Boundary Loss and the Born Rule

The Origin of Probability

Version 2 · September 13, 2026 · Part III of The Origin of Probability: A Complete Derivation in Three Parts

Quantum mechanics tells us how to calculate probabilities. This paper asks a prior question: how can a probability-bearing architecture arise inside a deterministic physical framework without treating randomness or probability as primitive?

The answer developed here is not that determinism fails. The key change is a loss of recoverability. Distinct admissible precursor states can leave the same later readable residue while differing in a physically relevant way. If the later description cannot recover which relevant precursor distinction obtained, then the evolution can remain lawful even though the determining distinction is no longer recoverable from that boundary description.

That loss is not yet probability. It creates the structural opening in which lawful unresolved alternatives can exist. The paper then shows what additional structure is required before those alternatives can support numerical probability, predictive use, and finally the local Born-form relation.

The endpoint is exact but local:

p_i = <J(Psi), P_i J(Psi)> = ||P_i J(Psi)||^2

This is the exact local source-linked Born form at the declared adopted-theory application scope. It is not an unrestricted global Born rule. The compatibility equality is not claimed to follow automatically from Boundary Loss, scalar grounding, normalization, or the independently constructed Hilbert bridges.

The locality is structural rather than provisional. The probability-bearing alternatives, event family, scalar measure, Hilbert bridges, and compatibility relation are all constructed relative to the same declared resolution context. The final Born relation therefore cannot lawfully extend beyond that domain without additional cross-context structure.

1. The Simple Idea

Start with an ordinary idea. Suppose two different past situations can lead to the same readable result. If the readable result no longer contains enough information to tell those past situations apart, then some distinction has become unrecoverable from that result.

The laws do not have to stop working for this to happen. The system can evolve deterministically while a later description preserves less recoverable information than the earlier one.

This is more precise than saying that “history is lost.” The formal claim is not that the universe erases its past in some absolute sense. It is that a particular later boundary or readout no longer preserves enough information to recover a particular distinction that matters to the question being asked.

That distinction matters because information loss by itself is not probability. A complete probability architecture requires several additional steps, and the paper keeps those steps separate.

2. Boundary Loss Comes Before Probability

The first step is a recoverability test, not a probability rule.

A many-to-one readout is not enough by itself. Several pre-boundary states may produce the same residue while agreeing on every distinction relevant to the active question. Recoverability-relevant Boundary Loss begins only when states compatible with the same residue disagree on a distinction that matters at the declared scope.

Boundary Loss is therefore not ordinary ignorance, not stochastic dynamics, and not a hidden probability distribution. It is the failure of a later readable residue to recover a distinction that is still represented among admissible precursors.

The paper begins in a deliberately weak realization-neutral setting. At this stage it does not assume a Hilbert space, a probability space, a measurement apparatus, a physical detector, a physical boundary occupant, or a specific microscopic theory.

3. A Deterministic Boundary Readout

Now the same idea can be written formally. Let a pre-boundary state space be mapped into a boundary-readable residue space:

B : S_pre -> S_B

For a pre-boundary state x, define the readable residue:

r = B(x)

All pre-boundary states that produce the same residue belong to the same compatibility class:

B^-1(r) = { y in S_pre : B(y) = r }

The analysis then fixes a nonempty tested domain:

empty != D subseteq B^-1(r)

The domain D is the boundary-compatible family actually under examination. The theorem does not automatically claim something about every state in every possible fiber.

The word boundary is used in a recoverability sense. The map may represent a quotient, projection, coarse-graining, export or readout map, horizon-like interface, restricted interface, or another admissible operation under which a distinction may cease to be recoverable. It need not initially be a literal physical surface or detector.

4. Guarantee, Exclusion, and Undetermination

For an admissible proposition P on the tested domain, the paper distinguishes three statuses:

Guaranteed_B(P,r;D) iff for all y in D, P(y)

Excluded_B(P,r;D) iff for all y in D, not P(y)

Undetermined_B(P,r;D) iff there exist y,z in D such that P(y) and not P(z)

If compatible states sharing the same residue disagree on the proposition, then the residue neither guarantees nor excludes it. The later readable description has lost recoverable determination of that distinction.

This theorem-level result is intentionally narrow. It does not yet assign alternatives, weights, measures, likelihoods, frequencies, or probabilities. It establishes proposition-level non-guarantee at the declared boundary/readout scope.

5. From Lost Guarantee to Lawful Alternatives

Loss of guarantee is still not enough. The unresolved distinctions must form a lawful alternative structure rather than an arbitrary list invented after the fact.

The next stage introduces a separately admitted scalar-neutral resolution relation ~K on the declared compatible domain:

F_K := D / ~K

The quotient must separately satisfy the required conditions for nontriviality, finiteness, non-artificial generation, redescription stability, nonfactorization through the single residue, and admissible presentation equivalence.

When those burdens are satisfied, the result is a representative-invariant scalar-neutral pre-probability context:

PreProbContext_B(r,K,D) := [F_K]_pres

The quotient classes are lawful unresolved alternatives. They are not probabilities. No magnitude, likelihood, frequency, measure, normalization, stochastic rule, or Born-form value has yet been introduced.

6. Physical Realization and Recurrent Persistence

The abstract Boundary-Loss result does not by itself establish that any quotient class is physically occupied or even physically realizable. Physical realization is a separate stage.

The paper carries qualified model recurrence into physical recurrence only through an explicit adopted physical-semantic bridge at the declared theory scope. The model recurrence has the form:

Psi(x,t + T) = Psi(x,t)

A classifier descends to represented physical states only after an independent preservation condition is supplied. Recurrence does not supply Boundary Loss. Classifier descent does not supply a probability variable. Integer-valued classifier structure does not establish physical quantization.

The physical Boundary-Loss roles are instantiated separately: physical carrier, physical boundary/readout, compatibility class, proposition architecture, scalar-neutral resolution relation, finite nontrivial quotient, and admitted presentation structure. Only after those burdens are independently satisfied does the physical pre-probability context exist at the adopted theory scope.

7. Source-Grounded Scalar Measure

Numerical structure enters only after the unresolved alternatives have been physically grounded at the declared scope.

Within the admitted source regime, the nonnegative scalar density is:

rho = |Psi|^2

A separately supplied carrier-grounding relation connects that source scalar content to the unresolved alternatives. For each alternative i, the source-grounded residue is:

s_i := integral over D_i of |Psi|^2 dx
s_i >= 0

Finiteness of the current alternative family together with finite-valued residue typing first establishes:

0 <= S < infinity
S := sum_i s_i

A separate positive-witness condition then establishes the denominator-licensing condition:

0 < S < infinity

The finite additive scalar measure is then:

m(J) := sum over i in J of s_i

and normalization gives:

W := m / S

At this point the construction has the mathematical form of a normalized measure on the earned finite event algebra. The manuscript does not yet call it predictive probability merely because it is normalized.

8. Local Predictive Probability

Normalization is not used to smuggle in predictive meaning. The theory states an explicit local finite probability-predictive admission criterion.

The normalized assignment is admitted for predictive use only when the current regime satisfies all required conditions:

• the event domain is exactly the earned finite nonempty unresolved-alternative event algebra;
• the assignment is fixed before any particular alternative is selected or occupied;
• the assignment is grounded only in the already earned source-grounded scalar residues, finite measure, and strictly positive finite normalization;
• the assignment is nonnegative;
• the assignment is normalized;
• the assignment is additive on the finite event algebra;
• admissible refinements and coarse-grainings preserve normalized totals;
• admitted presentation and redescription transport preserves the normalized values; and
• no external probability measure, hidden-variable completion, empirical-frequency rule, outcome-counting rule, ignorance postulate, equal-weighting postulate, or preferred representative is introduced.

Once the complete local criterion is verified, the theory admits:

p := W

p_i = s_i / S

Probability is therefore present at this locally admitted predictive scope before the final Hilbert-space comparison.

The status remains exact and limited. This is a conditional theory-level admission of predictive interpretation. It is not a theorem that predictive semantics follows source-neutrally from normalization, and it is not a claim that every physical prediction is probabilistic or that every Boundary-Loss context admits scalarization.

9. Source-Hilbert State and Event Bridges

The next stage does not rebuild probability from Hilbert space. Probability has already been admitted locally. The task is now to connect the source architecture to an independently constructed Hilbert representation.

A source-state bridge supplies the linked Hilbert state:

J(Psi)

A separate source-event bridge maps the relevant source events into Hilbert channels:

A_i -> H_i

Because the admitted H_i are closed Hilbert subspaces, ordinary Hilbert-space projection theory supplies the corresponding orthogonal projectors:

P_i

The two bridges are independent. Possessing both does not identify the locally admitted probability p_i with the Hilbert scalar <J(Psi), P_i J(Psi)>.

Common indexing does not supply that equality. Nonnegativity does not supply it. Additivity does not supply it. Coarse-graining does not supply it. Presentation invariance does not supply it. Projector structure does not supply it.

The compatibility burden remains.

10. Born Compatibility and the Exact Local Source-Linked Born Form

The paper isolates the minimum substantive compatibility commitment required at the exact finite linked source/Hilbert scope:

C_COMPAT_MIN

Its operative content is the eventwise equality together with preservation under the already earned admissible presentation, redescription, and finite coarse-graining structure:

p_i = <J(Psi), P_i J(Psi)>

The preceding architecture does not supply this equality source-neutrally. The theory therefore adopts C_COMPAT_MIN explicitly at the exact linked source/Hilbert scope and no stronger. Adoption is kept separate from application.

The adopted law is then applied to the already constructed local predictive probabilities and the independently established source-state, source-event, and projector structures. For every relevant event in the current finite family:

p_i = <J(Psi), P_i J(Psi)> = ||P_i J(Psi)||^2

This is the exact local source-linked Born form at the declared adopted-theory application scope.

The manuscript-specific result is not that the compatibility law was secretly contained in Boundary Loss. The result is that the two sides are constructed independently, the remaining compatibility burden is exposed rather than hidden, the minimum required law is isolated, and its adoption and application are stated exactly where they enter.

11. Why Probability and the Born Rule Are Local

This is one of the central structural conclusions of the paper.

Probability is local because the unresolved alternatives themselves are local. Boundary Loss does not generate one universal set of alternatives for the entire physical theory. It produces a lawful unresolved quotient only relative to a declared carrier, a boundary or readout map, a recoverability-relevant distinction, and a scalar-neutral resolution relation.

Change that resolution context and the lawful alternative space can change with it. The domain on which probability is defined is therefore not supplied globally in advance.

The same scope restriction propagates forward. Source-grounded scalar residues are defined on the current finite alternative family. Strict positivity, normalization, finite additivity, predictive admission, the source-state bridge, the source-event bridge, and the compatibility law are all established on that same linked context.

The final equality

p_i = <J(Psi), P_i J(Psi)>

therefore cannot lawfully outrun the domain on which its probability assignment, events, state bridge, projector bridge, and compatibility relation have actually been established.

Because the alternative space itself is generated relative to the resolution context, no single global probability-bearing domain is available within the derived architecture. Constructing one would require new cross-context structure capable of identifying or transporting alternatives, events, scalar assignments, state bridges, event bridges, and compatibility relations across distinct resolution contexts.

Local Born forms may therefore be exact within their lawfully constructed contexts without composing into one unrestricted global Born rule. Exactness inside each valid context does not create a universal rule across different contexts. Any attempt to glue the local forms together would require additional structure that is not supplied by the derivation and would constitute a new theoretical commitment rather than a consequence of the present architecture.

Accordingly, the paper does not merely stop at a local result for convenience. Within this architecture, locality is a structural constraint: the probability law cannot extend beyond the domain on which the unresolved alternatives and all subsequent compatibility structures are defined.

Here, “global Born rule” means a single source-neutral Born relation valid across all admissible physical resolution contexts, boundary maps, alternative spaces, and realizations. It does not mean the familiar use of the Born rule throughout one already fixed Hilbert-space model after its state space, event structure, and compatibility relations have been supplied.

12. The Full Sequence

The completed bounded route is:

persistent recoverable structure

-> boundary readability

-> local numerical probability architecture

-> recoverability-relevant Boundary Loss

-> lawful unresolved alternatives

-> physical recurrent persistence at adopted theory scope

-> source-grounded normalized scalar measure

-> locally admitted predictive probability

-> source-linked Hilbert structure

-> explicitly adopted compatibility law

-> exact local source-linked Born form

The arrows record dependency and handoff order. They do not represent one automatic implication chain. Each transition carries its own conditions, status, and burden.

The two probability stages in this sequence are different. The earlier local numerical probability architecture is the conditional Part II handoff: it identifies the regime in which boundary-readable deterministic structure can support probability-measure status. The later locally admitted predictive probability is the Part III result constructed on the physically realized Boundary-Loss alternatives after source grounding, strict positivity, finite measure, normalization, and the explicit local predictive-admission criterion have been satisfied. The later result does not travel backward and supply the earlier Part II conditions.

Part I establishes what can persist. Part II establishes the conditional architecture under which persistent deterministic structure can become boundary-readable, numerically weighted, measure-compatible, and probability-bearing. Part III establishes what lawful unresolved structure remains when recoverability-relevant Boundary Loss removes guarantee, then carries that structure through source grounding, finite measure, local predictive admission, independent source-Hilbert bridges, and explicit compatibility adoption to the exact local source-linked Born form.

13. Relation to Entanglement Compression Theory

The paper begins with realization-neutral Boundary-Loss mathematics and does not assume a specific physical carrier at that stage. It does not remain purely realization-neutral.

Later sections carry the protected structure into a declared adopted physical realization architecture. Qualified model recurrence is linked to physical recurrence at the declared theory scope. Physical carrier, readout, proposition, quotient, and presentation roles are supplied separately. Within the admitted source regime, the scalar density rho = |Psi|^2 is grounded to the unresolved alternatives and becomes the source of the finite scalar measure.

This ordering matters. ECT is not used to erase the distinction between Boundary Loss and probability. Boundary Loss first supplies recoverability failure and lawful unresolved alternatives. Source grounding supplies scalar content. Finite measure and normalization organize that content. The local predictive criterion supplies the theory-level admission of predictive probability. Independent Hilbert bridges supply the linked state and event representation. The final compatibility law is then adopted and applied explicitly.

The broader ECT mathematical framework preserves the same firewall between normalized scalar structure and probability interpretation. The present paper uses its own finite current-context admission criterion rather than treating the broader framework as an occupancy theorem for every ECT state or every physical prediction.

For the broader ECT mathematical treatment, see Mathematical Foundations of Entanglement Compression Theory: Causation, Compression Dynamics, Derived Probability, and Spacetime-Emergence Setup.

14. What This Paper Does and Does Not Claim

The paper establishes a bounded probability-origin architecture with an exact local endpoint.

• It establishes recoverability-relevant Boundary Loss as proposition-level non-guarantee at the declared boundary/readout scope.
• It establishes a finite, nontrivial, representative-invariant scalar-neutral unresolved context when the required quotient conditions are satisfied.
• It carries that architecture into a separately staged adopted physical realization.
• It constructs source-grounded nonnegative scalar residues and a finite additive scalar measure.
• It normalizes the measure only after strict positivity of the total is separately established.
• It admits the normalized measure as local predictive probability only after the explicit finite admission criterion is satisfied.
• It constructs independent source-state and source-event Hilbert bridges.
• It isolates, explicitly adopts, and applies the minimum compatibility law needed to obtain the exact local source-linked Born form.

It does not claim that Boundary Loss alone produces numerical probability. It does not claim that normalization alone produces predictive semantics. It does not claim that source-state and source-event bridges force Born compatibility. It does not establish measurement theory, collapse dynamics, event occupancy, empirical confirmation, physical quantization, necessity, fundamentality, or final physical theory.

It also does not establish an unrestricted global Born rule. Within the architecture developed here, the probability-bearing domain is generated locally from the resolution context, and every later step inherits that scope.

The claim is sharper: deterministic recurrent organization can coexist with a physical boundary/readout at which recoverability of a determining distinction fails while lawful structure survives. Under separately stated scalar, measure, predictive-admission, Hilbert-bridge, and compatibility conditions, that surviving architecture reaches local predictive probability and the exact local source-linked Born form.

15. Why It Matters

The paper changes the explanatory order of probability without hiding where additional theory content enters.

It does not begin with randomness. It does not begin with the Born rule. It begins with persistence and asks what a later readable boundary can fail to recover.

When recoverability fails in the relevant way, lawful alternatives can remain. Those alternatives can receive source-grounded scalar content. That content can support a finite measure. The measure can be normalized. Under an explicit local criterion, the normalized measure can be admitted as predictive probability. Independent source-linked Hilbert structures can then be built. The final scalar equality still does not appear for free, so the missing compatibility law is exposed, adopted, and applied.

This is the central discipline of the derivation: no downstream result is allowed to travel backward and pretend it supplied an earlier missing burden.

The second major lesson is equally important: locality is not merely a disclaimer attached to the final result. It follows from the fact that the lawful unresolved alternatives, event family, scalar measure, predictive assignment, Hilbert bridges, and compatibility relation are all context-bounded. The exact Born form is therefore local for the same reason the probability-bearing domain is local.

16. Source and Formal Paper

This webpage is an explanatory guide. The formal definitions, theorem stack, Boundary-Loss non-guarantee theorem, scalar-neutral quotient construction, representation-invariant realization, adopted physical instantiation, source-grounded scalar measure, predictive-admission criterion, source-Hilbert bridges, compatibility analysis, nonclaims, appendices, and references are in the Zenodo paper.

Source: Lawrence, W.A. (2026). Boundary Loss and the Born Rule: The Origin of Probability. Version 2, September 13, 2026. Zenodo. https://doi.org/10.5281/zenodo.20102172

Related CTI papers:
Finite Recurrent Stability Before Spacetime: Collapse, Emergence, Horizons, and Recurrence as Recoverability Boundaries | Stability, Boundary Observability, and Emergent Probability in Deterministic Systems | Mathematical Foundations of Entanglement Compression Theory | Theory of Derived Probability and Entanglement Compression | The Oscillation Principle | Quantum Gravity in a Deterministic Universe