structuralism beyond philosophy of science into general metaphysics: not just physical structure is fundamental, but logical, mereological, and existential structure. Against Sider’s structuralism, the contemporary literature includes Schaffer (Monism, 2010, on grounding-as-fundamental), Lewis (On the Plurality of Worlds, 1986, on modality-as-fundamental, engaged at §4.4.1), Williamson (Modal Logic as Metaphysics, 2013, on modal logic as metaphysically central, engaged at §4.4.6), and Fine (“The Question of Realism,” 2001, who develops a kindred structuralism with different commitments). The present reply sits closer to Sider on the empirical question (physics does carve at the joints in something like the way Sider claims) and presses the modal question: whether Sider’s structure-without-modal-scope can deliver what the present account’s predictions require.

Sider’s structuralism presses the central claim of §1.5 from a direction §4.4.1, §4.4.9, and §4.4.6 do not. Lewis multiplies the modal ontology; Modal Collapse worries about its propagation; Williamson endorses its scope. Sider does something subtler: he deflates it. Modality, on his picture, is a way of speaking about what the privileged structure looks like through the lens of various linguistic frameworks, not a feature the structure itself possesses. T, S, and Φ (the three structural Conditions, borne by Time, Space, and Substance, the trio the book’s argument derives) might be perfectly real, perfectly joint-carving features of the actual world without being necessary features in any deeper sense. The argument’s necessity claim (□(T∧S∧Φ), read as “necessarily, T and S and Φ obtain together”) is, on the structuralist’s reading, a claim about how the structure looks under a particular modal idiom rather than a claim about anything the structure itself contains. The threat is precise: if modal scope is deflated, the cycle prediction (§7.6), the consciousness-elsewhere claim (§7.8), and the foreclosure of absolute nothingness (§2.7) all lose the modal weight they were carrying. They become claims about this world’s structure rather than claims about the structure of any world.

A clarification before the moves. The dispute is not with structural realism as a methodological program. The argument shares with Sider (and with Ladyman & Ross’s naturalistic structural realism, engaged at §4.3.9) the commitment that fundamental ontology is given by structural relations rather than by hidden quiddities. The dispute is with whether the structural ontology can be specified without modal scope. Three forms of the move press the deflation at three resolutions: the bare deflation move, the redundancy move, and the no-modal-realism move. The space they canvass is the space modal-deflationary structuralism can occupy. The replies test whether any form can deflate modal scope without requiring it to specify what the structure is.

First. The bare deflation move holds that modal claims are not joint-carving. There are facts about what is the case (Sider’s privileged structure, the actual configuration of the world) and there are facts about what would have been the case, but the second class of facts is derivative. Possible-worlds talk tracks something useful; what it tracks is not modal reality but the relations between linguistic frameworks and the privileged structure. The argument’s modal claims, on this reading, are linguistic claims with structural correlates rather than structural claims about modal space.

The deflation requires specifying what the privileged structure is. Sider’s account answers in joint-carving notions: existence, parthood, fundamentality, identity, predication, plus whichever physical and logical primitives the final theory turns out to require. Two readings of joint-carvingness are available. On the first, joint-carvingness is exegeted modally: what the structure would recommend across counterfactual variation in our descriptive vocabularies. On the second, pressed most fully by the Sider–Fine–Schaffer hyperintensional-structuralism alliance (Fine, “The Question of Realism,” 2001; Schaffer, “On What Grounds What,” 2009), with Lewis’s perfectly-natural-properties as the precedent for a non-modal structural primitive in even the arch-modal-realist’s metaphysics, joint-carvingness is primitive structural-fundamentality, hyperintensional, not analyzable modally.

The argument’s reply does not require the first reading. Grant Sider the second: joint-carving is primitive structural-fundamentality, ontologically prior to modal facts. The argument’s modal scope is still required, but it comes not from the joint-carving exegesis; it comes from what the empirical predictions of §7.6, §7.8, and §2.8 carry across substrates. The redundancy move next makes the empirical-predictions ground precise.

A note on Sider’s positive view (Writing the Book of the World, §12). Sider’s modal deflationism is not mere skepticism but a positive Humean account: modal facts supervene on the actual-world mosaic of non-modal joint-carving patterns. What makes a modal claim true is whether the pattern it describes is joint-carving in the actual mosaic. This is a richer position than bare deflationism: it grounds modal claims in structural fabric rather than dismissing them. The structural argument’s reply takes the positive Humean account seriously and presses one question: does the mosaic’s joint-carving character itself hold necessarily or contingently? If contingently (if a different cosmos could have had entirely different joint-carving patterns) then Sider’s Humean modal grounding bottoms out in a contingent mosaic, and the structural argument’s necessity claim is not captured by it. If necessarily (if the joint-carving patterns of any mosaic are the patterns any mosaic must have by virtue of being a mosaic at all) then Sider’s positive Humeanism, fully developed, endorses the structural argument’s conclusion rather than deflating it. The structural argument offers the second reading: what makes T, S, and Φ joint-carving across any cosmos is that they are conditions on any fabric having causal joints at all. Sider’s positive Humeanism and the structural argument converge when the necessity of the structural mosaic is taken seriously rather than deferred.

Second. The redundancy move presses harder. Where the present argument makes modal predictions, the cycle (§7.6), substrate-independence of consciousness (§7.8), the preclusion of absolute nothingness (§2.7), Sider’s structuralism delivers them as descriptions of how this world’s structure works, without requiring modal scope. The cycle is what the actual cosmic dynamics deliver; the argument’s modal apparatus is metaphysical surplus. The substrate-independence claim is what neuroscience plus the physics of information processing delivers; the argument’s “in any world where T, S, Φ obtain…” formulation is bookkeeping. The preclusion of absolute nothingness is what cosmological inflation plus the structure of quantum field theory delivers; the modal scope adds nothing physics cannot say more carefully without it.

The redundancy diagnosis fails at the kind of generality it claims to absorb. The distinction the argument needs is not between a local forecast and a falsifiable cross-substrate empirical prediction: the substrate-neutral structural definition at §7.8 (closure-plus-throughput as a structural stance) and the cycle at §7.6 are structural-necessity claims about what any cosmos with T, S, and Φ must instantiate, rather than prediction-candidates in the narrow falsificationist sense. The reply therefore does not hostage itself to the standing of those predictions. A reader who suspends judgment on whether the cycle recurs (§7.6) or consciousness arises on other substrates (§7.8) still confronts the type-distinction, which turns on what kind of claim each makes (a structural-necessity claim across cosmological variation) rather than on whether either is ultimately borne out.

The distinction is between two kinds of structural generality: the Humean joint-carving description says ‘this universe’s constants and laws produce this pattern, and the pattern is joint-carving in this actual mosaic’; the modal structural-recurrence claim says ‘any cosmos hosting causal succession across a relational manifold will instantiate this structural pattern, because the pattern is what the structural conditions require, regardless of which particular constants or substrate produce them.’ Sider’s deflation can endorse the first kind of generality; it cannot endorse the second.

The across-substrate structural recurrence that §7.6 and §7.8 deliver is a scope of generality that local-universe joint-carving cannot capture: it identifies which features of our universe are universe-invariant (structural-necessity claims) versus which are local products of our particular physical constants (contingent-mosaic claims).

Whether this broader generality matters depends on what the structural argument is for. If only our actual universe is at stake, Sider’s deflation may suffice. If the structural conditions are genuinely necessary across any cosmos, as §§2.6–2.7 argue, then the modal scope is not surplus but constitutive; it is the difference between describing what this world happens to do and identifying the conditions any world must satisfy to be a world at all.

Third. The no-modal-realism move presses the boldest version. Modal facts are not facts about modal reality; they are facts about how the privileged structure relates to descriptive frameworks. The structural argument’s “necessarily, T∧S∧Φ” can be rewritten as “in the privileged structure, T∧S∧Φ are joint-carving features of the actual world’s fundamental ontology,” with no modal claim left over. The argument’s content is preserved, its modal commitment dissolved.

The rewriting fails to preserve content. “In the privileged structure, T∧S∧Φ are joint-carving features of the actual world’s fundamental ontology” is true if T, S, Φ obtain in the actual world and the structure carves at the joints. It does not entail that T, S, Φ obtain in any other world. But the argument needs the entailment. The cycle prediction, the consciousness-elsewhere claim, and the foreclosure of absolute nothingness (§7.6, §7.8, §2.7) hold across worlds, not just within the actual one, and the structural ground is what licenses the across-world reach. Sider’s rewriting eliminates the across-world reach. What it preserves is the actual-world structure; what it loses is the necessity that makes the structure structural in the load-bearing sense. Joint-carving features can be contingent; structural conditions of any cosmos are not. The bold rewriting trades the second for the first and calls the trade a translation.

A kindred move drops Sider’s full structuralism for the more cautious naturalistic structural realism of Ladyman & Ross. NSR holds that physics gives us structural facts about the actual world without committing to whether those facts are necessary across worlds. The reply runs as it does at §4.3.9: NSR is a methodological constraint on physics that is silent on the modal question; it is not a positive metaphysical alternative to the present argument. The argument welcomes NSR’s empirical structuralism and adds the modal architecture NSR’s caution declines to address. The two positions are not competitors but layered.

The Lewisian concrete-plurality alternative, which gives modal claims maximal ontological weight rather than minimal, is §4.4.1. The modal-collapse worry, that the argument’s modal scope over-generates necessity, is §4.4.9. Williamson’s necessitism, which licenses the modal scope at the type level, is §4.4.6. The Whiteheadian process alternative is §4.4.8.

What converges. Each of the three forms of structural deflation requires the modal scope it deflates. The bare deflation move can be granted its non-modal joint-carving primitive; the modal scope the argument needs is then carried by the across-substrate reach of its structural claims, where the redundancy move takes it up. The redundancy move loses the across-substrate reach the argument’s claims carry: the cycle, substrate-independence of consciousness, and the foreclosure of absolute nothingness are structural-necessity claims across cosmological variation, not local forecasts about our horizon’s future. The no-modal-realism move trades the across-world reach for actual-world joint-carvingness and calls the trade a translation, but the argument’s central claims hold across worlds, not just within the actual one.

The deepest agreement between Sider’s structuralism and the argument is real and worth naming explicitly. The structural-realist insight that physics names the constraints rather than discovers properties hidden behind them is already endorsed at §7.5. The Triconditional Cross-Entailment of §2.6.7 and the Ontological Identity of §2.6.8 operate within a framework that is recognizably continuous with Sider’s. What the argument adds (and what the present section makes explicit) is the modal floor under the structuralism. Joint-carving features can be contingent; necessary structural conditions cannot. The cycle, substrate-independence, and foreclosure claims have modal reach only if the prior necessity argument succeeds. Their empirical or explanatory usefulness cannot independently establish that reach without circularity. The book shares Sider’s joint-carving ambition but adds a contested modal thesis. A non-modal structuralist can accept the actual structural roles while withholding their cross-world necessity; later predictions inherit, rather than independently prove, the book’s answer to that dispute.

Sider’s defense of joint-carving structuralism is the most sustained one the analytic tradition has produced in this century: that there is a privileged structure, that metaphysics aims to identify it, and that the identification is genuinely revisionary: that we discover what the structure carves at, rather than choosing among equally legitimate frameworks. The defense is offered against a backdrop in which general metaphysics had begun to look, to many of its critics, like terminological housekeeping rather than discovery. The argument shares the project entirely. What it adds is the level of modal commitment on which the argument’s across-substrate claims depend: not just that this world has the structure it has, but that any possible cosmos must have at least this much structure to be a cosmos at all. Sider’s joint-carving is the local form; the argument’s modal necessity is the global form. The book of the world has more pages than one cosmos can fill.

4.4.8 Whiteheadian Process Metaphysics

Heraclitus held that nothing ever is, that everything becomes: what looks like a persisting thing is really a pattern maintained across constant material turnover, no drop of water in a river the same from moment to moment, only the shape and rate of its flow persisting. Alfred North Whitehead, twenty-four centuries later, worked out the most technically developed version of that ancient thought. If the process philosopher is right, what classical metaphysics called a substance is always a river, always just the apparent persistence of a pattern across occasions that arise and perish. The argument of this book asks the opposite: that what flows is Substance in verb form, and that asking a persistence question of a pattern only has content because something is carrying the pattern.

Whitehead (Process and Reality, 1929, from his Gifford Lectures of 1927–28) developed the classical statement of process metaphysics after leaving mathematical logic for philosophy at Harvard in his sixties. The position has three generations of defenders: Hartshorne (Creative Synthesis and Philosophic Method, 1970) develops process theology as the natural extension; Cobb (A Christian Natural Theology, 1965) and Griffin (Reenchantment Without Supernaturalism, 2001) extend it into theology and philosophy of mind (the process-theology branch named here is engaged directly at §5.3.3); Ford (The Emergence of Whitehead’s Metaphysics, 1984), Kraus (The Metaphysics of Experience, 1979), and Lango (Whitehead’s Ontology, 1972) develop the technical machinery. Rescher (Process Metaphysics, 1996) generalizes it beyond Whitehead’s specific terminology. Seibt (2017) gives the contemporary continuation, preserving process as primary while remaining silent on Whitehead’s more specific commitments. Against the process tradition stands the substance-realist line: Armstrong (A World of States of Affairs, 1997), Lowe (The Four-Category Ontology, 2006), and Heil (From an Ontological Point of View, 2003) all defend ontologies in which enduring substance is ineliminable. The present reply sits closest to Heil on the ontological question and adds what substance-realism alone does not: that Substance and process are not rivals but aspects. What the process carries, and what carries the process, are the same structural role described under two grammars.

The Whiteheadian challenge presses the specification in §2.6 of what Change requires and the identification at §2.6.10 of those requirements as Conditions-as-Nature. The argument of §2.6 is that succession requires temporal order, distinction requires extension, and exchange requires maintained causal capacity, with §2.6.8 specifying the Φ / Physicality role. Whitehead accepts the first two and contests the third: causal transmission, on his view, occurs through prehension (Whitehead’s term for the relational act by which one actual occasion takes account of and incorporates the character of a predecessor occasion) without requiring anything enduring. If prehension alone can do the work Substance does, the third role of §2.6 and the Conditions-as-Nature identification of §2.6.10 are incomplete: T, S, and Φ (the three structural Conditions the book’s argument derives, borne by Time, Space, and Substance) are not the three necessary structural roles; there are two, with the third replaced by a relational pattern among momentary occasions. This is a more refined threat than classical momentariness (§4.2.8), which denied persistence outright. Whitehead concedes the datum persistence captures (character does transmit across change) and proposes to capture the datum without the ontology.

A clarification before the moves. The dispute is not with the datum process philosophy starts from. The world plainly does flow, occasions plainly do arise and perish, pattern plainly does transmit across change. The dispute is with the ontological claim that the flow is fundamental and persistence is derivative. Three forms of the move press that fundamentality at three resolutions: the prehension move (pattern transmits without substrate), the superject move (what perishes leaves efficacious traces), and the process-all-the-way-down move (process is the final category). The space they canvass is the space Whiteheadian ontology occupies. Each relocates Substance under another name.

First. The prehension move holds that actual occasions perish entirely, and what we call persistence is the inheritance of character from predecessor to successor through prehension. No enduring substrate is required. The pattern is transmitted across a sequence of discrete creative acts, each complete in itself.

For a new actual occasion to inherit the character of its predecessor, there must be a structural connection between donor occasion and recipient occasion that transmits the character across the transition. That transmission is a directed dependence from donor occasion to recipient and requires differentiation plus causal efficacy: the predecessor’s character must genuinely condition the successor, not merely precede it. This directly realizes thin order, difference, and bearing roles. It does not by itself establish a spatial interval between occasions or a persisting substance; those loaded types require the separate role-to-type argument. This last requirement is where the prehension move must be met on its own terms rather than by positing a carrier it denies: the Whiteheadian is right that no occasion persists alongside the others to ferry the pattern across, but inheritance is not mere succession; an occasion that merely followed its predecessor without being causally conditioned by it would not have inherited anything. What makes prehension inheritance rather than succession is precisely that the donor’s character is causally efficacious in the constitution of the recipient, and that cross-transition efficacy is what Φ names (§2.6.8): not a substance enduring through the interval, but the causal capacity the inheritance-relation must instantiate to count as inheritance at all. Whitehead’s own account of prehension (the most technically developed account of how inheritance works) describes precisely this structural architecture. The prehensive relation does not float free of structural conditions; it instantiates them. The Whiteheadian account does not offer an alternative ontology to the structural picture. It offers an alternative vocabulary for the same structural facts. The roles are instantiated between occasions rather than within a persisting thing, but instantiated nonetheless.

Second. The superject move goes deeper. Each occasion perishes into what Whitehead calls the ‘superject’: a mode of being in which it becomes an object for future prehension. The superject is not a continuing subject but a contribution to subsequent occasions. The causal pattern is carried in the way the superject functions in the prehensions of its successors, not in any ongoing existence of the occasion itself.

Superject ontology relocates Substance rather than removing it. Even granted in full, objective immortality as a relational-inheritance fact rather than an entity persisting through an interval (the most thoroughgoing non-substantialist reading, parallel to the relational-availability move classical momentariness pursues at §4.2.8), the superject’s contribution to subsequent occasions instantiates the structural roles at the level of the transition itself. The inheritance of character from O₁ to O₂ has direction (Time’s asymmetry: from donor to recipient, not the reverse), differentiation (Space’s distinction: O₁ and O₂ are distinguishable contributors and recipients), and causal efficacy across the transition (Φ’s role, engaged as substrate at §2.6.8 in the form the transition itself must take). The ontological burden Substance carries in the present account (being the carrier of causal efficacy across the transition) is carried by the inheritance-relation in the Whiteheadian account whether that relation is glossed as persistence-through-interval or as availability-for-prehension. What changes is the name. What does not change is the work the role does at the level of the transition structure. What is settled here is the bearer-role: inheritance instantiates cross-transition causal efficacy under whatever grammar names it. The step from that role to a persistent causal being, the Φ-type, is neither carried here nor needed here; it is the Stage-2 work of the realizer biconditional Φ ↔︎ (T∧S) (§2.6.4), secured at §3.5.3. The Whiteheadian who grants cross-transition efficacy while denying that any occasion persists is therefore met on the persistence-type question there, not in this section, which secures only that no process ontology escapes the role.

Third. The process-all-the-way-down move presses the boldest version. Prehension is insufficient as a final ontology; what is needed is process as the fundamental category. Reality is not events in prehensive relation; reality is process: the dynamic, self-organizing flow of becoming that constitutes all finite entities. There is no substance because there is no static layer at any depth. Process is ontologically basic.

‘Process all the way down’ requires exactly the three structural roles at every level. For reality to be process, there must be transitions, differentiated phases or occasions, and some determinate continuity or efficacy linking them. These support the three roles at thin resolution. Requiring in addition a distinct thing that is processed would beg the question against process ontology, which treats the process or relation itself as fundamental. The deepest Whiteheadian commitment (that reality is nothing but process) does not escape the structural roles. It makes them the content of what the process is. Process ontology is Substance ontology articulated in the grammar of verbs rather than nouns. What the process consists in and what carries the process are the same structural role under two descriptions.

A kindred move drops the Whiteheadian apparatus for Bergson’s (Creative Evolution, 1907) élan vital (French: “vital impulse”): reality is not composed of substances but of a creative evolutionary impulse, durée (French: “duration”, lived time as continuous flow rather than discrete instants), whose continuous becoming precedes any static ontology. The reply is the same in French vocabulary. Creative becoming requires differentiation of the created (S), asymmetric direction of the creation (T), and continuity of the creating impulse across its acts (Φ). Bergsonian duration is the structural roles rendered in the grammar of flow.

The classical momentariness objection is §4.2.8. The emergent-time objection, that temporal flow is a derived feature, is §4.2.3. The modal-collapse worry that necessary process propagates necessity everywhere is §4.4.9.

What converges. Each of the three forms relocates Substance under another name rather than replacing it. The prehension move describes the structural architecture of inheritance, direction from donor to recipient (Time), differentiation of donor occasion from recipient (Space), and causal efficacy across the transition that makes inheritance more than mere succession (Substance), under alternative vocabulary. The superject move relocates Substance’s work into objective immortality: what the superject is and what it continues to do is exactly what Substance names, in non-substantialist grammar. ‘Process all the way down’ requires the three structural roles at every level (transitions, differentiated phases, and determinate continuity or efficacy linking them) with process ontology revealed as Substance ontology articulated in the grammar of verbs rather than nouns.

Whitehead delivered the lectures at Edinburgh in a form dense enough that the eventual book is read more often in commentaries than in the original. Process and Reality is the last great systematic metaphysics of the twentieth century. What it offers is a grammar (a way of redescribing the world under the verb rather than the noun), not an alternative to what the world structurally is. The Whiteheadian who argues that each occasion perishes entirely sustains the argument across multiple cognitive occasions, each thought inheriting the reasoning of its predecessor, carrying the argument forward to its conclusion. That inheritance is Substance described in verb form: the structural role persistence names, operating through the very transmission the argument describes.

Heraclitus said no one steps in the same river twice; the observation is acute and has never been refuted. But stepping, river, and twice are themselves structural descriptions: Substance moving through Time across Space. The flux Heraclitus saw and the structure Parmenides saw are one thing seen from inside and from outside; their dispute was over which vocabulary tracks it, not over which was illusion.

4.4.9 Modal Collapse Objection

The modal-collapse objection holds that a necessary ground generates necessary outputs: an argument that makes the structural conditions of reality inescapable thereby makes everything that happens within them inescapable too. Spinoza reached this conclusion by tracing out the consequences of substance monism, and drew it as a triumph: every mode follows with the force of geometrical demonstration, and the appearance of contingency is a mark of ignorance. Most of his successors reached the same conclusion and tried to find a way to block it. The present account, the objector presses, appears to rebuild the same trap from a different direction: with T, S, and Φ (the three structural conditions the book’s argument derives, borne by Time, Space, and Substance) in the place of Deus sive Natura (Spinoza’s identification: “God or Nature”) and the history of the world in the place of the modes.

The modal-collapse objection has a long lineage. Spinoza (Ethics)²⁶⁰ drew it as a positive conclusion: substance is necessary, its modes follow with geometrical necessity, and contingency is an illusion of perspective. Leibniz (Theodicy)²⁶¹ reached a softer form: God necessarily chooses the best possible world, and the specific history of this world is therefore fixed with a necessity one step removed from Spinoza’s. Modern defenses: Pruss (The Principle of Sufficient Reason)²⁶² presses the Leibnizian version against necessitarian accounts of grounding. Van Inwagen (An Essay on Free Will)²⁶³ sharpens the deterministic relative, the Consequence Argument: if the laws and the past are necessary, the future is necessary too. Rasmussen (How Reason Can Lead to God)²⁶⁴ applies the modal-collapse worry to necessitarian alternatives to classical theism. Against the collapse, the contemporary literature runs thick: Plantinga (The Nature of Necessity)²⁶⁵ deploys the de dicto/de re distinction in ways that block various propagation moves; Swinburne (The Christian God)²⁶⁶ defends a libertarian space within theism; the open theists (Hasker, Sanders) reject the collapse by denying the necessity of the antecedents. The contemporary grounding literature establishes the framework the present reply uses. Fine (“Guide to Ground,”)²⁶⁷ and Rosen (“Metaphysical Dependence: Grounding and Reduction,”)²⁶⁸ set out grounding as a relation of constitutive determination and defend the principle that a full ground necessitates what it grounds; Schaffer (On What Grounds What;²⁶⁹ Monism)²⁷⁰ and Dasgupta²⁷¹ develop the apparatus the present reply turns on: above all the distinction, owed to Fine, between full and partial ground. The structural account sits closest to Dasgupta’s line on grounding and adds: the specific grounds at issue are T, S, and Φ, and their necessity is compatible with the contingency of what they ground because they are only a partial, structural ground of any particular change, never its full ground.

The objection bears on two of the argument’s central theses. The Modal Identity is established at §2.6.6: the necessity of the structural conditions (□(T∧S∧Φ)) entails □◇C, the necessary possibility of change, not □C, the necessity of change itself. The Necessity of Nature is then established at §2.8.5: the specific character of the natural order is constrained by, but not reducible to, the structural conditions. If the modal-collapse objection succeeds, one of two things goes wrong: either the distinction between □◇C and □C dissolves (collapsing the stage into the play), or the grounding relation between structure and specifics is productive (collapsing the contingent specifics into the necessary structure). Both collapses entail Spinozism. The argument would have generated, by a different line, the very view Hume thought Spinoza’s necessitarianism reduced to absurdity.

A clarification before the moves. The objection is not that contingency is difficult to square with the argument: it is that contingency is incompatible with it. Three forms of the move press that incompatibility at three resolutions: the deterministic version, the Leibnizian best-of-all-worlds version, and the Spinozan propagation version. The space they cover is the space in which modal collapse could be forced. Each misreads what the argument establishes and misreads what grounding does.

First. The van-Inwagen-style propagation move presses the deterministic version. Necessary laws plus necessary initial conditions yield necessary outcomes. If T, S, and Φ are necessary, and the laws of physics follow from the structural roles, the laws are necessary too. Necessary laws plus a fixed past fix the future, so every event is necessary. The Consequence Argument delivers the collapse.

The Modal Identity of §2.6.6 draws a line the objection ignores. What the argument establishes is □◇C (the possibility of change is necessary), not □C (change occurs necessarily). Structural necessity fixes what is structurally possible, not which possibilities are actualized. A necessary stage does not script the contingent play. The necessity of T, S, and Φ constrains the type of physical law (any law must be expressible as a structural relation among the roles), but it does not fix the specific content. Which symmetry groups govern, which constants obtain, which particles compose the physical domain: these are contingent facts within the necessary architecture. Quantum theory supplies interpretations on which token history is genuinely open, but deterministic interpretations remain live. The modal-collapse reply does not depend on settling that dispute: the structural roles do not by themselves fix laws, initial conditions, or outcomes, so modal collapse does not follow from their necessity alone. The Consequence Argument targets determinism. The present account is not committed to determinism.

Second. The Leibnizian best-possible-world version goes deeper. A necessary, perfectly good ground must produce the best possible outcome. If the structure is necessary and it generates a universe, that universe must be the best possible one. Anything less contradicts the ground’s perfection. So the specific universe is necessary after all. Modal collapse returns through the back door.

The Leibnizian move presupposes a necessary evaluative subject: something that ranks possible universes and selects the best. The present account posits no such subject. T, S, and Φ are structural roles, not agents. They permit contingency; they do not rank it. The Leibnizian version gets traction against classical theism, where a necessary perfect Being must choose the best; it loses traction against a structural account with no such Being to rank outcomes. An impersonal version of the same demand fares no better. Axiarchism (Leslie, Value and Existence)²⁷² drops the chooser and grounds the demand in value itself: the good, on this view, tends toward its own actuality with no one selecting it. But a value that actualizes itself still actualizes some determinate configuration rather than the field of alternatives it leaves unactualized, and picking one over the rest is exactly the ranking operation the reply above blocks, run without an agent to perform it; value monism inherits the subject’s job without inheriting a subject to do it. Structure permits contingency; it does not rank it, personally or impersonally. Sharper still: the Leibnizian objection, applied to classical theism, generates exactly the collapse the theist wanted to avoid. If the Maximally Great Being is necessarily perfect and necessarily creates the best, the MGB’s freedom collapses into predetermined output. The objection is a better argument against classical theism than against the present account.

Third. The Spinozan propagation version goes severest. Necessity propagates through grounding relations. If the ground is necessary and the relation is grounding, the grounded is necessary. A necessary structural base cannot ground contingent content, because grounding transmits modal status. The argument has smuggled in contingency it has no right to.

Constitutive Grounding Non-Propagation · L9 · Lemma toward Modal Collapse Resistance

Two relations must be kept apart. □(T∧S∧Φ) and □◇C are not ground and grounded but one fact under two readings: by the Modal Identity, the necessity of the structural conditions just is the necessary possibility of change read modally, so □◇C carries its necessity by identity, not by transmission across a grounding relation. The structural conditions also ground the particular changes, but only in part: to be a change at all is to instantiate T, S, and Φ, so the roles enter every change as a necessary structural constituent, while what makes a given change occur rather than another, its initial conditions, its open quantum outcome (§7.7), lies outside them. The necessity of the conditions is compatible with the contingency of what is conditioned. □(T∧S∧Φ) ⊢ □◇C, not □C. (Modal Identity: §2.6.6)

Spinozan propagation holds only if T, S, and Φ are the full ground of each particular change: only if fixing the structural conditions fixes which changes occur. They are not. The roles ground the possibility of change and enter each actual change as a necessary structural constituent, but they are never the whole of what grounds a given change: which change occurs turns on contingent specifics (initial conditions, the open quantum outcome of §7.7) that the roles leave undetermined. Grant even the strongest grounding-necessitarian principle, that a full ground necessitates what it grounds; a merely partial ground does not, and T, S, and Φ are at most a partial, structural ground of any particular change. That T, S, and Φ are necessary entails any actual change instantiates them; it does not entail any particular change had to occur.

A kindred move presses through mereology. Schaffer’s²⁷³ priority monism holds that the cosmos is the fundamental object and all its parts derive from it by ontological grounding. If grounding transmits necessity, every part is necessary. The reply is the same in mereological vocabulary. Priority monism by itself does not decide whether the cosmos fully or only partly grounds each configuration of its parts. Blocking collapse requires the additional thesis that the necessary structural base underdetermines those configurations; it cannot be read off monism alone. Priority monism does not generate modal collapse once the partial/full ground distinction is drawn.

The Lewisian plurality, which proposes a different reduction of modality, is §4.4.1. Williamson’s necessitism, which gives the necessity claim its formal vocabulary, is §4.4.6. Sider’s structuralism, which deflates the necessity claim rather than embracing it, is §4.4.7. The Whiteheadian process alternative, which relocates Substance into verb form, is §4.4.8. The simultaneous-causation response, which grounds an atemporal production of the world, is §5.1.3.

What converges. Each of the three forms misreads what the argument establishes and what grounding does. The Consequence Argument targets determinism, which the present account is not committed to: □◇C fixes what is structurally possible, not which possibilities are actualized, and the structural roles do not by themselves fix laws, initial conditions, or outcomes, whichever quantum interpretation obtains. The Leibnizian best-world version presupposes a necessary evaluative subject the present account does not posit, and is, in fact, a sharper argument against classical theism than against the present account. Spinozan propagation holds only if T, S, and Φ fully ground each particular change rather than partly. The roles ground the possibility of change and enter each actual change as a necessary constituent, but the contingent specifics that fix which change occurs lie outside them, so a merely partial ground does not propagate modal status in the Spinozan sense.

Spinoza wrote the Ethics more geometrico (Latin: “in the geometrical manner”, his method of presenting metaphysics in the form of definitions, axioms, and propositions). The necessitarian conclusion he reached (God is nature, nature is necessary, every mode follows with the force of geometrical demonstration) has haunted rationalist metaphysics ever since. Most of his successors tried to restore contingency by loosening the grounding relation between substance and mode. The structural account restores it by a different line: by clarifying what was grounded in the first place. T, S, and Φ ground the possibility of change, not the necessity of any particular change. The stage may be necessary without the argument thereby making every detail of the play necessary. Spinoza saw what happens if the necessary base fully grounds its modes, so that necessity transmits unrestricted; he saw correctly. The mistake was not in following the logic. It was in assuming the base must be the full ground of what follows from it.

Necessity at the base permits contingency at the leaves.

4.4.10 Gasking’s Parody of the Ontological Argument

Gaunilo, a Benedictine monk of Marmoutiers, replied to Anselm’s Proslogion in the late eleventh century: the ontological argument’s form, if sound, would prove the existence of a perfect lost island (an island than which no more perfect could be conceived), and since no such island exists, the argument’s form cannot be sound. His parody was the first great demonstration that an argumentative form can be tested by seeing what else the form delivers. Nine centuries later, Douglas Gasking extended the tradition in a manuscript circulated in the late 1970s and published after his death: the greatest conceivable achievement, he argued, would be creating the universe while not existing, since creation from nothing without resources or powers is a greater feat than creation by an omnipotent being. If maximal greatness entails necessary existence in the Anselmian tradition, parody-greatness entails necessary non-existence. The parody is deliberately absurd. Its point is methodological: an argument whose form licenses its own reductio never really licenses its conclusion.

Gaunilo’s Pro Insipiente (c. 1078) set the template. Gasking’s “Proof of the Non-Existence of God” (commonly cited 1978; published posthumously in Language, Logic and Causation, 1996) modernized it. Mackie (The Miracle of Theism, 1982) and Oppy (Ontological Arguments and Belief in God, 1995; Arguing About Gods, 2006) developed the parody tradition into a general diagnostic: an argumentative form that delivers one impressive conclusion can usually be adapted to deliver its opposite, revealing that the form’s apparent productivity rests on hidden premises about greatness, predication, or modal reasoning. Plantinga (God, Freedom, and Evil, 1974a) responded that the modal ontological argument survives parody because the parody premises are not genuinely possibly instantiated in the way the original premises are. Grim (“There Is No Set of All Truths,” 1984) opened a different parody-adjacent front via Cantorian diagonalization (the powerset technique that, given any proposed totality, generates an element outside it) against omniscience. Adams (“Flavors, Colors, and God,” 1987) and van Inwagen (“Ontological Arguments,” 1977) canvassed the conditions under which a parody does and does not mirror its target. The tradition’s shared assumption is that arguments with modal conclusions are particularly susceptible to parody because modal space is abundant enough to contain mirror-image premises the original argument cannot exclude.

Brought to bear on the present argument, the parody tradition tests whether □(T∧S∧Φ) (the necessity of Time, Space, and Substance together) is earned by content or by form. If the argument’s machinery, the Modal Identity from §2.6.6, S5 / Necessary Possibility from §2.1 (the modal-logic system in which whatever is necessary or possible at any world is necessary or possible at every world), and the ∅ result from §2.7 (the null state, the absolute absence of being), could be redeployed to produce inverted or absurd conclusions, the machinery is suspect. A parody that successfully mirrored the argument’s form would deliver a structural result the argument cannot accept, revealing the necessity claim rests on hidden premises rather than on the form the argument presents. The question is whether the formal structure is neutral between its intended conclusion and a parody-conclusion, or whether the content of T, S, and Φ does work the form alone cannot carry.

A clarification before the moves. The argument’s conclusion is not ‘T, S, and Φ are great and therefore necessary.’ It is ‘T, S, and Φ are the structural requirements of any change, any extension, any persistence, and therefore no medium can be without them.’ The content is functional, not evaluative. A parody that substitutes an evaluatively defined alternative (‘maximal Non-Being,’ ‘greatest achievement of non-existence’) is not running the argument’s form with new content. It is running a different form entirely, one the present argument never uses. Three forms of the parody; each breaks at a point where the present argument differs from the Anselmian one.

First. The inversion move runs Gasking’s exact trick on the present argument. The greatest conceivable being, the objector presses, is one whose existence is not presupposed by the categories of existence: a being that grounds the structure without instantiating it. This is the exact move classical theists make for God. By parallel reasoning, there must be a maximally great Non-Being that grounds the structure while instantiating no structural role. If the ontological argument’s form is legitimate, the parody’s form is equally legitimate.

The parody fails at the point the original argument is made. Gasking’s greatest-achievement move trades on a category confusion the structural account explicitly rejects. Greatness in the ontological tradition is defined by impressiveness — by what it would be more impressive for a being to be. The structural argument does not work with this concept. It works with Functional Potency: the capacity to act, cause, and affect. Non-existence possesses zero potency. A non-existent cause cannot produce a manifest effect. ‘Necessary Non-Existence’ is not a form of greatness: it is the ultimate state of causal inertness, which is what ∅ names. The structural roles possess the full functional architecture that makes causal efficacy possible at all. The parody inverts the argument’s definition of greatness and then treats the inversion as a reductio. The inverted definition, not the argument, generates the absurdity. Swap ‘impressiveness’ for ‘potency’ and the parody cannot even begin: a non-existent ground is not more potent than a structural one. It is maximally impotent.

Second. The S5 symmetry move targets the logical machinery directly. The argument moves from ◇C to □◇C via S5, applied to change-permission. By parallel reasoning, ¬◇C should yield □¬◇C: the necessary impossibility of change. If the logical machinery licenses the first move, it licenses the second. But the second contradicts the first. The S5 system powers both, so the argument’s conclusion rests on a prior decision about which modal premise to start with.

The parallel fails because the two premises are not symmetric. ◇C (change is possible) is defended by the Cartesian certainty that change is occurring in the very act of doubt and the subsequent derivation that change requires T, S, and Φ. ¬◇C (change is impossible) has no equivalent step. There is no Cartesian certainty that change is impossible, because the certainty of doubting would itself refute the certainty of impossibility. The parody’s starting premise conflicts with the book’s Cartesian datum if doubting is correctly analyzed as change. That is a substantive premise about the act, not an analytic truth and not yet a freestanding cross-world proof. S5 preserves whichever modal premise is supplied; it does not establish the asymmetry by itself. The machinery is the same. The inputs are not. This is where the present reply converges with Plantinga’s (1974a): a modal argument survives parody only when the parody’s mirror premise is not genuinely possibly instantiated. On the book’s act analysis, ¬◇C is dialectically unstable because asserting or considering it occurs through an ordered cognitive performance. The force of that reply remains act-side and depends on the argument connecting the performance to real change.

Third. Oppy’s general indictment presses the crispest version. Any argument attempting to derive substantive ontological conclusions from conceptual premises is suspect in principle, because conception does not deliver existence. The structural argument is such an argument: it conceives change and derives structural necessity. The parody argues, equally validly, that it conceives non-existence and derives structural absence. Both conclusions rest on the same epistemic error: confusing what can be coherently thought with what must ontologically obtain.

The argument does not derive existence from conception. It derives structural conditions from the content of any possible thought. The Cartesian datum (that thought is occurring) is not a conceptual premise but a phenomenally given fact. The subsequent derivation proceeds from what any thought-event structurally requires, not from the conceptual content of any particular thought. Oppy’s indictment applies to arguments of the form ‘there is a concept of X; therefore X exists.’ The argument has a different form: ‘there is a phenomenally given fact F; F requires structural conditions T, S, Φ; therefore T, S, Φ obtain.’ The parody cannot be reformulated in the same way because there is no phenomenally given fact of non-existence: non-existence is not a fact that can be phenomenally given, because phenomenal givenness requires the very structural conditions non-existence is defined as lacking. The parody can be conceived. It cannot be given. Oppy’s indictment targets the wrong kind of argument.

A kindred move drops Gasking’s inversion for Grim’s Cantorian diagonalization. In its original form (Grim 1984) the target is omniscience: if there is a set of all truths, the powerset argument generates a truth outside it, so the set (and the omniscience that would survey it) is incoherent.

The Incomplete Universe (Grim 1991) presses the result much harder, generalizing it past omniscience to any totality of facts: if a world is conceived as a maximal consistent collection of truths, the same diagonalization shows there is no such collection, and the universe considered as a totality of facts is essentially open and incomplete.

This is the more aggressive form, and it bears on the present argument more directly than the omniscience version, since the exhaustion claim can look like just such a totality. Applied here rather than to theism, the Grim-style move argues: if T, S, and Φ exhaust necessary existence, there is a fact about that exhaustion standing outside the structure, and the exhaustion claim is incoherent.

The reply follows the same path against both forms. Facts about exhaustion are statements about the structural roles, made by reasoners instantiating the roles. The meta-fact is not outside the structure; it is within the structure, at a higher level of description. What the diagonalization defeats is the pretension of a totality of all truths surveyed from beyond the domain it totalizes; the exhaustion claim is not that. It is the narrower claim that exactly three structural conditions are necessary for any being, and its own meta-statements are themselves further instances of T, S, and Φ at work, not truths inspecting the structure from outside. The account never asserts a closed totality of facts; it asserts the necessity of the conditions under which any fact is stated at all.

The Lewisian modal realism the parody sometimes implicates is §4.4.1. The modal-collapse worry that the necessity claim generates absurd consequences is §4.4.9. Williamson’s necessitism, which embraces necessity at the type level rather than rejecting it, is §4.4.6. Sider’s structuralism, which deflates rather than embraces necessity, is §4.4.7. The Whiteheadian process alternative is §4.4.8. Modal fictionalism’s attempt to deflate the whole apparatus is §4.4.4. The chained-origin counterexamples, which press the modal logic itself rather than the argument’s form, are §4.4.11.

What converges. Each of the three parodies breaks at a point where the present argument differs from the Anselmian one. Gasking’s inversion trades on a category confusion the structural account rejects: greatness on the present view is Functional Potency rather than impressiveness, and necessary non-existence is maximal impotence rather than maximal greatness. The S5 symmetry move fails because the two premises are not symmetric: ◇C is defended by a Cartesian certainty that change is occurring, and ¬◇C is dialectically unstable because asserting or considering it occurs through an ordered cognitive performance. Oppy’s indictment targets arguments that move from concept to existence; the present argument moves from a phenomenally given fact of change to its structural conditions, and the parody cannot be reformulated in the same way because non-existence cannot be phenomenally given.

Gaunilo’s island vanishes beneath a wave the moment it is conceived: not because islands cannot be perfect, but because an argument that proves an island by its mere conceivability has confused what is contained in a concept with what is contained in the world. Gasking’s non-existent creator performs an identical trick, translated into the idiom of twentieth-century analytic theism. The argument survives both because it never performs the trick either parody is designed to expose. It does not move from the concept of change to the existence of structure. It moves from the phenomenal fact of change — the thinking now going on — to the structural conditions that make the thinking possible. Constructing the parody supplies an actual structured performance and therefore rebuts a denial aimed at the performance itself. It does not alone establish that every possible world realizes the same loaded types; that broader result depends on the separate modal and role-to-type arguments.

4.4.11 Chisholm’s Paradox: Chained Origins

Roderick Chisholm asked a question in 1967 about Adam and Noah (Chisholm 1967). Adam lived 930 years, Noah 950. Surely Adam could have lived a year longer, and Noah a year less: nothing in either man’s essence hangs on a single year. Grant the concession once and a world opens in which Adam lives 931 and Noah 949. Grant it again from there, and again, and the two lifespans trade places entirely — a world in which the man who lives 950 years is Adam and the man who lives 930 is Noah, which is to say a world in which each is the other. Every step was innocent. The destination is not. Chisholm’s Paradox is the ancestor of every later argument that possibility is chained: that what is possible depends on where the question is asked from, and that iterating small admissible variations carries one somewhere no single leap could have reached.

Hugh Chandler (1976) ran the structure on artifacts rather than patriarchs, and aimed it at exactly the machinery this book uses — his paper is titled “Plantinga and the Contingently Possible,” and the modal apparatus of Plantinga (1974b) turns on the same transfer principle the present argument turns on. A bicycle assembled from a given set of parts could have been assembled with one part swapped; from that world, another swap is admissible; and the chain terminates in an assembly too distant from the original to have been possible for the original bicycle. Nathan Salmon developed the sharpest version on Kripke’s table: the table made from a given hunk of wood H could have been made from the slightly different hunk H′, and a table made from H′ could have been made from H″ — which was not possible for the original table (N. Salmon 1981, 1989; the example descends from Kripke 1980). Salmon then drew the conclusion the others resisted. Metaphysical accessibility — the relation a world bears to those worlds that count as possible relative to it — is not transitive; so S4, the axiom licensing the collapse of a two-step modal chain into a single step, fails for the metaphysical operator, and the S5 transfer fails with it. Graeme Forbes answered by keeping the logic and locating the trouble in vagueness, first through degree-theoretic treatments and then through counterpart-theoretic ones (Forbes 1984, 1985). Lewis answered through the inconstancy of counterpart relations: on counterpart theory, no individual literally exists in more than one world, and its modal properties are read off the individuals elsewhere that most resemble it — and which resemblance relation is in play shifts with context (Lewis 1968, 1986). Timothy Williamson supplies the epistemicist option, on which such a series has a sharp cutoff that is unknowable rather than absent (Williamson 1994). Behind all of them stands Kripke’s origin essentialism, the thesis that a thing could not have come from a wholly different origin and still be that thing (Kripke 1980). The dispute is sixty years old and has produced no settled verdict.

Brought to bear on the present argument, the objection targets the logic directly rather than the ontology, which makes it more dangerous than most of this section’s entries. Necessary Possibility is established at §2.1.10: whatever is possible, its possibility is necessary (◇P → □◇P), with the bilateral companion that whatever is impossible, its impossibility is necessary (¬◇P → □¬◇P). It is packaged into S5 at §2.1.4; the move from the actuality of change to its necessary possibility at §2.3 runs on it, and so does the Modal Identity of §2.6.6. If possibility is chained, the transfer is unsound, and no amount of care about T, S, and Φ — Temporality, Spatiality, and Physicality, the three structural Conditions — repairs it. The objector here concedes the framework, concedes possible-worlds semantics, concedes origin essentialism, and produces cases. That is the hardest kind of objection to answer, because there is nothing exotic to reject.

A clarification before the moves, and a distinction the replies turn on. The argument’s ◇ means real possibility: membership in one fixed global domain W, the complete range of ways Reality permits (§1.2.1, §2.1). And two readings of any modal claim about a particular thing have to be kept apart. A de dicto claim is about a proposition — “it is possible that some table was made from different wood.” A de re claim is about a thing — “this table is such that it possibly was made from different wood.” The first says something about what the range of possibilities contains; the second says something about what a particular object could have survived. Three forms of the objection press at three different resolutions: the chained-origin move as Chisholm and Chandler state it, Salmon’s principled rejection of transitivity, and the counterpart-theoretic inconstancy version. Each reply below diagnoses the cases rather than denying them, which is the more demanding standard: the intuitions driving the series are genuine, and an account that merely contradicted them would have explained nothing.

First. The chained-origin move holds that the series of small variations is individually innocent and jointly unacceptable, so possibility must be relative to a starting point. Each link is possible relative to its predecessor; the endpoint is not possible relative to the origin; therefore what is possible varies with where the question is asked.

The reply begins by separating the two readings, because on one of them there is no case at all. Read de dicto, “a table made from slightly different wood” is among the least exotic possibilities ever entertained in philosophy: furniture workshops witness it by the thousand daily. Every such table is already among the ways Reality permits, and no act of carpentry puts it there — the carpenter never once consults the accessibility relation. The cases become cases only on the de re reading: could this very table have originated in H″ and still be it? That is a question about how much origin-swapping identity tolerates. It is a substantive and interesting question. It is not a question the transfer principle was ever answering, because the transfer ranges over the modal status of propositions and not over the persistence of a token individual across worlds. The distinction is the same one §4.4.1 draws against counterpart theory, running in the other direction: what the argument needs across worlds is instantiation of a structural role, never a numerically identical token recurring, and the chained-origin cases press exclusively on tokens.

Second. Salmon’s version blocks that reply by refusing the relabeling. His ◇, he insists, is the metaphysician’s own unrelativized operator and not a disguised accessibility relation; the cases show that operator to be non-transitive, and calling the variation “accessibility” redescribes the problem rather than answering it.

The reply takes the global question directly. Ask of the far link: is a table made from H″ among the ways Reality permits, or not? The question admits one answer. If yes, it was in W all along, and what varied along the chain was never membership but tolerable distance from a shifting anchor — a fact about the essentialist premise, not about the logic. If no, then no sequence of intermediate worlds puts it there, since traversing intermediates is not a way of enlarging the range. Either way, nothing has entered or left the domain.

The chain also cannot travel without paying. To move from H to H′ to H″, the argument must concede at every link that the intermediate origin is possible; that concession is what licenses the step. But a conceded possibility, run through the transfer, is a necessary possibility: each step certifies its own destination as a permanent member of the domain. Meanwhile no step removes anything, because unreachability is not eviction — that an origin cannot be reached from a given anchor is silent on whether it is in the range at all. So the chain deposits members and withdraws none. Pursued to its end, the argument has populated the global domain it set out to fragment, and the only thing still varying at the far link is the de re label: which of those necessarily possible tables counts as this one.

One further pressure, the difference-maker point of §2.1.5. If “possible at one world, impossible at another” states a fact about a single unrelativized operator, then something must make the difference between the two worlds. That something would have to concern what possibility itself answers to, and a fact of that kind is arrangement-independent and structural — which is precisely the cross-world stability the objection was contesting. The variation hypothesis consumes the resource it needs.

The candor the rest of this chapter keeps is owed here too. Salmon can plant on the disputed point and hold that the unrelativized operator simply is non-transitive, taking the cases as primitive data about it rather than as evidence to be diagnosed. Against a philosopher who does that, none of the above is a formal refutation, because its load-bearing premise is the fixed-domain reading he declines. What the reply secures is narrower and still worth having: the objection cannot be run without conceding, link by link, the very possibilities whose global standing is at issue, and the position it defends turns out to be a position about a relation between worlds rather than about the range those worlds jointly compose.

Third. The counterpart-theoretic version drops non-transitivity in favor of inconstancy. On Lewis’s treatment, a de re modal claim about the table is made true by what happens to the table’s counterparts — the objects in other worlds that resemble it most in whatever respects the conversation is weighting. Weight origin heavily and the H″-table is no counterpart of this table; weight function or appearance instead and it is. So “possibly made from H″” comes out true in one conversation and false in another, and any transfer principle stated over such claims inherits the instability.

The reply is that inconstancy of this kind is variation in representation, not variation in the domain. What shifts with context is which similarity respects a de re predication tracks — a fact about the semantics of counterpart talk. Nothing shifts in which worlds are among the ways Reality permits, and the whole plurality remains available to every context that draws on it. The inconstancy defense is also an awkward ally for the objector, since it concedes that the chained-origin data are produced by a context-sensitive representational device rather than discovered in the logic of possibility. A phenomenon that varies with conversational salience is not the sort of thing a modal axiom is answerable to.

A kindred move presses the vagueness directly rather than through counterpart semantics, and it is the most honest form of the objection. Each origin-swap is tolerable, the endpoints are not, and no step is the step at which identity lapses. That is a sorites — a chain of individually harmless steps whose endpoints differ in a way no single step seems to introduce, as in the grain-by-grain removal that turns a heap into a non-heap — run on essence rather than on heaps. Forbes proposes degrees of counterpart-hood; supervaluationism proposes that the claim be assessed on every admissible way of sharpening the vague predicate; Williamson’s epistemicism proposes a sharp cutoff that no one is in a position to locate. The present account takes no stand among the three, and does not need to: sorites series do not show that the predicates involved are world-relative. They show that the predicates are tolerant — that they survive small changes in their subject matter. Treating tolerance as evidence against a modal transfer principle mislocates the finding, though it is an interesting mislocation, and one worth the sixty years it has occupied. What the counterexamples establish is that de re origin essentialism gradates. Reading them as a refutation of the transfer requires one additional step: letting ◇ mean “reachable from the world of evaluation.” That is a legitimate operator with a genuinely non-S5 logic, and it is not the operator in which the conclusions of metaphysics are stated. When Salmon writes that the far origin was not possible for the table originally, that adverb is carrying the argument, and it is a word about starting points.

The status reading that makes this diagnosis available is fixed at §1.2.6 and §2.1.7: the claim is not that a necessity operator scopes over a possibility operator by climbing an accessibility relation, but that whatever is possible has its possibility as a settled status. The Bottleneck at §2.1.5 marks the single point at which a rival can disagree. Lewis’s concrete plurality, whose counterpart theory supplies the third move’s machinery, is §4.4.1. Williamson’s necessitism is §4.4.6. Sider’s structuralism is §4.4.7. The modal-collapse worry is §4.4.9. Gasking’s parody, which tests the argument’s form rather than its logic, is §4.4.10.

What converges. Each of the three forms turns on a relation between worlds rather than on the range the worlds compose. The chained-origin move is a case only on the de re reading, and the transfer never ranged over token identity. Salmon’s principled version asks a global question that admits one answer either way, and cannot advance its chain without certifying, at every link, a permanent member of the domain it means to fragment. The counterpart-theoretic version relocates the variation into context-sensitive representation, where it is no longer the sort of fact a modal axiom answers to. Underneath all three is a sorites on essence, and tolerance is not world-relativity.

Chisholm’s Adam and Noah do not trade places. What the series shows is that “same man” bears a great deal of counterfactual stretching without anyone being able to say where it snaps, which is a fact about the elasticity of identity under supposition and not a discovery about the range of ways Reality permits. The lifespans were always in that range. The question the chain makes vivid is only ever which of them we are still entitled to call Adam’s.

4.5 What the Rivals Carried

This is the chapter’s reckoning. The engagements do not all end with one verdict. Some rivals are internally incoherent on the book’s analysis. Some realize the same structural roles under another vocabulary. Some are compatible specifications of those roles. Some remain coherent only by taking on explanatory cost. Some do not contest the ground at all but the machinery used to reach it. Some identify a question the chapter does not settle. And one is no rival at all: it converges on the same modal floor and stops short of the structural conclusion. Keeping these verdicts distinct is part of the argument’s discipline.

Verdicts by rival type

1. Denial attempts

Null-state and domain-external absence. The proposal that nothing at all obtains dissolves rather than supplies a rival state: ∅ is not a world with unusual contents but the failure of the conditions under which a state or world could be specified (§2.7). This is an incoherence verdict.

Onset from absence. A first temporal order, first placement, or first causal production presupposes the domain it claims to originate (§2.1.5). This is also an incoherence verdict, but it is limited to onset from domain-external absence. It does not by itself prove that every possible world realizes the full Conditions.

Humean and deflationary accounts. Humean regularity, Best-System accounts, conventionalism, and modal deflation can preserve a real manifold, patterns, or local reachability. Where they do, they retain directed order, differentiation, or bearing at least at role-resolution. Their strongest forms are therefore ordinarily relocation verdicts, not contradictions. They remain live where they reject the book’s type-level or necessity claims (§4.1.4).

Strict temporal unreality. Two theses must be kept apart here. A thesis denying any ordering whatever does defeat the possibility of a world-state, but it does so by collapsing into the null-state case already treated above, not by any independent result about time. McTaggart’s own strict-unreality reading is weaker, and its verdict is weaker to match: the retorsion from 1908 diagnoses the act side, since the paper’s writing, its publication, and its century of reception deploy succession, distinction, and persistence in the course of denying them, but the object side is left untested, because an error theorist may still propose a non-temporal ground of the apparent order. The verdict is therefore act-side incoherence with the object-side error theory open, not incoherence outright — the shape of §3.3.3 rather than of §2.7. B-theory is neither thesis: it retains a temporal ordering and receives a different verdict below (§4.2.5).

2. Priority attempts

Changeless and non-temporal grounds. A ground that claims to produce, sustain, or answer to the changing order must stand in some real dependence relation to it. The chapter establishes that this relation cannot be wholly structureless. The full theological assessment of divine eternity, simplicity, and simultaneity belongs to Chapter 5. The verdict here is structural burden plus deferred theological engagement, not a complete refutation of every classical-theist account (§4.2.1).

Source-of-possibility accounts. No candidate whose own status is stated in modal terms can ground the global modal domain it already inhabits. A committed essence-prior or transcendence theorist may nevertheless posit a primitive external ground. The verdict is internal modal-ground exclusion and an open grounding gloss, not universal contradiction (§4.1.10).

Dispositionalism about modality. The nearest rival and the closest ally: like the present account, it locates modality in reality rather than in a plurality of concrete worlds, and its objections to Lewis are largely the book’s own. It grounds possibility in the powers of actual objects, and powers are borne. The global case has no bearer, and iterating potentiality does not supply one, since the possibility that there be bearers presupposes a bearer. The verdict is shared diagnosis of the rivals, with grounding that reaches every case but the one at issue (§4.1.12).

Lewisian plurality. Concrete Lewisian worlds must meet concrete spatiotemporal burdens internally. Their numerical plurality, if treated as primitive, remains a coherent but costly posit rather than a contradiction. The verdict is structured realization plus explanatory cost (§4.4.1).

Necessitism. Fixed domains of individuals are compatible with the fixed global domain of real possibilities. Necessitism neither refutes the structural conclusion nor automatically supplies its metaphysical content. The verdict is formal compatibility with a competing ontological reading (§4.4.6).

3. Insufficiency attempts

B-theory and block universe. The block realizes temporal ordering, differentiation, and cross-time structure. It is compatible with the role-level framework while remaining in dispute over objective passage. The verdict is compatibility at role-resolution, open disagreement about the passage feature (§4.2.6).

Whiteheadian process. Process ontology captures cross-transition efficacy without an enduring subject. It therefore realizes the bearer-role while disputing persistent causal being as a type. The verdict is relocation at role-resolution, type-level dispute deferred (§4.4.8).

Mathematical Universe Hypothesis and structural realism. Pure mathematical structures and abstract relational fabrics need not instantiate temporal, spatial, and causal types merely by being abstract structures. When proposed as physical reality, observer-hosting reality, or a genuine change-realizer, they incur the richer realizer burden. The verdict is conditional compatibility: abstract structures remain live, physical-realizer identity claims inherit structural requirements (§4.3.7, §4.3.9).

Hard problem of consciousness. Experience does not establish a structureless domain. Even irreducible qualitative features would be features of a structured ground. Whether phenomenal interiority is identical with structural self-modeling remains a constructive question for later chapters. The verdict is no escape from structure, positive explanatory seam left open (§4.3.1).

4. Apparatus attempts

A fourth kind of objection leaves the ground alone and goes after the machinery. It offers no rival world, no rival source, and no rival realizer; it argues that the semantics or the epistemology the argument runs on will not bear the weight placed on it. These verdicts are narrower than the others, and two of them cost the book something.

Impossible worlds and relative modality. Admitting impossible worlds multiplies representations rather than genuine alternatives, and representations are themselves structured. A Kratzer-style modal base restricts the domain rather than populating it, and every subset of the global domain has only members of the global domain. The hyperintensionality complaint is conceded in full: necessary equivalence is too coarse for many metaphysical distinctions, which cuts below the necessity claim without cutting into it. Non-normal worlds are the real exception, since invalidating necessitation is a genuine way to block the lift; the reply there is redirection to the non-S5 route of §3.5.4 rather than defense. The verdict is representational enlargement without domain enlargement, with one semantic route answered by redirection (§4.1.14).

Chained-origin counterexamples. The tolerance chaining that generates Chisholm’s paradox makes what is possible for an individual vary from world to world, which is genuine pressure on S4 and S5 and is not dismissed here. It reaches de re identity claims about contingent individuals; the structural claim is de dicto and quantifies over none. The verdict is live pressure on a scope the argument does not occupy (§4.4.11).

The remoteness objection. That the modal faculty grows unreliable at metaphysical distances is granted without reservation, and then declined as beside the point, because the method it impugns is not the method in use: the book’s negative results come from attempted specification and diagnosed failure, not from imaginative survey. The verdict is the faculty conceded, the method refused, and a standard incurred (§4.1.13).

5. Convergence

Pizzi’s impossibility of nothingness. Not a rival but the literature’s nearest cousin: the same closure of modal space against ∅, reached by kindred reductio machinery, ending at the bare-existential □∃x x rather than the structural □(T ∧ S ∧ Φ). The verdict is convergence at the modal floor, with the naming of the necessary content left to Chapter 5 (§4.1.11).

What the pattern establishes

The chapter’s positive finding is not that every objection secretly proves the full Necessity of Nature. It is more exact: no rival surveyed supplies a fully articulated alternative that both avoids the relevant structural roles and performs the work for which it was introduced. Some candidates dissolve. Many relocate roles. Several remain compatible specifications. A few remain coherent at a price or mark an explanatory question that the book carries forward.

The claim remains falsifiable. A genuine defeater would specify a real ground, world, or change-realizer that performs its stated work while instantiating none of the relevant roles, without merely appealing to primitive exemption or shifting to a different relative modality. Producing such a case would defeat the corresponding structural claim. None of the positions considered here has done so.

What counts as having failed to produce one is now fixed rather than left to judgment. The Specification Test (§4.1.13) holds that a failure to specify is evidence of impossibility only where the failure is diagnosed — where the attempt can be shown to deploy what it denies — and not where the attempt merely stalls. The test binds this ledger as much as it binds the rivals: any verdict above that rests on a failure encountered rather than exhibited is owed a downgrade, not a defense.

The chapter therefore ends without a victory lap. It leaves a ledger: structural absence fails where it is supposed to describe a world; relocation does not displace the roles; compatibility does not amount to refutation; explanatory cost is not incoherence; a restriction cannot manufacture a member, and a doubt about the faculty is not a defeat of the method; convergence at the floor is not the structural conclusion; and open seams remain open until the chapters that bear them do their work.

---

1. Ontological Arguments and Belief in God (Cambridge University Press, 1995); Arguing about Gods (Cambridge University Press, 2006); The Best Argument Against God (Palgrave Macmillan, 2013).

2. Naming and Necessity (Harvard University Press, 1980), Lecture I.

3. “The Meaning of ’Meaning’,” in Minnesota Studies in the Philosophy of Science 7: 131–193 (University of Minnesota Press) (1975), https://hdl.handle.net/11299/185225.

4. “Essence and Modality,” Philosophical Perspectives, 1994, https://doi.org/10.2307/2214160.

5. Alan Sidelle, Necessity, Essence, and Individuation: A Defense of Conventionalism (Cornell University Press, 1989); Ross P. Cameron, “On the Source of Necessity,” in In Bob Hale & Aviv Hoffmann (Eds.), Modality: Metaphysics, Logic, and Epistemology (Oxford University Press), Pp. 137–152 (2010), https://academic.oup.com/book/11764/chapter/160788905; Theodore Sider, Writing the Book of the World (Oxford University Press, 2011).

6. The Conscious Mind: In Search of a Fundamental Theory (Oxford University Press, 1996), https://global.oup.com/academic/product/the-conscious-mind-9780195117899; The Character of Consciousness (Oxford University Press, 2010), https://academic.oup.com/book/6996.

7. From Metaphysics to Ethics: A Defence of Conceptual Analysis (Clarendon Press / Oxford University Press (Oxford), 1998).

8. “Conceptual Analysis and Reductive Explanation,” The Philosophical Review 110 (3): 315–360, 2001, https://doi.org/10.1215/00318108-110-3-315.

9. Constructing the World (Oxford University Press, 2012), https://global.oup.com/academic/product/constructing-the-world-9780199608577.

10. “Does Conceivability Entail Possibility?” in In Conceivability and Possibility (Eds. Gendler & Hawthorne), Oxford University Press (2002), https://academic.oup.com/book/49672/chapter/422537279.

11. “Does Conceivability Entail Possibility?”

12. Chalmers, Constructing the World.

13. “Conceptual Analysis and Reductive Explanation.”

14. On the Plurality of Worlds (Blackwell, 1986).

15. Constructing the World.

16. “Essence and Modality.”

17. “Indexicality and Actuality,” Philosophical Review 89(3): 403–426, 1980, https://doi.org/10.2307/2184396.

18. “Indexicality and Actuality,” 416.

19. “Theories of Actuality,” Noûs 8(3): 211–231, 1974, https://doi.org/10.2307/2214751.

20. “Conceptual Role Semantics,” in Routledge Encyclopedia of Philosophy (Online Edn 1998), 1998, https://doi.org/10.4324/9780415249126-W037-1.

21. “(Nonsolipsistic) Conceptual Role Semantics,” in In e. LePore (Ed.), New Directions in Semantics (Academic Press), Pp. 55–81 (1987), https://philpapers.org/rec/HARNCR.

22. “Monadology,” in Philosophical Essays, Translated by r. Ariew and d. Garber. Indianapolis: Hackett ((1714) 1989), https://philpapers.org/rec/ARIGWL.

23. The Principle of Sufficient Reason: A Reassessment (Cambridge University Press, 2006).

24. The Principle of Sufficient Reason.

25. Necessary Existence (Oxford University Press, 2018), https://doi.org/10.1093/oso/9780198746898.001.0001.

26. Edward Feser, Five Proofs of the Existence of God (Ignatius Press, 2017), https://ignatius.com/five-proofs-of-the-existence-of-god-fpegp/; Robert C. Koons, “A New Look at the Cosmological Argument,” American Philosophical Quarterly 34 (2): 171–192, 1997, https://www.jstor.org/stable/20009892; Joshua Rasmussen, How Reason Can Lead to God: A Philosopher’s Bridge to Faith (IVP Academic, 2019).

27. An Enquiry Concerning Human Understanding (Oxford University Press, (1748) 1999), sec. VII.

28. On the Plurality of Worlds, ix–xvi.

29. “Humean Supervenience,” Philosophical Topics 24 (1): 101–127, 1996, https://doi.org/10.5840/philtopics199624112.

30. Hume on Causation (Routledge, 2006), https://doi.org/10.4324/9780203966600.

31. Real Time II (Routledge, 1998).

32. The Metaphysics Within Physics (Oxford University Press, 2007).

33. The Metaphysics Within Physics.

34. Lewis, On the Plurality of Worlds; Loewer, “Humean Supervenience.”

35. D. M. Armstrong, What Is a Law of Nature? (Cambridge University Press, 1983), https://doi.org/10.1017/CBO9781139171700, What Is a Law of Nature?

36. F. I. Dretske, “Laws of Nature,” Philosophy of Science 44 (2): 248–268, 1977, https://doi.org/10.1086/288741; M. Tooley, “The Nature of Laws,” Canadian Journal of Philosophy 7 (4): 667–698, 1977, https://doi.org/10.1080/00455091.1977.10716190.

37. Lewis, On the Plurality of Worlds, ix–xvi.

38. “On the Notion of Cause,” Proceedings of the Aristotelian Society 13: 1–26, 1913, https://doi.org/10.1093/aristotelian/13.1.1, “On the Notion of Cause.”

39. “Causation as Folk Science,” Philosophers’ Imprint 3 (4): 1–22, 2003, https://doi.org/10.3998/3521354.0003.004.

40. G. E. M. Anscombe, Causality and Determination (Cambridge University Press, 1971), https://search.worldcat.org/title/Causality-and-determination:-an-inaugural-lecture./oclc/844685222, “Causality and Determination”; James Woodward, Making Things Happen: A Theory of Causal Explanation (Oxford University Press, 2003), Making Things Happen.

41. “A Mathematical Theory of Communication,” Bell System Technical Journal 27 (3): 379–423; (4): 623–656, 1948, https://doi.org/10.1002/j.1538-7305.1948.tb01338.x.

42. “Irreversibility and Heat Generation in the Computing Process,” IBM Journal of Research and Development 5 (3): 183–191, 1961, https://doi.org/10.1147/rd.53.0183.

43. J. A. Wheeler, “Information, Physics, Quantum: The Search for Links,” in In Complexity, Entropy, and the Physics of Information, Edited by w. H. Zurek, 3–28. Redwood City, CA: Addison-Wesley (1990), https://philarchive.org/rec/WHEIPQ.

44. Steps to an Ecology of Mind (Ballantine Books, 1972), https://search.worldcat.org/title/3162139.

45. “A Mathematical Theory of Communication.”

46. M. Tegmark, Our Mathematical Universe: My Quest for the Ultimate Nature of Reality (Knopf, 2014).

47. Our Mathematical Universe.

48. Programming the Universe: A Quantum Computer Scientist Takes on the Cosmos (Knopf, 2006).

49. The Fabric of Reality (Viking, 1997), https://search.worldcat.org/title/263628359.

50. A New Kind of Science (Wolfram Media, 2002).

51. “Irreversibility and Heat Generation in the Computing Process.”

52. Every Thing Must Go: Metaphysics Naturalized (Oxford University Press, 2007).

53. “Two Dogmas of Empiricism,” The Philosophical Review 60 (1), 1951, https://doi.org/10.2307/2181906.

54. “Reference and Modality,” in In from a Logical Point of View (Cambridge, MA: Harvard University Press), 139–159 (1953).

55. Word and Object (MIT Press, 1960).

56. “Essentialism and Reference,” in In the Philosophy of w. V. Quine (Library of Living Philosophers, Open Court) (1986).

57. “Quinus Ab Omni Naevo Vindicatus,” Canadian Journal of Philosophy, Supplementary Volume 23, 1997, https://doi.org/10.1080/00455091.1997.10715961.

58. “Contingent Identity,” Journal of Philosophical Logic 4 (2): 187–221, 1975, https://doi.org/10.1007/BF00693273.

59. Modal Logic as Metaphysics (Oxford University Press, 2013), ch. 2.

60. “Contingent Identity.”

61. “Reference and Modality.”

62. “Essentialism and Reference.”

63. “Quinus Ab Omni Naevo Vindicatus.”

64. Modal Logic as Metaphysics, ch. 2.

65. “Natural Kinds,” in In Ontological Relativity and Other Essays (New York: Columbia University Press), 114–138 (1969), https://philpapers.org/rec/QUINK.

66. Modal Logic as Metaphysics, ch. 2.

67. Modal Logic as Metaphysics.

68. Peter van Inwagen, “Modal Epistemology,” Philosophical Studies 92 (1): 67-84, 1998.

69. Necessity, Essence, and Individuation.

70. Language, Truth and Logic (Victor Gollancz (London), 1936), https://philpapers.org/rec/AYELTA-2.

71. “Empiricism, Semantics, and Ontology,” Revue Internationale de Philosophie 4 (11): 20–40, 1950, https://philpapers.org/rec/CARESA.

72. “Two Dogmas of Empiricism.”

73. “Modality and Objects,” The Philosophical Quarterly 60 (238): 109–125, 2010, https://doi.org/10.1111/j.1467-9213.2008.604.x.

74. “On the Source of Necessity.”

75. Norms and Necessity (Oxford University Press, 2020).

76. Necessary Beings: An Essay on Ontology, Modality, and the Relations Between Them (Oxford University Press, 2013), https://global.oup.com/academic/product/necessary-beings-9780198748038.

77. “Laws for Metaphysical Explanation,” Philosophical Issues 27(1): 302–321, ahead of print, 2017, https://doi.org/10.1111/phis.12111.

78. Necessity, Essence, and Individuation.

79. Necessary Beings, ch. 4.

80. “On the Source of Necessity.”

81. Norms and Necessity.

82. “Laws for Metaphysical Explanation.”

83. “Empiricism, Semantics, and Ontology.”

84. Quantifier Variance and Realism: Essays in Metaontology (Oxford University Press (New York), 2011).

85. Ontology Made Easy (Oxford University Press, 2015).

86. “Ontological Anti-Realism,” in In Chalmers, Manley & Wasserman (Eds.), Metametaphysics: New Essays on the Foundations of Ontology (Oxford University Press), Pp. 77–129 (2009), https://academic.oup.com/book/50169/chapter/422047610.

87. “On Carnap’s Views on Ontology,” Philosophical Studies 2(5): 65–72, ahead of print, 1951, https://doi.org/10.1007/BF02199422.

88. “On What There Is,” Review of Metaphysics 2 (5): 21–38, 1948, https://www.jstor.org/stable/20123117.

89. Ontology Made Easy.

90. “Ontological Anti-Realism.”

91. “The Picture of Reality as an Amorphous Lump,” in In Theodore Sider, John Hawthorne and Dean w. Zimmerman (Eds.), Contemporary Debates in Metaphysics, Malden, MA: Blackwell, 2008: 382–396 (2008), https://philpapers.org/rec/EKLTPO.

92. Ontology Made Easy, ch. 3.

93. Tractatus Logico-Philosophicus (Translated by D. F. Pears; B. F. McGuinness. London: Routledge, (1921) 1961).

94. The Realistic Spirit: Wittgenstein, Philosophy, and the Mind (MIT Press (Cambridge, MA), Bradford Books / Representation; Mind series, 1991), https://doi.org/10.7551/mitpress/5797.001.0001.

95. “The Method of the Tractatus,” in In e. H. Reck (Ed.), from Frege to Wittgenstein: Perspectives on Early Analytic Philosophy (Oxford University Press), Pp. 374–462 (2002), https://doi.org/10.1093/0195133269.003.0015.

96. With Baker, Wittgenstein: Understanding and Meaning (Analytical Commentary on the Philosophical Investigations, Vol. 1) (Blackwell, 1980), https://philpapers.org/rec/BAKWUA-16; Wittgenstein: Rules, Grammar and Necessity (Analytical Commentary on the Philosophical Investigations, Vol. 2) (Basil Blackwell, 1985), https://philpapers.org/rec/BAKWRG; Hacker, Wittgenstein: Connections and Controversies (Clarendon Press / Oxford University Press (Oxford), 2001), https://global.oup.com/academic/product/wittgenstein-connections-and-controversies-9780199271139.

97. Mind and World (Harvard University Press, 1994).

98. The Claim of Reason: Wittgenstein, Skepticism, Morality, and Tragedy (Clarendon Press (Oxford) / Oxford University Press (New York), 1979), https://global.oup.com/academic/product/the-claim-of-reason-9780195131079.

99. Hale, Necessary Beings.

100. Lewis, On the Plurality of Worlds.

101. Williamson, Modal Logic as Metaphysics.

102. Barbara Vetter, Potentiality: From Dispositions to Modality (Oxford University Press, 2015).

103. R. M. Adams, “Divine Necessity,” Journal of Philosophy 80 (11): 741–752, 1983, https://philpapers.org/rec/ADADN; A. R. Pruss, Actuality, Possibility, and Worlds (Continuum, 2011); Brian Leftow, God and Necessity (Oxford University Press, 2012).

104. Fine, “Essence and Modality.”

105. Williamson, Modal Logic as Metaphysics.

106. Kit Fine, “Guide to Ground,” in In Metaphysical Grounding: Understanding the Structure of Reality, Edited by Fabrice Correia and Benjamin Schnieder, 37–80. Cambridge: Cambridge University Press (2012), https://www.cambridge.org/core/books/metaphysical-grounding/guide-to-ground/3C2135AA04D01F8821602F4667F56C1C.

107. Jonathan Schaffer, “On What Grounds What,” in In Metametaphysics: New Essays on the Foundations of Ontology, Edited by David j. Chalmers, David Manley, and Ryan Wasserman, 347–383. Oxford: Oxford University Press (2009), https://academic.oup.com/book/50169/chapter/422047737; Fine, “Guide to Ground.”

108. C. S. Jenkins, “Is Metaphysical Dependence Irreflexive?” The Monist 94 (2): 267–276, 2011, https://doi.org/10.5840/monist201194213.

109. God and Necessity.

110. David Pizzi, “The Impossibility of Nothingness: A Foundational Argument for Necessary Existence,” 2025, https://philarchive.org/rec/PIZTIO-2.

111. Robert Merrihew Adams, Finite and Infinite Goods: A Framework for Ethics (Oxford University Press, 1999), https://doi.org/10.1093/0195153715.001.0001.

112. Pruss and Rasmussen, Necessary Existence; Pizzi, “The Impossibility of Nothingness.”

113. Pruss and Rasmussen, Necessary Existence.

114. Rasmussen, How Reason Can Lead to God.

115. On the Plurality of Worlds.

116. Potentiality.

117. “A Dispositional Theory of Possibility,” Dialectica 62 (1): 21-41, 2008.

118. “A Powers Theory of Modality: Or, How i Learned to Stop Worrying and Reject Possible Worlds,” Philosophical Studies 151 (2): 227-248, 2010.

119. Actuality, Possibility, and Worlds.

120. Nature’s Metaphysics: Laws and Properties (Oxford University Press, 2007), https://doi.org/10.1093/acprof:oso/9780199227013.001.0001.

121. Laws in Nature (Routledge, 2004).

122. Getting Causes from Powers (Oxford University Press, 2011).

123. “The Modal Limits of Dispositionalism,” Nous 49 (3): 454-469, 2015.

124. Potentiality.

125. Modal Logic as Metaphysics.

126. “Modal Epistemology.”

127. “Conceivability and de Re Modal Knowledge,” Nous 45 (1): 22-49, 2011.

128. “Van Inwagen’s Modal Skepticism,” Philosophical Studies 153 (3): 351-364, 2011.

129. The Philosophy of Philosophy (Blackwell, 2007), ch. 5.

130. Graham Priest, Towards Non-Being: The Logic and Metaphysics of Intentionality (Oxford University Press, 2005); Berto, Francesco; Jago, Mark, Impossible Worlds (Oxford University Press, 2019); Daniel Nolan, “Impossible Worlds: A Modest Approach,” Notre Dame Journal of Formal Logic 38 (4): 535-572, 1997.

131. Angelika Kratzer, Modals and Conditionals: New and Revised Perspectives (Oxford University Press, 2012).

132. “Impossible Worlds.”

133. Impossible Worlds.

134. Kratzer, Modals and Conditionals.

135. Summa Theologiae (Benziger Brothers, (1265–1274) 1947).

136. The Consolation of Philosophy (Penguin, (524) 1969), https://search.worldcat.org/title/328107608.

137. The Nature of Necessity (Oxford University Press, 1974); Warranted Christian Belief (Oxford University Press, 2000).

138. “Eternity,” The Journal of Philosophy 78(8): 429–458, ahead of print, 1981, https://doi.org/10.2307/2026047.

139. Time and Eternity (Cornell University Press (Ithaca, NY & London), 1991).

140. The End of Time: The Next Revolution in Physics (Weidenfeld & Nicolson (London), 1999), https://philpapers.org/rec/BARTEO-12.

141. “The Unreality of Time,” Mind 17 (68): 457–474, 1908, https://doi.org/10.1093/mind/XVII.4.457.

142. Time and Eternity.

143. “Time Without Change,” The Journal of Philosophy 66(12): 363–381, ahead of print, 1969, https://doi.org/10.2307/2023892.

144. Aristotle, The Complete Works of Aristotle (Princeton University Press, (4th c. BCE) 1984), IV.10–14, https://press.princeton.edu/books/hardcover/9780691016504/the-complete-works-of-aristotle-volume-one.

145. Philosophiæ Naturalis Principia Mathematica (Jussu Societatis Regiae (Royal Society), 1687).

146. “The Unreality of Time.”

147. “Changes in Events and Changes in Things,” in In Papers on Time and Tense (Oxford: Clarendon Press), 1–14 (1968), https://academic.oup.com/book/50182/chapter/422051860.

148. The Structure of Time (Routledge & Kegan Paul, 1980).

149. Change, Cause, and Contradiction (Macmillan (London) / St. Martin’s Press (New York), 1991).

150. Events: A Metaphysical Study (Routledge & Kegan Paul (London & Boston), 1986).

151. “Time, Events, and Modality,” in In Robin Le Poidevin and Murray MacBeath (Eds.), the Philosophy of Time (Oxford Readings in Philosophy). Oxford: Oxford University Press (1993), https://philpapers.org/rec/FORTEA-5.

152. Real Time II.

153. The Tensed Theory of Time: A Critical Examination (Kluwer Academic (Synthese Library, vol. 293), Dordrecht, 2000), https://doi.org/10.1007/978-94-015-9345-8.

154. Asymmetries in Time: Problems in the Philosophy of Science (MIT Press / Bradford Books (Cambridge, MA), 1987).

155. Four-Dimensionalism: An Ontology of Persistence and Time (Clarendon Press (Oxford University Press), 2001).

156. Time and Space (Acumen (Durham) / McGill-Queen’s University Press, 2nd ed., 2010), https://doi.org/10.1017/UPO9781844654437.

157. A Brief History of the Philosophy of Time (Oxford University Press, 2013), https://global.oup.com/academic/product/a-brief-history-of-the-philosophy-of-time-9780199301089.

158. Philosophiæ Naturalis Principia Mathematica.

159. The Structure of Time.

160. The reply that absolute-time substantivalism relocates the structural roles rather than eliminating them is not neutral between conceptions of time, and the dependency is worth marking. A purist who takes time to be bare duration (mere ordering, with influence occurring within time rather than constituting it) can decline the identification of a moment’s individuation with T ∧ S ∧ Φ. Two costs attend the exit, and the relational account pays neither. First, individuating the moments of an empty, changeless absolute time without relational extension requires primitive thisness: the haecceity Leibniz’s identity of indiscernibles and the hole-argument literature press, which sophisticated substantivalism can absorb only by surrendering haecceitism; the relational individuation §2.6 supplies (S) needs no such posit. Second, by the lights of the change-theoretic conception defended here (a cost the purist will not grant is one) a duration stripped of all influence is ordering without succession, the §2.6.7 ordered-set structure rather than Time. The line can be held, but at the price of primitive identity and the divorce of time from succession: a trade the relocation reply exposes rather than the neutral refutation it can be mistaken for.

161. Le Poidevin, Change, Cause, and Contradiction; earlier Ayer, Language, Truth and Logic.

162. “Evolution Without Evolution: Dynamics Described by Stationary Observables,” Physical Review D 27(12): 2885–2892, ahead of print, 1983, https://doi.org/10.1103/PhysRevD.27.2885.

163. The End of Time.

164. The Order of Time (Riverhead Books, 2018).

165. Time Reborn: From the Crisis in Physics to the Future of the Universe (Houghton Mifflin Harcourt, 2013).

166. Time and Chance (Harvard University Press, 2000), https://doi.org/10.4159/9780674020139.

167. Time’s Arrow and Archimedes’ Point: New Directions for the Physics of Time (Oxford University Press, 1996).

168. The Direction of Time (University of California Press, 1956).

169. The Metaphysics Within Physics.

170. The Problem of Time: Quantum Mechanics Versus General Relativity (Springer (Fundamental Theories of Physics), 2017), https://doi.org/10.1007/978-3-319-58848-3.

171. Die Grundlagen Der Arithmetik (Verlag von Wilhelm Koebner (Breslau), 1884), https://search.worldcat.org/title/13477085.

172. “Russell’s Mathematical Logic,” in In p. A. Schilpp (Ed.), the Philosophy of Bertrand Russell (Library of Living Philosophers, Vol. 5), Northwestern University, Pp. 123–153 (1944); “What Is Cantor’s Continuum Problem? (Revised),” in In p. Benacerraf & h. Putnam (Eds.), Philosophy of Mathematics: Selected Readings (Prentice-Hall), Pp. 258–273 (1964), https://www.cambridge.org/core/books/philosophy-of-mathematics/what-is-cantors-continuum-problem/D60596C86B5B9B99014D3CB54DF45ABD.

173. Philosophy of Logic (Harper & Row, 1971); Mathematics, Matter and Method: Philosophical Papers, Volume 1 (Cambridge University Press (2nd ed.), 1979), https://doi.org/10.1017/CBO9780511625268.

174. Philosophy of Mathematics: Structure and Ontology (Oxford University Press, 1997).

175. Defending the Axioms: On the Philosophical Foundations of Set Theory (Oxford University Press, 2011).

176. Naturalism in Mathematics (Clarendon Press / Oxford University Press (Oxford), 1997).

177. Platonism and Anti-Platonism in Mathematics (Oxford University Press, 1998), https://global.oup.com/academic/product/platonism-and-anti-platonism-in-mathematics-9780195122305.

178. Ludwig Wittgenstein, Philosophical Investigations (Blackwell (trans. G. E. M. Anscombe), 1953), secs. 65–71.

179. John Dupré, The Disorder of Things: Metaphysical Foundations of the Disunity of Science (Harvard University Press (Cambridge, MA), 1993), https://search.worldcat.org/title/25746325.

180. Richard Boyd, “Realism, Anti-Foundationalism and the Enthusiasm for Natural Kinds,” Philosophical Studies 61 (1/2): 127–148, 1991, https://doi.org/10.1007/BF00385837.

181. Process and Reality (Macmillan, 1929).

182. Creative Evolution (l’évolution Créatrice) (Félix Alcan (Paris), 1907).

183. Lewis, On the Plurality of Worlds.

184. Frege, Die Grundlagen Der Arithmetik, sec. 63.

185. An Examination of McTaggart’s Philosophy (Cambridge University Press, 1938), https://search.worldcat.org/title/182775429.

186. Past, Present and Future (Oxford University Press (Clarendon), 1967).

187. Real Time (Cambridge University Press, 1981).

188. Real Time II.

189. Time, Tense, and Causation (Clarendon Press, 1997).

190. “A Defence of McTaggart’s Proof of the Unreality of Time,” The Philosophical Review 69 (4): 497–504, 1960, https://doi.org/10.2307/2183483.

191. Time and Space.

192. “The River of Time,” Mind 58(232): 483–494, ahead of print, 1949, https://doi.org/10.1093/mind/LVIII.232.483.

193. Philosophy and Scientific Realism (Routledge & Kegan Paul, 1963).

194. “The Myth of Passage,” The Journal of Philosophy 48(15): 457–472, ahead of print, 1951, https://doi.org/10.2307/2021694.

195. Real Time.

196. Real Time II.

197. Four-Dimensionalism.

198. “Time and Physical Geometry,” The Journal of Philosophy 64(8): 240–247, ahead of print, 1967, https://doi.org/10.2307/2024493.

199. “A Rigorous Proof of Determinism Derived from the Special Theory of Relativity,” Philosophy of Science 33(4): 341–344, ahead of print, 1966, https://doi.org/10.1086/288106.

200. The Metaphysics Within Physics.

201. Time’s Arrow and Archimedes’ Point.

202. “How to Teach Special Relativity,” in Speakable and Unspeakable in Quantum Mechanics (Cambridge University Press, 1987); Orig. Progress in Scientific Culture 1 (1976) (1976), https://www.cambridge.org/core/books/speakable-and-unspeakable-in-quantum-mechanics/how-to-teach-special-relativity/74F7023E4F9101356459BDC6069B804B.

203. Quantum Non-Locality and Relativity (Blackwell (Malden, MA), 2nd ed., 2002), ch. 7.

204. Time and the Metaphysics of Relativity (Kluwer Academic Publishers, 2001), https://doi.org/10.1007/978-94-017-3532-2.

205. Howard Stein, “On Einstein–Minkowski Space–Time,” The Journal of Philosophy 65(1): 5–23, ahead of print, 1968, https://doi.org/10.2307/2024512; Howard Stein, “On Relativity Theory and Openness of the Future,” Philosophy of Science 58(2): 147–167, ahead of print, 1991, https://doi.org/10.1086/289609.

206. “Presentism and Relativity,” The British Journal for the Philosophy of Science 54 (2): 327–346, 2003, https://doi.org/10.1093/bjps/54.2.327.

207. On the Plurality of Worlds.

208. Four-Dimensionalism.

209. How Things Persist (Clarendon Press / Oxford University Press (Oxford), 2001), https://doi.org/10.1093/acprof:oso/9780199275434.001.0001.

210. On the Plurality of Worlds.

211. Philosophical Papers, Volume II (Oxford University Press (New York & Oxford), 1986).

212. Word and Object.

213. The Ontology of Physical Objects: Four-Dimensional Hunks of Matter (Cambridge University Press, 1990), https://doi.org/10.1017/CBO9781139166409.

214. A Materialist Metaphysics of the Human Person (Cornell University Press (Ithaca, NY), 2001).

215. Material Beings (Cornell University Press, 1990), https://doi.org/10.7591/9781501713033.

216. Sameness and Substance (Blackwell, 1980).

217. The Possibility of Metaphysics: Substance, Identity, and Time (Oxford University Press, 1998).

218. Four-Dimensionalism, §4.3.

219. The Art and Thought of Heraclitus (Cambridge University Press, 1979).

220. Summa Theologiae, I.104.1.

221. De Potentia (In On the Power of God, translated by E. English. Westminster, MD: Newman Press, (1265–1266) 1932), q.5 a.1, https://search.worldcat.org/title/On-the-power-of-God-(Quaestiones-disputatae-de-potentia-Dei)/oclc/965714.

222. “God’s General Concurrence with Secondary Causes: Why Conservation Is Not Enough,” Philosophical Perspectives 5: 553–585, 1991, https://doi.org/10.2307/2214109.

223. “God’s General Concurrence with Secondary Causes: Pitfalls and Prospects,” American Catholic Philosophical Quarterly 68 (2): 131–156, 1994, https://doi.org/10.5840/acpq199468224.

224. Aquinas: A Beginner’s Guide (Oneworld, 2009), https://search.worldcat.org/title/421814088.

225. Five Proofs of the Existence of God.

226. “The World’s Continuance: Divine Conservation or Existential Inertia?” International Journal for Philosophy of Religion 61 (2): 83–98, 2007, https://doi.org/10.1007/s11153-007-9113-1.

227. “Existential Inertia and the Aristotelian Proof,” International Journal for Philosophy of Religion 89 (3): 201–220, 2021, https://philpapers.org/rec/SCHEIA-20.

228. Existential Inertia and Classical Theistic Proofs (Springer (Cham), 2022), https://doi.org/10.1007/978-3-031-19313-2.

229. “Divine Conservation and the Persistence of the World,” in In Thomas v. Morris (Ed.), Divine and Human Action: Essays in the Metaphysics of Theism, Ithaca, NY: Cornell University Press, Pp. 13–49 (1988), https://philpapers.org/rec/KVADCA.

230. Summa Theologiae, I.104.1.

231. “God’s General Concurrence with Secondary Causes,” 1991.

232. “God’s General Concurrence with Secondary Causes,” 1994.

233. Existential Inertia and Classical Theistic Proofs.

234. “The World’s Continuance.”

235. Eliot Deutsch, Advaita Vedānta: A Philosophical Reconstruction (University of Hawaii Press, 1969), https://philpapers.org/rec/DEUAVA-5.

236. Robert K. C. Forman, The Problem of Pure Consciousness: Mysticism and Philosophy (Oxford University Press, 1990), https://academic.oup.com/book/50260.

237. George Berkeley, A Treatise Concerning the Principles of Human Knowledge (Aaron Rhames, for Jeremy Pepyat (1710), 1710).

238. George Berkeley, Three Dialogues Between Hylas and Philonous (Printed by G. James for Henry Clements (1713), 1713).

239. I. Kant, Critique of Pure Reason (Translated by P. Guyer; A. W. Wood. Cambridge: Cambridge University Press, (1781) 1998).

240. Barry Stroud, “Transcendental Arguments,” Journal of Philosophy, 1968, https://doi.org/10.2307/2024395.

241. Rae Langton, Kantian Humility: Our Ignorance of Things in Themselves (Oxford University Press, 1998).

242. Edmund Husserl, Logical Investigations (M. Niemeyer (2 vols.; Eng. trans. J. N. Findlay, Routledge & Kegan Paul, 1970), 1900–01).

243. Edmund Husserl, Ideas Pertaining to a Pure Phenomenology (Ideas i) (M. Niemeyer (Ideen I; Eng. trans. Boyce Gibson 1931 / Kersten 1983), 1913).

244. M. Heidegger, Being and Time (Translated by J. Macquarrie; E. Robinson. New York: Harper & Row, (1927) 1962), https://search.worldcat.org/title/76901440.

245. Maurice Merleau-Ponty, Phenomenology of Perception (Gallimard (Phénoménologie de la perception; Eng. trans. Colin Smith, Routledge & Kegan Paul, 1962), 1945).

246. Bernardo Kastrup, Why Materialism Is Baloney (Iff Books (Winchester, UK), 2014).

247. Bernardo Kastrup, The Idea of the World: A Multi-Disciplinary Argument for the Mental Nature of Reality (Iff Books (Winchester, UK), 2019).

248. Bernardo Kastrup, Analytic Idealism in a Nutshell (Iff Books (Winchester, UK), 2024).

249. Philip Goff, Galileo’s Error: Foundations for a New Science of Consciousness (Pantheon Books (New York), 2019), https://search.worldcat.org/title/1089841262.

250. Philip Goff, Consciousness and Fundamental Reality (Oxford University Press (New York), Philosophy of Mind series, 2017), https://doi.org/10.1093/oso/9780190677015.001.0001.

251. David J. Chalmers, Reality+: Virtual Worlds and the Problems of Philosophy (W. W. Norton & Company, 2022), https://wwnorton.com/books/9780393635805.

252. Keith Frankish, Illusionism as a Theory of Consciousness (Imprint Academic (Exeter), 2017).

253. Daniel Stoljar, “Two Conceptions of the Physical,” Philosophy and Phenomenological Research 62(2): 253–281, ahead of print, 2001, https://doi.org/10.1111/j.1933-1592.2001.tb00056.x.

254. Sam Coleman, “The Real Combination Problem: Panpsychism, Micro-Subjects, and Emergence,” Erkenntnis 79, No. 1: 19–44, 2014, https://doi.org/10.1007/s10670-013-9431-x.

255. Kastrup, The Idea of the World, ch. 1.

256. A. Schopenhauer, The World as Will and Representation (Dover, (1819) 1969).

257. Coleman, “The Real Combination Problem.”

258. Chalmers, Reality+.

259. Nick Bostrom, “Are You Living in a Computer Simulation?” The Philosophical Quarterly 53(211): 243–255, 2003, https://doi.org/10.1111/1467-9213.00309.

260. Ethics (Penguin, (1677) 1996).

261. Theodicy: Essays on the Goodness of God, the Freedom of Man and the Origin of Evil (Essais de Théodicée (Amsterdam, 1710); standard Eng. ed. trans. E. M. Huggard, ed. A. Farrer (Routledge & Kegan Paul, 1951; repr. Open Court, 1985), 1710).

262. The Principle of Sufficient Reason.

263. An Essay on Free Will (Clarendon Press, 1983).

264. How Reason Can Lead to God.

265. The Nature of Necessity.

266. The Christian God (Clarendon Press, 1994).

267. “Guide to Ground.”

268. “Metaphysical Dependence: Grounding and Reduction,” in In Modality: Metaphysics, Logic, and Epistemology, Edited by Bob Hale and Aviv Hoffmann, 109–135. Oxford: Oxford University Press (2010), https://academic.oup.com/book/11764/chapter/160788198.

269. “On What Grounds What.”

270. “Monism: The Priority of the Whole,” Philosophical Review 119 (1): 31–76, 2010, https://doi.org/10.1215/00318108-2009-025.

271. “On the Plurality of Grounds,” Philosophers’ Imprint 14 (20): 1–28, 2014, https://quod.lib.umich.edu/p/phimp/3521354.0014.020.

272. Value and Existence (Basil Blackwell, 1979).

273. “Monism.”