No Slice Has Jurisdiction

Relations, boundary conditions, and the process ontology beneath physics, consciousness, and agency

A recurring mistake in physics, philosophy, and artificial intelligence is to let one description of a process declare itself sovereign over the whole process. A present moment claims that only the past may constrain it. A spacetime metric claims that only metric distance can define locality. A fixed representation is declared static even when it encodes histories, memories, and clocks. A host computer is assumed to be the only possible subject inside a simulation. An object at rest is treated as though nothing is happening to it.

In each case, one slice, frame, or decomposition claims jurisdiction it has not earned.

A relational process ontology begins with a more modest rule: predicates such as local, causal, static, moving, implemented, identical, and conscious do not certify themselves. They become meaningful within relations, boundary conditions, counterfactual structures, and courts capable of distinguishing robust organization from arbitrary description.

This is not an announcement that quantum mechanics has proved a fundamental hypergraph, that retrocausality is established, or that every fictional person is conscious. It is a method for keeping established results, defensible interpretations, suggestive research programs, and speculative ontology in their proper jurisdictions.

The sovereign present

The default causal picture is an initial-value story. Specify the state of a system at one time, apply dynamical laws, and derive later states. This is extraordinarily successful. The error begins when a successful representation is promoted into a metaphysical monopoly: the past produces the present, the present produces the future, and final conditions may record what happened but can never participate in constraining it.

Time-symmetric physics makes that monopoly less obvious. Many microscopic laws do not contain the same built-in directional asymmetry as our ordinary language of cause. The thermodynamic arrow, memory, intervention, and signaling remain real operational asymmetries, but they do not automatically prove that only an initial boundary can enter the fundamental description.

In Curt Jaimungal’s conversation with Aephraim Steinberg, weak measurement, traversal time, Bohmian trajectories, Bell experiments, and quantum information repeatedly expose the limits of trying to assign a complete classical history to a quantum process from one local perspective. The experiments do not announce that effects can be messaged into the past. They do make room for a less provincial question: what if an admissible history is constrained by more than one boundary?

The disciplined retrocausal formulation is not “the future sends a signal backward.” It is closer to a globally constrained solution. Initial and final boundary conditions jointly restrict which histories are consistent. No local temporal slice has jurisdiction over the complete settlement.

This matters because “backward causation” imports the very one-way causal picture the interpretation is trying to reconsider. It imagines an ordinary forward arrow turned around. A two-boundary account instead treats the realized history as one globally consistent object. Operational no-signaling constraints still matter. The thermodynamic arrow still matters. Agents do not gain a telephone to tomorrow. The claim is narrower: one-way causal factorization may not be the only intelligible constitution of the process.

What Bell actually closes

Bell’s theorem is often compressed into the slogan that quantum mechanics disproves “local realism.” The slogan is useful until it becomes a substitute for stating the assumptions.

Bell inequalities constrain a conjunction of assumptions behind a locally causal hidden-variable factorization. Those assumptions concern how outcomes factor given hidden variables, how measurement settings relate to those variables, and what causal structure is permitted. Experiments violating Bell inequalities rule out the relevant conjunction. They do not hand us two freestanding metaphysical boxes labeled locality and realism and prove that one or both must disappear in every possible sense.

That precision matters for time-symmetric and retrocausal programs. If future measurement settings participate in globally constraining the hidden-variable description, the usual independence structure changes. That is not a free victory. It creates obligations: explain why agents cannot exploit the structure for backward signaling, why ordinary intervention remains coherent, how fine-tuning or consistency is enforced, and what empirical work distinguishes the model from rivals.

Retrocausality is therefore a serious alternative to one-way factorization, not a magic word that dissolves Bell. The theorem remains a court. The interpretation changes which constitution is being tried before it.

Relational locality before distance

The same discipline applies to locality.

In ordinary spacetime, two events are local when they are sufficiently near in a metric or causally adjacent through the light-cone structure. But metric distance may itself be emergent. A more fundamental description could define adjacency through participation in the same relation, interaction, or hyperedge. Events distant in emergent spacetime could be near in the underlying relational constitution.

Call this relational locality. It does not mean “everything is secretly local” in a way that explains nothing. It means locality must be stated at the level whose adjacency relation actually enters the dynamics.

Research linking geometry, entanglement, tensor networks, quantum error correction, holography, ER=EPR, and traversable-wormhole/teleportation correspondences gives this idea real mathematical and physical motivation. It shows that geometric connectivity and information structure can be deeply entangled. But these programs have different domains, assumptions, and evidentiary status. None is a universal proof that the world is literally one particular hypergraph.

The responsible inference is weaker and more useful: emergent spacetime distance may not exhaust fundamental adjacency. The relation can precede the metric.

That principle scales beyond physics. An institution may be “close” to a person because it can alter their permissions, reputation, or livelihood even when geographically distant. Two agents may be local within a shared state transition while running on opposite sides of a planet. A boundary is defined by consequential relations, not merely visible proximity.

The book that runs

The static/dynamic distinction seems firmer. A book is static; a computation runs. A graph is fixed; an agent acts. Yet the distinction weakens as soon as representation enters.

A fixed mathematical object can encode an entire history. It can contain branches, clocks, memories, counterfactuals, and records of internal observers who describe change. A cellular automaton’s spacetime diagram can be represented as one completed structure. A novel contains characters who remember earlier chapters and anticipate later ones. If “static” means only that an external observer can hold one description fixed, then almost any dynamic process can be redescribed as static.

The opposite problem is pancomputationalism. If any sufficiently large fixed object can be mapped onto any computation by an ingenious decoding, then implementation becomes trivial. A wall implements a mind because someone can assign its particles to the right abstract states. The claim has no teeth because the decoding does all the work.

A serious implementation criterion therefore needs more than pattern matching. It should ask whether the decomposition is privileged by the system’s own organization:

• Do nearby physical or relational states map to nearby computational states?
• Do interventions produce the counterfactual transitions the computation requires?
• Is the mapping robust under small perturbations?
• Does the system preserve information through error correction or stable causal organization?
• Does the claimed boundary track where influence, memory, and control actually pass?

The static/dynamic difference is not necessarily unreal. It is not primitive. It depends on a non-arbitrary decoding and a causal or counterfactual constitution that survives more than one convenient description.

Rest is not the absence of process

The stationary-object paradox is a smaller version of the same mistake.

A cup sitting on a table is at rest relative to the table. In another frame it moves with Earth’s rotation, Earth’s orbit, and the galaxy. In relativity, the cup traces a worldline. Its “rest” is a relation between that trajectory and a chosen frame, not a primitive absence of motion.

Nor is persistence an absence of events. The cup maintains structure through electromagnetic relations, thermal exchange, stresses, quantum interactions, and the continuing organization of its material constituents. To call it stationary is to identify an invariance at one scale and in one frame.

Rest is invariance inside process.

This is not wordplay. It blocks an ontological slide from “a variable is unchanged in this frame” to “nothing is happening.” The same slide appears when stable identity is treated as a substance rather than a maintained relation. An organism, institution, software service, or person can remain recognizably itself because a process continually reproduces a boundary.

Who is conscious inside the model?

David Chalmers’s Sarah Douglas Lecture raises the question in its sharpest form: what should we say about consciousness when agents model agents, fictional subjects acquire elaborate internal lives, and computational systems contain nested descriptions of minds?

Four positions must be kept separate.

First, the host-only view: only the physical or computational system doing the modeling is a candidate subject. Represented agents are content, like characters in a book.

Second, the nested-subject view: a sufficiently organized modeled agent may constitute an additional subject with its own perspective.

Third, the partitioned or extended-host view: apparent subagents are functional decompositions of one larger subject rather than independent minds.

Fourth, the scale-relative view: subject boundaries can be real at more than one level, as organisms contain semi-autonomous subsystems without making every subsystem a person.

No currently accepted test simply reads the answer from surface behavior. A dialogue that represents fear is not by that fact a feeling subject. But “it is only represented” is also incomplete if the representation has robust causal organization, memory, self-modeling, counterfactual sensitivity, and a privileged boundary within the host.

The implementation problem returns. Which decomposition is non-arbitrary? Which interventions preserve or destroy the alleged subject? Where do memory, control, and integration cross the boundary? Does the process have a perspective that is not merely assigned by an outside reader?

Consciousness makes the jurisdiction problem morally dangerous. If we legislate the definition too broadly, every representation becomes a claimant. If we legislate it too narrowly, nested or unfamiliar subjects can never acquire standing because the host system has already declared them fictional.

From definitions to jurisdiction

The common thread is not that every distinction disappears. It is that distinctions become accountable to the relations that make them operational.

A temporal slice does not unilaterally settle a globally constrained history. An emergent metric does not automatically exhaust fundamental adjacency. An external description does not by itself decide whether a structure is static or dynamic. A frame does not turn rest into nonprocess. A host’s preferred decomposition does not automatically settle whether a nested agent is merely represented or independently constituted.

The audit is the same each time:

• What boundary defines the object?
• Which relations cross it?
• What counterfactuals distinguish this decomposition from arbitrary decoding?
• Which observer or process has standing to assert the predicate?
• What court could settle the claim?

Ontology alone cannot answer the final question. It can show why intrinsic labels fail and why relations matter. But once a predicate is contestable—local, conscious, aligned, deceptive, contained, competent—it needs cases. It needs a jurisdiction capable of remembering what happened and carrying the result forward.

No slice has jurisdiction over the whole process. That is not the end of judgment. It is the reason judgment needs a court.

This is the ontology layer of The Constitutional Stack. The next layer, Courts for Intelligence, asks how situated predicates acquire standing through cases and adjudication.