Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

SFIT Stochastic-Braid Substrate: Part III - Finite-Carrier Record Formation and the Conditions for Born-Weighted Outcomes: A Structural Analysis

Luiz PUODZIUS

Published Nov 2, 2026• 0 citations
Publisher

Abstract

This deposit contains the manuscript and the three analysis scripts (with outputs and SHA-256 hashes) for Part III of the SFIT Stochastic-Braid Substrate series. The paper separates three objects that are often conflated in attempts to obtain quantum measurement statistics from a classical substrate: (i) the substrate configuration (positions, velocities, helicities of the Part I defects), (ii) the channel population \(x_k \ge 0\), \(\sum_k x_k = 1\), and (iii) the record probability \(P(k)\), i.e. the probability that a physical record names channel \(k\). The Born rule relates (iii) to (ii); it says nothing by itself about how (ii) arises from (i). The paper asks three questions and reports the status of each: 1. Can a shared, finite carrier enforce exclusive records? In a coupled-mode model with explicit local detectors, detector-local back-reaction lowers the gated coincidence parameter from \(\alpha = 1.00\) to \(\alpha = 0.57\text{–}0.60\) at the strongest back-reaction tested, but only when detector thresholds are comparable to the carrier energy. At thresholds one-twentieth of the carrier energy, \(\alpha\) stays at \(0.96\text{–}1.00\). Strong exclusion (\(\alpha < 0.1\)) was not reached in any of 136 configurations. Result: partial suppression only. 2. Under what conditions does record selection reproduce Born weights? Two sufficient routes are identified, each with a precisely quantified fragility. Route 1: exactly memoryless (exponential) detector thresholds give Born frequencies by a common-time-change argument; gamma thresholds of shape 2 depart by TV distance \(0.08\text{–}0.15\). Route 2: a bounded martingale for channel populations absorbed at a vertex gives Born weights for any functional form of the exchange law; a first-passage bound and a converse are proved (given absorption, zero conditional drift is necessary), and a biased exchange kernel is shown to follow the Kimura fixation formula, so zero drift must hold to order \(1/M\) in the number of exchange units. Common-mode noise on the shared carrier cannot substitute for detector-local randomness: with fixed thresholds the largest channel fires first in every run, at every reversal rate. 3. Does Part I supply these conditions? No. The Part I helicity representation acts by permutation matrices, and no permutation action yields zero conditional drift for any filtration unless it is the identity (for pairwise-distinct populations). Measured on the Part I crossing process, the drift is nonzero at every test state, and populations are rearranged but never concentrated; sector populations are exactly invariant. This identifies braiding as neutral with respect to record selection. Scope and caveats. No derivation of the Born rule or of quantum superposition from the substrate is claimed. The internal state \(\psi\) is an input; the exchange kernel of Route 2 is global and postulated; memoryless detector thresholds (Route 1) are assumed; strong exclusion was not achieved; and the first-passage record of Proposition 3 (\(\alpha = 0\) by conservation) has not been combined with local detectors in simulation. Real amplitudes suffice for the single-system statistics studied, but real-amplitude quantum theory is experimentally distinguishable from complex quantum theory in network scenarios. Each result in the manuscript is tagged [DERIVED], [ASSUMED], or [TESTED]. Three scripts produce every number in the paper; each states its model, metrics, controls, predictions and decision rules in its header, fixed before the runs reported. The scripts were not preregistered; this deposit supplies them together with their outputs and hashes to support reproducibility.

DOI
10.5281/zenodo.23899562

Related papers

SFIT Stochastic-Braid Substrate: Part II - Conditions for Born-rule record statistics in a classical absorber race

Luiz PUODZIUS

Part II of the SFIT Stochastic-Braid Substrate series asks which additional structure would be required for outcome records to occur with Born-rule frequencies \(|\psi_k|^2\), and how much of that structure the substrate itself could plausibly supply. Part I established that persistent stochastic defects generate braid-group structure with orthogonal representations preserving \(\|\psi\|^2\), but made no claim about outcome probabilities; pure braiding cannot produce definite records, since orthogonal invertible maps cannot drive the internal state to a basis state. This paper studies a classical absorber race. Each channel \(k\) has an environment oscillator driven at resonance by a force linear in the carrier amplitude \(\psi_k\); a record is produced by the first channel whose absorbed drive work exceeds a random threshold. The quadratic dependence of absorption on \(\psi_k\) is derived rather than assumed: a harmonic absorber gives \(P_k = g^2\psi_k^2/(2\gamma)\), while an anharmonic (Duffing) absorber yields fitted exponents of 1.65 at weak drive and 0.74 at strong drive, showing that the quadratic law holds only in the linear-response regime. With exponentially distributed thresholds, record frequencies agree with \(|\psi_k|^2\) within sampling error for all tested states in \(n = 2, 3\) and 4 channels. With fixed, uniform, or gamma-distributed thresholds of shape \(k > 1\), frequencies depart from Born values toward the largest weight - already at a total-variation distance of 0.09 at \(k = 1.5\). This is the behaviour characterized by Yellott's theorem: within races with independent, identically distributed thresholds, choice probabilities proportional to rates for three or more alternatives require exponential thresholds. The paper distinguishes this \(n \geq 3\) condition (number of competing alternatives) from the \(n \geq 3\) condition in Gleason's theorem (state-space dimension). Oscillator simulations match the exact race integral throughout, and thermal noise up to \(k_BT = 0.1\) leaves frequencies unchanged when records are triggered by drive work. The result is explicitly conditional. Linear coupling, steady drive, channel independence and symmetry, memoryless thresholds, and mutual exclusion of records are assumptions of the model, not consequences of Part I; Table 2 lists all conditions and their status. The paper identifies memoryless thresholds (potentially supplied by the exponentially timed velocity reversals of Part I) as the most promising substrate contribution, states a pre-registered decision rule for testing it, and proposes further tests (start-from-rest ramps, total-energy triggers, over-dispersed thresholds). No derivation of the Born rule from the substrate is claimed; the mechanism is presented as a candidate, with the task for SFIT made specific. Simulation code (stage_c_absorber.py) reproducing all numbers from fixed random seeds accompanies the preprint.

•Oct 25, 2026•0 citations

SFIT Stochastic-Braid Substrate: Part IV - Ring Crossings Cannot Supply a Fair Exchange Sign: A Topological Obstruction and a Prediction of the Record Bias from Substrate Statistics

Luiz PUODZIUS

This paper investigates whether the one-dimensional stochastic dynamics of the SFIT stochastic-braid substrate can supply the conditionally fair exchange sign required for Born-weighted record formation. Building on Part III, which established that Born-weighted outcomes follow when channel populations form a bounded martingale absorbed at a vertex, we examine two possible sources of exchange signs in the one-dimensional ring dynamics: velocity reversals and defect-crossing orientations. The analysis identifies two distinct obstructions. First, velocity-reversal signs alternate for each defect, producing systematic second-order drift rather than conditionally fair exchange. Second, crossing orientations are constrained by the cyclic order of defects on the ring. We prove that this orientation is determined by the preceding crossing history and therefore cannot supply a conditionally fair sign. This is a topological obstruction, independent of the particular population jump law whenever that law responds to the crossing sign. For the carrier-mediated exchange law studied here, the obstruction produces a measurable departure from Born-weighted outcomes. At reversal rate γ = 1, the resulting bias is approximately +0.056 in the probability of the initially largest channel, remains essentially unchanged across the tested exchange amplitudes, and is statistically distinct from the martingale prediction. We then construct a reduced predictive model using only the substrate's measured pair-crossing statistics and the topological orientation rule. The model predicts record probabilities without fitting parameters to carrier outcomes. Across multiple reversal rates, initial population states, and time-step settings, its predictions agree with independent carrier simulations within the stated uncertainty. An ablation retaining only marginal pair frequencies fails to reproduce the reversal-rate dependence, demonstrating the importance of pair-sequence memory. At γ = 4, convergence of the finite-history model is not established, and that case is explicitly reported as a model limit rather than a successful validation. The results establish that the one-dimensional ring-crossing mechanism does not provide the fair-sign condition needed for the martingale route to Born-weighted records. Instead, the substrate's deterministic topological orientation constraint, together with the memory of its crossing sequence, predicts the resulting departure from Born statistics. The study is limited to the one-dimensional ring configuration and the specified carrier-mediated jump law. In particular, it does not settle whether the braid-generator sign in the planar dynamics of Part I can be conditionally fair. That remains a separate question for genuine two-dimensional braid geometry. Part IV therefore provides a topological no-go result for the one-dimensional fair-exchange route and a quantitative, independently tested account of the record bias that replaces it.

•Nov 4, 2026•0 citations

An Audited Born Calculus on a Pre-Physical Stochastic Substrate

Ioannis Tsiokos

Can a Born-type probability structure arise on a substrate that is only a finite stochastic process? A stochastic program built from six primitive operations supplies the final matrix; a specified layered architecture constructs, from its outputs, a record layer, a two-dimensional real carrier and a quadratic law, and supports an adopted Born calculus. The six operations alone do not give this structure, and the square-root encoding and the projection are construction choices. Our main contribution is the architecture: a directed acyclic graph whose arrows are each marked as derived, chosen, adopted or standard mathematics, with the kind of check behind each. On registered populations, a snapshot predictor whose rank-two case is the law predicts the substrate's sampled label frequencies with lower aggregate loss than named competitors, and, in the construction's canonical basis, its cross terms carry its advantage over the diagonal rule. In each of 2,122 census cases the encoded state is not a function of the anonymous initial record description: two point preparations sharing all three labels differ by at least 1/25 in the trace probability of raw label 0, a statement about exact data, not operational access. This bound is retrospective: the registered bound 3/20 fails in 117 cases of the full census. On the original worked case the updates by the final matrix and its square do not lift to the carrier. The measurement rules are adopted: Lüders tests, all densities and effects, and additivity; on them the trace rule is unique, and finite ontic models with additive responses, and measurable preparation-noncontextual representations preserving the specified mixtures, are excluded. We do not claim that the substrate produces the measurement rule, that the layer forms in the run, or a derivation of quantum theory. The source code of the stochastic program, the constructions, the reproduction pipeline, the verification commands and the Lean 4 proofs are available at https://github.com/ioannist/six-birds-born (release v1.0). AI systems were used throughout this work, as the manuscript describes; the author directed the research, made the scientific decisions and is responsible for the content.

•Oct 14, 2026•0 citations

Finite-Observer Martingale Record Selection

Itay Priiz

This theoretical and computational preprint integrates standard QND measurement and posterior-martingale results with finite-confidence decisions and physically retained observer records. Under a declared quantum instrument and its standard outcome probabilities, the posterior is a bounded martingale on an increasing observation filtration. A precise sufficient identification condition yields asymptotic terminal weights; a binary Gaussian benchmark gives exact mean stopping times, finite-threshold latch probabilities, and a controlled population-transfer bias bound. Version 2.3 is the first public release. It distinguishes accumulated observations from lossy current memory; assumed Born/instrument probabilities from derived consistency results; expected entropy from pathwise ambiguity; continuous confidence endpoints from discrete overshoot and unresolved runs; and Kramers first escape from endpoint readout error. It also clarifies the conditional D18/D19 persistence-selector interface and the separate D1092 finite-resource ledger, with direct attribution to established QND convergence and filtering literature. The companion archive includes the editable manuscript, PDF, equation sources, a fixed validation protocol, executable independent checks, numerical results, a revision ledger, and unchanged historical inputs. Historical simulation summaries are explicitly distinguished from new executions; some older executable sources were unavailable. The new audit includes exact finite-kernel and stopped-posterior checks, Gaussian likelihood-mixture calculations, five neutral and four population-transfer boundary-value cases, and retention/accounting controls. All examples are analytic or synthetic. No laboratory measurements, independent Born-rule derivation, measurement-problem solution, universal physical selector, or new empirical anchor are claimed.

•Sep 14, 2026•0 citations

Born-Convergence and Decision-Level Memory: A Δt-Sweep Operational Exclusion Criterion for Readout-Only Models

Emre Koçana

Quantum mechanics predicts outcome statistics via the Born rule, yet remains operationally under-specified with respect to how, and on what timescale, a single outcome becomes definite in an experimental record. In many platforms, identical Born-compliant outcome histograms can be produced either by a time-extended physical closure process, in which competing channels accumulate influence until a decision event, or by purely readout-level mappings and post-processing applied to measurement records. This under-determination motivates an operational question: which broad classes of readout-based explanations remain viable given finite-time, resolution-dependent data alone? Here we formulate an operational exclusion criterion relative to a broad null family of readout-only, memoryless models. In this null family, for each temporal resolution Delta t, the reported outcome is generated by a (possibly stochastic and nonlinear) mapping from a single-shot record R to an outcome Y, with no decision-level internal state persisting across Delta t settings and no dependence on sweep direction or parameter history. Under these structural assumptions, sweep-history dependence is excluded by construction. Using Delta t as a single experimentally controlled parameter, interpreted as an effective integration window, we define three operational quantities. The first is a Born distance D_B(Delta t), measuring the deviation between the observed outcome distribution at resolution Delta t and the corresponding Born-rule prediction. The second is a convergence-rate proxy gamma(Delta t), defined as the negative derivative of D_B with respect to the logarithm of Delta t, which quantifies the tendency toward convergence to Born statistics as temporal resolution is increased. The third is a hysteresis-based memory witness M(Delta t), defined as the absolute difference between Born distances obtained from increasing and decreasing Delta t sweeps. We then state a conditional trade-off for stateful closure explanations. Over Delta t ranges where these quantities are well defined, a positive convergence tendency gamma(Delta t) greater than zero is generically expected to be accompanied by nonzero path dependence M(Delta t) greater than zero. This expectation can be summarized by an upper bound of the form: gamma(Delta t) is less than or equal to C times M(Delta t), plus a correction term epsilon(Delta t). Here C depends on coarse structural features of the closure process, such as the number of competing channels and the sharpness of decision thresholds, while epsilon collects finite-sample and finite-resolution effects. An efficiency ratio eta(Delta t), defined as gamma(Delta t) divided by M(Delta t), is introduced as a diagnostic. Within the null family, where M(Delta t) is identically zero, any sustained claim of finite positive gamma necessarily drives eta to diverge, signaling incompatibility with the readout-only, memoryless assumptions. The framework does not propose new dynamics or modify quantum theory. Instead, it specifies an operational protocol and a coupled signature set whose robust observation would render the readout-only, memoryless null family insufficient as an explanation of finite-time Born convergence. Robustness controls, including sweep-rate variation, randomized Delta t ordering, and mapping-invariance checks, are outlined to separate decision-level path dependence from trivial instrumental artifacts.

•Jan 5, 2026•0 citations

Martingale Collapse in Quantum Measurement A Record-Gated Dynamical Framework for the Emergence of Single Outcomes and Born-Rule Weights

Fahad Almukairish

Abstract Standard quantum mechanics asks us to live with two laws of time. The first is the continuous, linear, unitary evolution of the Schrödinger equation. The second is the discontinuous, non-unitary reduction that appears when a measurement yields one result rather than many. We have learned to use this double description with remarkable success, yet the success of the practice does not remove the tension in the principle. This paper proposes a conditional framework, Record-Gated Martingale Collapse (R-MC), for a narrower and sharper question: how can a state that has branched through ordinary unitary interaction among a system, an apparatus, and an environment settle into a single recorded outcome without adding an external projection rule? The framework begins where ordinary dynamics leaves us: unitary branching and decoherence. It then adds three elements. First, a non-selective record gate opens the collapse mechanism only after inter-branch overlap has weakened and a sufficiently independent environmental record has accumulated. Second, the collapse rate is amplified by the effective number of independent recording channels, not by the raw number of environmental degrees of freedom. Third, an uncontrollable objective seed generates the stochastic path along which branch weights move. Under quantum non-demolition (QND) conditions these weights are bounded martingales that diffusively localize toward a single branch, and Born-rule weights follow from conservation of their mean rather than from a postulated projection. We separate this gradual convergence from the operational recording time, and we distinguish a safe external form, in which the gate depends on local environmental variables and the ensemble mean closes into a linear Lindblad equation, from a more ambitious internal form that requires operational locality of the gate and locality of the apparatus pointers. A distinctive prediction, record-threshold collapse delay, states that two environments with comparable decoherence may nevertheless yield different outcome times if they differ in their capacity for independent, amplifiable recording.

•Jun 28, 2026•0 citations
This deposit contains the manuscript and the three analysis scripts (with outputs and SHA-256 hashes) for Part III of the SFIT Stochastic-Braid Substrate series. The paper separates three objects that are often conflated in attempts to obtain quantum measurement statistics from a classical substrate: (i) the substrate configuration (positions, velocities, helicities of the Part I defects), (ii) the channel population (x_k \ge 0), (\sum_k x_k = 1), and (iii) the record probability (P(k)), i.e. the probability that a physical record names channel (k). The Born rule relates (iii) to (ii); it says nothing by itself about how (ii) arises from (i). The paper asks three questions and reports the status of each: 1. Can a shared, finite carrier enforce exclusive records? In a coupled-mode model with explicit local detectors, detector-local back-reaction lowers the gated coincidence parameter from (\alpha = 1.00) to (\alpha = 0.57\text{–}0.60) at the strongest back-reaction tested, but only when detector thresholds are comparable to the carrier energy. At thresholds one-twentieth of the carrier energy, (\alpha) stays at (0.96\text{–}1.00). Strong exclusion ((\alpha < 0.1)) was not reached in any of 136 configurations. Result: partial suppression only. 2. Under what conditions does record selection reproduce Born weights? Two sufficient routes are identified, each with a precisely quantified fragility. Route 1: exactly memoryless (exponential) detector thresholds give Born frequencies by a common-time-change argument; gamma thresholds of shape 2 depart by TV distance (0.08\text{–}0.15). Route 2: a bounded martingale for channel populations absorbed at a vertex gives Born weights for any functional form of the exchange law; a first-passage bound and a converse are proved (given absorption, zero conditional drift is necessary), and a biased exchange kernel is shown to follow the Kimura fixation formula, so zero drift must hold to order (1/M) in the number of exchange units. Common-mode noise on the shared carrier cannot substitute for detector-local randomness: with fixed thresholds the largest channel fires first in every run, at every reversal rate. 3. Does Part I supply these conditions? No. The Part I helicity representation acts by permutation matrices, and no permutation action yields zero conditional drift for any filtration unless it is the identity (for pairwise-distinct populations). Measured on the Part I crossing process, the drift is nonzero at every test state, and populations are rearranged but never concentrated; sector populations are exactly invariant. This identifies braiding as neutral with respect to record selection. Scope and caveats. No derivation of the Born rule or of quantum superposition from the substrate is claimed. The internal state (\psi) is an input; the exchange kernel of Route 2 is global and postulated; memoryless detector thresholds (Route 1) are assumed; strong exclusion was not achieved; and the first-passage record of Proposition 3 ((\alpha = 0) by conservation) has not been combined with local detectors in simulation. Real amplitudes suffice for the single-system statistics studied, but real-amplitude quantum theory is experimentally distinguishable from complex quantum theory in network scenarios. Each result in the manuscript is tagged [DERIVED], [ASSUMED], or [TESTED]. Three scripts produce every number in the paper; each states its model, metrics, controls, predictions and decision rules in its header, fixed before the runs reported. The scripts were not preregistered; this deposit supplies them together with their outputs and hashes to support reproducibility.
Part II of the SFIT Stochastic-Braid Substrate series asks which additional structure would be required for outcome records to occur with Born-rule frequencies ∣ψk∣2|\psi_k|^2∣ψk​∣2, and how much of that structure the substrate itself could plausibly supply. Part I established that persistent stochastic defects generate braid-group structure with orthogonal representations preserving ∥ψ∥2\|\psi\|^2∥ψ∥2, but made no claim about outcome probabilities; pure braiding cannot produce definite records, since orthogonal invertible maps cannot drive the internal state to a basis state. This paper studies a classical absorber race. Each channel kkk has an environment oscillator driven at resonance by a force linear in the carrier amplitude ψk\psi_kψk​; a record is produced by the first channel whose absorbed drive work exceeds a random threshold. The quadratic dependence of absorption on ψk\psi_kψk​ is derived rather than assumed: a harmonic absorber gives Pk=g2ψk2/(2γ)P_k = g^2\psi_k^2/(2\gamma)Pk​=g2ψk2​/(2γ), while an anharmonic (Duffing) absorber yields fitted exponents of 1.65 at weak drive and 0.74 at strong drive, showing that the quadratic law holds only in the linear-response regime. With exponentially distributed thresholds, record frequencies agree with ∣ψk∣2|\psi_k|^2∣ψk​∣2 within sampling error for all tested states in n=2,3n = 2, 3n=2,3 and 4 channels. With fixed, uniform, or gamma-distributed thresholds of shape k>1k > 1k>1, frequencies depart from Born values toward the largest weight - already at a total-variation distance of 0.09 at k=1.5k = 1.5k=1.5. This is the behaviour characterized by Yellott's theorem: within races with independent, identically distributed thresholds, choice probabilities proportional to rates for three or more alternatives require exponential thresholds. The paper distinguishes this n≥3n \geq 3n≥3 condition (number of competing alternatives) from the n≥3n \geq 3n≥3 condition in Gleason's theorem (state-space dimension). Oscillator simulations match the exact race integral throughout, and thermal noise up to kBT=0.1k_BT = 0.1kB​T=0.1 leaves frequencies unchanged when records are triggered by drive work. The result is explicitly conditional. Linear coupling, steady drive, channel independence and symmetry, memoryless thresholds, and mutual exclusion of records are assumptions of the model, not consequences of Part I; Table 2 lists all conditions and their status. The paper identifies memoryless thresholds (potentially supplied by the exponentially timed velocity reversals of Part I) as the most promising substrate contribution, states a pre-registered decision rule for testing it, and proposes further tests (start-from-rest ramps, total-energy triggers, over-dispersed thresholds). No derivation of the Born rule from the substrate is claimed; the mechanism is presented as a candidate, with the task for SFIT made specific. Simulation code (stage_c_absorber.py) reproducing all numbers from fixed random seeds accompanies the preprint.
This paper investigates whether the one-dimensional stochastic dynamics of the SFIT stochastic-braid substrate can supply the conditionally fair exchange sign required for Born-weighted record formation. Building on Part III, which established that Born-weighted outcomes follow when channel populations form a bounded martingale absorbed at a vertex, we examine two possible sources of exchange signs in the one-dimensional ring dynamics: velocity reversals and defect-crossing orientations. The analysis identifies two distinct obstructions. First, velocity-reversal signs alternate for each defect, producing systematic second-order drift rather than conditionally fair exchange. Second, crossing orientations are constrained by the cyclic order of defects on the ring. We prove that this orientation is determined by the preceding crossing history and therefore cannot supply a conditionally fair sign. This is a topological obstruction, independent of the particular population jump law whenever that law responds to the crossing sign. For the carrier-mediated exchange law studied here, the obstruction produces a measurable departure from Born-weighted outcomes. At reversal rate γ = 1, the resulting bias is approximately +0.056 in the probability of the initially largest channel, remains essentially unchanged across the tested exchange amplitudes, and is statistically distinct from the martingale prediction. We then construct a reduced predictive model using only the substrate's measured pair-crossing statistics and the topological orientation rule. The model predicts record probabilities without fitting parameters to carrier outcomes. Across multiple reversal rates, initial population states, and time-step settings, its predictions agree with independent carrier simulations within the stated uncertainty. An ablation retaining only marginal pair frequencies fails to reproduce the reversal-rate dependence, demonstrating the importance of pair-sequence memory. At γ = 4, convergence of the finite-history model is not established, and that case is explicitly reported as a model limit rather than a successful validation. The results establish that the one-dimensional ring-crossing mechanism does not provide the fair-sign condition needed for the martingale route to Born-weighted records. Instead, the substrate's deterministic topological orientation constraint, together with the memory of its crossing sequence, predicts the resulting departure from Born statistics. The study is limited to the one-dimensional ring configuration and the specified carrier-mediated jump law. In particular, it does not settle whether the braid-generator sign in the planar dynamics of Part I can be conditionally fair. That remains a separate question for genuine two-dimensional braid geometry. Part IV therefore provides a topological no-go result for the one-dimensional fair-exchange route and a quantitative, independently tested account of the record bias that replaces it.
Can a Born-type probability structure arise on a substrate that is only a finite stochastic process? A stochastic program built from six primitive operations supplies the final matrix; a specified layered architecture constructs, from its outputs, a record layer, a two-dimensional real carrier and a quadratic law, and supports an adopted Born calculus. The six operations alone do not give this structure, and the square-root encoding and the projection are construction choices. Our main contribution is the architecture: a directed acyclic graph whose arrows are each marked as derived, chosen, adopted or standard mathematics, with the kind of check behind each. On registered populations, a snapshot predictor whose rank-two case is the law predicts the substrate's sampled label frequencies with lower aggregate loss than named competitors, and, in the construction's canonical basis, its cross terms carry its advantage over the diagonal rule. In each of 2,122 census cases the encoded state is not a function of the anonymous initial record description: two point preparations sharing all three labels differ by at least 1/25 in the trace probability of raw label 0, a statement about exact data, not operational access. This bound is retrospective: the registered bound 3/20 fails in 117 cases of the full census. On the original worked case the updates by the final matrix and its square do not lift to the carrier. The measurement rules are adopted: Lüders tests, all densities and effects, and additivity; on them the trace rule is unique, and finite ontic models with additive responses, and measurable preparation-noncontextual representations preserving the specified mixtures, are excluded. We do not claim that the substrate produces the measurement rule, that the layer forms in the run, or a derivation of quantum theory. The source code of the stochastic program, the constructions, the reproduction pipeline, the verification commands and the Lean 4 proofs are available at https://github.com/ioannist/six-birds-born (release v1.0). AI systems were used throughout this work, as the manuscript describes; the author directed the research, made the scientific decisions and is responsible for the content.
This theoretical and computational preprint integrates standard QND measurement and posterior-martingale results with finite-confidence decisions and physically retained observer records. Under a declared quantum instrument and its standard outcome probabilities, the posterior is a bounded martingale on an increasing observation filtration. A precise sufficient identification condition yields asymptotic terminal weights; a binary Gaussian benchmark gives exact mean stopping times, finite-threshold latch probabilities, and a controlled population-transfer bias bound. Version 2.3 is the first public release. It distinguishes accumulated observations from lossy current memory; assumed Born/instrument probabilities from derived consistency results; expected entropy from pathwise ambiguity; continuous confidence endpoints from discrete overshoot and unresolved runs; and Kramers first escape from endpoint readout error. It also clarifies the conditional D18/D19 persistence-selector interface and the separate D1092 finite-resource ledger, with direct attribution to established QND convergence and filtering literature. The companion archive includes the editable manuscript, PDF, equation sources, a fixed validation protocol, executable independent checks, numerical results, a revision ledger, and unchanged historical inputs. Historical simulation summaries are explicitly distinguished from new executions; some older executable sources were unavailable. The new audit includes exact finite-kernel and stopped-posterior checks, Gaussian likelihood-mixture calculations, five neutral and four population-transfer boundary-value cases, and retention/accounting controls. All examples are analytic or synthetic. No laboratory measurements, independent Born-rule derivation, measurement-problem solution, universal physical selector, or new empirical anchor are claimed.
Quantum mechanics predicts outcome statistics via the Born rule, yet remains operationally under-specified with respect to how, and on what timescale, a single outcome becomes definite in an experimental record. In many platforms, identical Born-compliant outcome histograms can be produced either by a time-extended physical closure process, in which competing channels accumulate influence until a decision event, or by purely readout-level mappings and post-processing applied to measurement records. This under-determination motivates an operational question: which broad classes of readout-based explanations remain viable given finite-time, resolution-dependent data alone? Here we formulate an operational exclusion criterion relative to a broad null family of readout-only, memoryless models. In this null family, for each temporal resolution Delta t, the reported outcome is generated by a (possibly stochastic and nonlinear) mapping from a single-shot record R to an outcome Y, with no decision-level internal state persisting across Delta t settings and no dependence on sweep direction or parameter history. Under these structural assumptions, sweep-history dependence is excluded by construction. Using Delta t as a single experimentally controlled parameter, interpreted as an effective integration window, we define three operational quantities. The first is a Born distance D_B(Delta t), measuring the deviation between the observed outcome distribution at resolution Delta t and the corresponding Born-rule prediction. The second is a convergence-rate proxy gamma(Delta t), defined as the negative derivative of D_B with respect to the logarithm of Delta t, which quantifies the tendency toward convergence to Born statistics as temporal resolution is increased. The third is a hysteresis-based memory witness M(Delta t), defined as the absolute difference between Born distances obtained from increasing and decreasing Delta t sweeps. We then state a conditional trade-off for stateful closure explanations. Over Delta t ranges where these quantities are well defined, a positive convergence tendency gamma(Delta t) greater than zero is generically expected to be accompanied by nonzero path dependence M(Delta t) greater than zero. This expectation can be summarized by an upper bound of the form: gamma(Delta t) is less than or equal to C times M(Delta t), plus a correction term epsilon(Delta t). Here C depends on coarse structural features of the closure process, such as the number of competing channels and the sharpness of decision thresholds, while epsilon collects finite-sample and finite-resolution effects. An efficiency ratio eta(Delta t), defined as gamma(Delta t) divided by M(Delta t), is introduced as a diagnostic. Within the null family, where M(Delta t) is identically zero, any sustained claim of finite positive gamma necessarily drives eta to diverge, signaling incompatibility with the readout-only, memoryless assumptions. The framework does not propose new dynamics or modify quantum theory. Instead, it specifies an operational protocol and a coupled signature set whose robust observation would render the readout-only, memoryless null family insufficient as an explanation of finite-time Born convergence. Robustness controls, including sweep-rate variation, randomized Delta t ordering, and mapping-invariance checks, are outlined to separate decision-level path dependence from trivial instrumental artifacts.
Abstract Standard quantum mechanics asks us to live with two laws of time. The first is the continuous, linear, unitary evolution of the Schrödinger equation. The second is the discontinuous, non-unitary reduction that appears when a measurement yields one result rather than many. We have learned to use this double description with remarkable success, yet the success of the practice does not remove the tension in the principle. This paper proposes a conditional framework, Record-Gated Martingale Collapse (R-MC), for a narrower and sharper question: how can a state that has branched through ordinary unitary interaction among a system, an apparatus, and an environment settle into a single recorded outcome without adding an external projection rule? The framework begins where ordinary dynamics leaves us: unitary branching and decoherence. It then adds three elements. First, a non-selective record gate opens the collapse mechanism only after inter-branch overlap has weakened and a sufficiently independent environmental record has accumulated. Second, the collapse rate is amplified by the effective number of independent recording channels, not by the raw number of environmental degrees of freedom. Third, an uncontrollable objective seed generates the stochastic path along which branch weights move. Under quantum non-demolition (QND) conditions these weights are bounded martingales that diffusively localize toward a single branch, and Born-rule weights follow from conservation of their mean rather than from a postulated projection. We separate this gradual convergence from the operational recording time, and we distinguish a safe external form, in which the gate depends on local environmental variables and the ensemble mean closes into a linear Lindblad equation, from a more ambitious internal form that requires operational locality of the gate and locality of the apparatus pointers. A distinctive prediction, record-threshold collapse delay, states that two environments with comparable decoherence may nevertheless yield different outcome times if they differ in their capacity for independent, amplifiable recording.