The argument begins from a distinction that is easy to lose if rules, possibilities, and realized states are treated as one thing.
The underlying rules do not change. They constrain what can occur, but constraint is not prescription: a rule can remain invariant while permitting enormous variation in organization, interaction, and realization. Nothing about invariance alone requires every possible realization to belong to one common inventory of states, and nothing about it requires different regimes to share the same objects, measurements, or forms of description. That inference is often drawn from invariance and is declined here, because it adds a commitment the invariance does not contain.
Separating the ideas requires three terms that are usually allowed to run together: representation, admissibility, and realization. Something may be representable without being admissible, and admissible without being realized. The ability to construct or describe a possibility does not establish that it can occur under the constraints of a particular system, and the fact that something could occur does not establish that it does. The rules restrict admissibility; they do not dictate realization.
These distinctions become consequential at regime boundaries. A regime determines more than which values happen to occur. It also determines what can function as an object, what can count as a state, which transformations are available, and which operations can produce evidence. Two regimes therefore need not differ merely by occupying different locations in an otherwise unchanged description; the forms through which their underlying relations become observable may themselves differ. None of that requires the rules to change. It means only that invariant rules do not imply invariant realization.
The distinction is relational before it is object-level. Distinguishable roles may enter into relation without requiring independently persistent object identities, and the coupled condition may realize a configuration not attributable to either role considered independently.
Let B stand here only for the underlying relational structure whose rules are invariant. A regime supplies its own objects, states, transformations, and evidentiary operations. No assumption is made that two regimes share these, that either regime’s description contains the other’s, or that a natural translation exists between them.
Local License and Export License
Every observation is made from somewhere. The observer, the instrument, the measurement protocol, and the resulting record are instantiated within a regime, so evidence obtained there establishes, first, a local license: warrant for a claim within the conditions under which that evidence can be produced. An export license is something stronger. It warrants carrying the claim beyond those conditions, into a regime that need not share the objects or operations that certified it. The two should not be confused, and the first does not contain the second.
A relation may hold with extraordinary precision within one regime and still depend on objects or operations peculiar to that regime. Mathematical necessity does not by itself resolve this. A conclusion can follow necessarily from a formal structure while the applicability of that structure remains regime-local; the proof is valid while its export is unlicensed. And the mathematical type in which the necessity was stated may fail to carry over before any theorem in it does.
The question is therefore not simply whether a relation is invariant. It is what the invariance belongs to.
Object-Agnostic Invariance
If an apparent invariant depends on the particular objects, states, coordinates, or evidentiary operations supplied by one regime, its invariance may be local however exact it appears. The stronger candidate is a relation whose statement does not depend on those regime-specific objects at all. The criterion proposed here has two parts.
A relation I is object-agnostic if it can be specified at the level of B, without reference to the objects, states, or evidentiary operations of any particular regime. A relation established within a regime is a candidate for export only if it is the local expression of an object-agnostic I, so that its holding in a second regime, whose objects may bear no correspondence to the first’s, is a consequence of I and not of the objects through which it was first observed.
Object-agnostic does not mean object-free observation. Observations necessarily occur through particular systems; the requirement is that the relation proposed for export is not defined by the identity of the objects through which it happens to become observable. Different realizations therefore need not be identical for a relation to survive between them. Invariance, in this sense, is not an unchanging state. It is a relation preserved through change.
Object-agnostic invariance therefore asks whether relational structure survives where object identity does not.
What a Candidate Relation Would Have to Preserve
The criterion specifies the form a successful candidate would have to take. The objects themselves are not what is proposed to survive. A, B, and C may belong only to one realization and may cease to be meaningful across a regime boundary. What would have to survive is some relational quantity I that can be specified without depending on those particular objects.
For illustration only, consider the following schematic relational form. It is a conceptual aid, not a defining equation, derivation, or representation of the proposed formal mathematics:
Δᵣᵢ(A, B) ≠ 0, A ℛᵢ B → C
Here, ℛᵢ denotes coupling or relation as expressed in regime i, while Δℛᵢ(A, B) ≠ 0 marks relational distinction within that coupling. It does not require A and B to possess independently persistent identities across regimes. The labels identify distinguishable roles within this realization only.
The point of the schematic is that relation need not erase distinction, and the coupled condition may support a realized organization not attributable to either term considered independently.
A second realization need not contain the same objects or state descriptions. Its terms may be entirely different:
A′ ℛⱼ B′ → C′
with
(A, B, C) ≢ (A′, B′, C′).
If something exports between these realizations, the proposed invariant is therefore not the identity of A, B, or C. The required form is instead a relational quantity I, specified from relational structure rather than from the identities of the participating objects, such that visually, the required result would take the form:
I[ℛᵢ] = I[ℛⱼ]
This expression is a template for the test, not an exhibited invariant. No functional form for I is supplied here, and no equality across a genuine regime boundary is claimed as an empirical result. What has been identified is the job such an invariant would have to perform: remain specifiable and comparable even where the objects through which the relation is realized do not.
The resulting test is therefore more specific than the assertion that “some relation must survive.” It asks whether a quantity I can actually be defined from relational structure independently of the object vocabulary of either regime, computed in both realizations, and shown to satisfy
I[ℛᵢ] = I[ℛⱼ] across a genuine regime transition. Until such an I is identified and evaluated, this remains a specification of what a candidate footprint must look like, not a candidate footprint itself.
This also identifies the characteristic failure of export. A relation can appear fundamental from inside a regime because the regime provides no vantage from which its dependence on its own objects is visible. Where those objects change, disappear, or cease to be meaningful across a boundary, an invariant tied to them cannot be carried through by assumption. What survives, if anything does, must be specified without depending on the objectification peculiar to either side.
The criterion therefore carries a condition on its own use. It can be applied only if the underlying relational structure B can itself be specified, from within a regime, in terms that do not depend on that regime’s objects. Whether this is possible is not settled here. The present argument requires only the distinction: a relation established within a regime is not thereby established across regimes.
Quantum Theory as a Near Case
Quantum theory does not solve the export problem; it provides some of its clearest working examples.
The transition from classical to quantum mechanics shows that successful physics need not preserve its objects or even its mathematical type. Classical observables are functions on phase space; quantum observables are operators. Yet, where canonical correspondence is licensed, the classical Poisson-bracket structure and quantum commutator structure are related schematically by
{f, g}ₚᵦ ↔ (1/iℏ)[f̂, ĝ]
This is not an unrestricted exact translation. The point is narrower: a licensed structural correspondence can survive while the object inventory and mathematical representation change. Classical mechanics is recovered only under specified limits, approximations, or correspondence regimes, not by assuming that classical objects pass unchanged into quantum theory.
Quantum theory also makes the distinction between representation, admissibility, and realization difficult to avoid. A mathematical state may be representable without being dynamically or physically admissible, and an admissible state need not be the state realized in a particular preparation or measurement. Entanglement supplies a near case for the relational form above: a joint description can contain structure not exhausted by properties assigned independently to the participating subsystems. This does not identify the required invariant or support an ontological conclusion about the base; it shows that relational structure need not reduce to independently persistent object properties.
When quantum physics crosses descriptions successfully, it normally does so under a specific license. Classical limits and correspondence arguments provide controlled overlap. Effective field theories explicitly restrict claims to a domain and decline uncontrolled export beyond a cutoff. Bell-type arguments take a different route: a proposed class of underlying accounts is required to leave a measurable constraint on records available here, allowing the proposed ontology to be audited without requiring direct access to its objects. Dualities raise the complementary possibility that very different object-level descriptions may nevertheless preserve common structure.
These are powerful licenses, but they are particular licenses. None supplies a general rule saying that a relation established as necessary within one formal regime must survive wherever the objects, operations, or mathematical type of that regime no longer apply. Quantum theory therefore sharpens rather than removes the question pursued here: when the objects do not export, what licenses the relation to do so?
Alive
One ontological proposal is added, and it is to be kept separate from the epistemic result above. Alive names the proposal that relational activity is intrinsic to the underlying structure: coupling, response, and differentiated realization are expressions of the activity of the base rather than operations supplied to an otherwise passive substrate.
The schematic relation introduced above does not establish this proposal. The same observable coupling can be described under an inert ontology. The additional claim made here concerns the source of that coupling: R is proposed not merely as a description of what occurs between otherwise inert terms, but as an expression of relational activity intrinsic to the base from which the realized configuration arises.
Alive does not mean that the rules change. It does not mean that the substrate rewrites its laws, that anything whatsoever is permitted, or that constraint disappears. Activity occurs under invariant constraint. Constraint and activity are different things: the rules constrain what the base can realize without being posited as an external source of its activity.
Whether that distinction is empirically available is a separate question. If intrinsic relational activity and an inert account generate exactly the same possible records, no local experiment discriminates between them and Alive remains an ontological claim rather than an empirical result. If the distinction leaves a footprint, some relation among records must be expected under one account and excluded by the other. No such discriminator is claimed here.
What Remains
The position is narrower than a theory of what lies beyond every regime, and stricter for it. Stable rules do not guarantee identical realizations. Local invariance does not guarantee export. Mathematical necessity does not guarantee that its mathematical type survives a boundary. And an ontological proposal does not become an empirical conclusion merely by being compatible with the same architecture.
What remains is the question with which the paper began: when every observer, instrument, and record is instantiated within a regime, what can warrant the claim that a relation established here survives where the objects, operations, or conditions that established it no longer do? The answer cannot be that the relation worked here. The footprint, if there is one, must survive what the object does not.