The measurement-theory core: what a record is, why records are classical, and how the record/response split becomes theorems.
For computer scientists
Algebra Before Bit: A Computer Scientist's Account of a Finite Record Universe
Quantum physics for the computer scientist: the whole finite record framework explained through a programmer's console, with no quantum mechanics assumed. A short primer teaches the four ideas needed to read the display — qubits, superposition, entanglement and measurement — and the paper then opens the data structures: quantum states as typed structures, interactions as typed rules, measurement as a state change plus a record write, and the cell, register and gauge-bridge layout of the substrate drawn as an entity model. Two questions organise the tour. Where is the information behind a Bell correlation stored — in the two local records, in the description of the pair, or in the procedure that produces the answers? And how much memory does a description need — a flat table for a 300-qubit state has 2³⁰⁰ entries, while a structured family of states on the same qubits packs into under 50 kB, and the paper says exactly which states fit and why. Measurement is treated as a commit protocol, nonlocality as the price of a clock, and the closing sections ask which parts of the console a classical computer can run and which physical rules the framework still has to be given. A small executable Bell example, a numerical verifier and an SSADM-style data, process and life-history specification in Appendix A make the description checkable. It is a representation, not a hypothesis: nothing here claims the universe is a simulation, and nothing depends on it being one. Published here in full as searchable HTML as well as PDF.
Technical note
Described Twice? A Finite Audit of a Possible Gravity–Measurement Seam
The programme's most direct approach to the quantum-gravity question, and the one that has moved furthest. The premise is a disciplined version of an old move: physics has repeatedly unified descriptions through conversion laws and invariants, so when quantum measurement and weak-field gravitational sourcing keep attaching their bookkeeping to the same events, is that two mechanisms or one in two languages? On 50,401 frozen histories the commit and billing ledgers agree in number and timing under the declared one-bill-per-commit convention, and the later audits sharpen that count into an exact jump–tape–commit correspondence — but a correspondence of events is not a billing law: per-commit, per-tape and erasure-only billing remain distinct conventions, none is selected, and the additive history currents that now exist, the full ℤ[i]²²⁴ of Gaussian-integer weightings on the 224 exit events, carry weights the event count leaves free. The finite-C4 current ring is ℤ[i], but no absolute pair-unit-to-bill map follows: the pair unit is nondefinable from the retained algebra and observables and stands as one declared normalization input. The sharper development is the discriminator. Two exact pre-commit completions agree on the committed face — quantum-until-commit Q and records-only R — and their two-source extensions can be told apart: R is separable throughout its frozen class, Q is entangled exactly when an invariant cross phase Ξ is nontrivial, and a single common 36-outcome local instrument separates them, now with an exact entanglement threshold for independent local dephasing (this revision corrects the earlier threshold). What remains underdetermined is the number, not the procedure: two admissible control-to-phase maps give opposite decisions at the same control and noise point. Further audits fence the space — the service ceiling bounds record-coherence attenuation rather than entanglement growth, additive rational control composition is impossible in the Gaussian-rational phase carrier, and the system-only action grammar cannot reproduce either the nonunitary record channel or intercell record transport — and inside the declared four-element enlargement lattice the local-environment and record-transfer sectors are each necessary and jointly the unique minimal completion, a minimality relative to that lattice that claims no uniqueness among all open-system actions. New in this version is a stated boundary on repetition and time: the first-cycle channel does not determine repeated-use dynamics, a pointer read need not be a new commit, and physical timing, service scheduling and history weights all remain unselected. The paper still establishes neither unification nor a null result, and its closing section says precisely what would change that. Four arrows have to close first: a source-bound calibration and amount link, since the exact event bijection supplies neither an amount nor a bill; a physical action, phase law and clock, the minimal-enlargement theorem having typed the sectors without adopting their maps, selecting coefficients or deriving laboratory time; physical invocation and empirical selection, since a laboratory implementation must establish that its locality, mediator and noise premises actually hold and the programme must justify the step from the proved finite record correspondence to physical increments; and the continuum and covariance lift, an emergent-Lorentz proposal inheriting rather than evading the obligation to recover an adequate relativistic response within the Weinberg–Witten constraints. Only once those close does the stronger sentence become available — “measurement is the quantum description and gravity the macroscopic accounting of one record process” — and the paper is explicit that, for now, that sentence is a destination rather than a conclusion.
Technical note
When Does a Quantum Transition Become a Record? Instruments, export, and cadence in finite quantum dynamics
A finite transition table is not yet a measurement record. This report separates six objects that are often collapsed in discrete models — transition support, effects, instruments, environmental export, invocation cadence, and autonomous dissipative dynamics — and gives exact finite witnesses for the failed implications between them. Exact calibration fixes the visible effects but not the post-measurement instrument; positive-semidefinite environment Gram matrices realise the same basis dynamics with no, partial, or projective record export; endpoint maps do not determine cadence; and a lawful Lindblad jump family still has an independent rate. The worked example is explicitly the eight-register cell from the finite-QEC substrate programme, not an unmotivated toy, but the programme-level service-clock, gravitational, and cosmological interpretations are kept as non-claims. Its 256 states and 2,048 addressed events are exhaustively checked: 48 physical states, nine truthful record sectors, and distinct positive-control, complementary-control, and polarity-complete exit supports of 104, 120, and 224. The Zenodo deposit includes the PDF, source, verifier, and machine-readable JSON/CSV reports.
Technical note
Pointer States Are Not Enough: Physical Monitor Selection in a Finite Quantum Register
Decoherence does not select a basis in the abstract — it selects stable sectors through a particular system/environment interaction, so the open-system channel is part of the statement. This report separates three things usually run together: a candidate pointer basis, the algebra generated by the microscopic operations actually available, and the pointer structure of a channel that has been physically selected. The distinction is made exact on an eight-qubit register whose qubits sit on the triangular faces of a bond-centred oblate square bipyramid, with the cube graph Q₃ as adjacency: twelve commuting edge-parity observables have rank seven, leaving 128 two-dimensional complement sectors, which on the 48 valid register labels resolve into eight rank-two and 32 rank-one sectors. What the service operations then achieve turns on which records are physically exported. Full polarity-resolved J/K support individuates 32 of 48 labels; the banked elementary positive-branch event leaves J alone individuating only 22, with residual multiplicity in all three generations — yet the generation-00 noiseless factor survives both, support-robust on the same fixed 48-state closure. The sharp result is an exact no-go: coherent evolution, an address-only L record, and polarity-resolved J/K records can share the same 224 exit edges while preserving different coherences and generating different resolving algebras. Only polarity-resolved export yields the conditional finite hierarchy 01 < 10 < 00 — executing the missing edges is not enough. So the open task is not rate-tuning but deriving a polarity-complete schedule, keeping its fired-arm flag, exporting that flag to a fresh orthogonal environment record, and licensing branchwise recovery. Three interference/rate experiments separate the clauses exactly.
Technical note
The Ledger Is Not Enough: Counts, Memory and Observable Content in a Finite Record Framework
The natural sequel to Described Twice? and Pointer States Are Not Enough, and the paper that asks the question those two kept deferring: once an exact ledger of reads, commits and service events exists, does it also fix what happens next? In a finite quantum model with 48 preparations and a ten-outcome apparatus the answer is no, on three separate counts. Two reversible evolution rules reproduce the same first-use specification exactly and then disagree the moment the apparatus is used again, so the first-use map does not select the law. With one law fixed, two histories can carry identical counts of reads, commits and bills and still assign different probabilities to the next outcome, so the count ledger is not a predictive state. And what survives depends on what is looked at: at eight preparations the two laws agree on every single-read distribution and are told apart only by the correlations between reads — an exact six-label cancellation that leaves the whole contrast in a zero-marginal correlation block. Two declared families of observations are classified, with exact ranges for the simpler one, separating event accounting from the choice of evolution, from retained memory, and from the response, weighting and sampling rules that turn records into measured numbers. The results are conditional finite-model counterexamples and classifications — not new quantum principles, and not a derivation of a dynamics or a coupling — and the closing table lists precisely which inputs a predictive physical interpretation still has to supply. Every calculation is exact and reproduced by a bundled verifier. This paper is also published here in full as searchable HTML, not only as a PDF.
Technical note
Records and Responses
The rigorous operator-theoretic core behind the measurement discipline: a class of monitored quantum systems — latched instruments — in which every observable splits exactly into a record class (diagonal in a fixed, dynamically absorbing pointer decomposition) and a response class (all the off-diagonal support). Five short, machine-certified theorems follow. (T1) Record statistics form a single classical probability space — coincidence rates factorize at diagonal Born weights and every retarded correlator built from records vanishes identically. (T2) Pointer-conditional dressing leaves the entire record algebra invariant as an operator identity while renormalising responses by Franck–Condon overlaps: the same interaction cloud leaves the bills fixed and changes only the responses. (T3) The record channel's LSZ pole residue is exactly 1 and cannot be renormalised, whereas a bare response excitation carries residue Z < 1 with the deficit living entirely in the dressing continuum. (T4) Latched accumulation is a functional of the source alone, independent of any washout applied to unlatched components — each event is billed once. (T5) The additive contact term to which fluctuation–dissipation and Kramers–Kronig reconstructions are provably blind is fixed by the Euclidean zero-frequency value, so local subtraction constants belong to the record side of the split. Each theorem ships with a self-checking program, and the paper is explicit about which parts are standard material made exact and which readings are new. Applications: measurement statistics, Quantum-Darwinism redundancy, spectator-charge accumulation, and the normal-ordering constants in dressed couplings — the same counter/kernel split that separates the bare α₀ = 1/137 from its dressed value.
Technical note
Records and Responses in the World: A Derived Fine-Structure Boundary, Certified Confinement Gaps, and Registered Discriminators for a Finite Record Substrate
The physics companion to the operator-theoretic Records-and-Responses note above: what happens when a research programme takes the record/response split seriously as an executable accounting rule — bills cannot be renormalised, meters always are. Working over the finite error-corrected substrate whose record layer was compressed and certified in the record-grammar note, every claim is tier-labelled and every number re-derivable from a named self-checking program. Four confrontation fronts. (i) A derivation of the fine-structure boundary, α⁻¹_FW = 137.035999107 — the record-grade count of 137 interface channels read through a computed dressing branch (an endpoint-current identity, a single contact term, and the second-order service kernel K₂, assembled Euclideanly exactly as the companion's T5 instructs, with no free choices). The number lands on the contested side of metrology's own 5.5σ caesium–rubidium recoil disagreement (+2.3σ from Cs, −9.0σ from Rb, −3.9σ from the electron anomaly through an unmodified perturbative series) — next-generation recoil metrology adjudicates a stake the framework can no longer move. (ii) An electroweak one-anchor route reduced to three response legs, with a registered top mass M_t = 172.69 GeV, a Higgs boundary narrowed to λ(M_Pl) = −Cα₀ with C ∈ {1,2}, and a vacuum-value candidate v = (15/16)α₀⁸M_Pl/√λ_eff whose distance to lock is fully quantified — the experimental endpoint is δm_H ≲ 55 MeV, the same HL-LHC measurement two independent parts of the programme bottom out on. (iii) A confinement statement pushed to certificate grade: the mirror-sector gap of the substrate's chiral (symmetric-mass-generation) embedding carries rigorous coupling-uniform floors 2.63–3.00 across an exact volume ladder, a conditional infinite-volume floor Δ∞ ≥ 2.35, and exactly two named limit legs remaining. (iv) A gravitational and transport response atlas — black-hole thermodynamics and greybody structure reproduced with superradiance reread as unbilled response, a zero-parameter substrate-noise fingerprint Γ = D(2π/L)² at fixed D = 1.20×10⁻⁷ m² s⁻¹, and quantified nulls protecting the framework from its own most convenient stories. Five predictions are pre-registered with timestamps, eleven branches the framework killed itself are listed with dates, and the paper closes with a falsification map naming which experiment ends which claim.
Technical note
From Counts to Observables: The Response Layer of a Finite Record Substrate — a Bridge for Physicists, Information Scientists, and Engineers
The bridge paper of the records-and-responses family, written to be readable from three directions at once — physics, information science, and engineering. Every measurement chain an engineer has ever calibrated divides an indication by a transfer function to recover an invariant, and since the 2019 SI redefinition the invariants at the bottom of every such chain are counts — fixed integers — with everything instrument-shaped being transfer. This paper takes that architecture seriously as physics. In a finite record substrate, dimensionless constants arise as exact rational counts — shares of a finite service ledger — while experiments only ever read responses: in-in (closed-time-path) correlators driven by a probe. Records say what can be known; responses say what experiments measure. The bridge is formalised as five theorems, each verified by a self-asserting computation: (i) the monitored theory splits into a commuting, copyable record algebra and a non-commuting response algebra; (ii) a collapse theorem — for a latched record channel the spectral function is an equal-time contact whose residue is exactly the count, so every probe observable factorises as O = T(probe; scheme) × r; (iii) rigidity — counts have no anomalous dimension: scheme changes move the transfer factor T, never the count r; (iv) a five-front ledger showing that the observables where bare counting historically stalled — QED, the electroweak sector, black-hole emission, the CMB, and continuum QCD — are exactly the fronts where T ≠ 1; and (v) a reading law — Born weights are count shares at the latch, with interference confined to the response layer. The classical limit recovers calibration practice — Wheatstone bridges, lock-in detection, Kalman observability — as the commutative special case, which is why engineering developed the split without ever naming it. It closes with a forward path for quantum information theory: latching as a free operation, record capacity as an operational measure of classicality, a conjectured record/response complementarity, and stabiliser codes re-read as engineered record algebras with calibration-free syndrome statistics.
Technical note
A Selection-Rule Calculus for Finite Record Physics: Static Predicates, Monitored Recovery Instruments, and Physics-Bearing Environment Records
The newest member of the records-and-responses family: a calculus for deciding when a selection rule is physics and when it is only bookkeeping. Selection rules are usually treated as allowed/forbidden labels — a state is admissible if it obeys the rule and absent if it does not. In a finite record-bearing substrate that is not enough: a static predicate can define a codespace, but it writes no environment record, carries no phase, produces no entropy, and cannot by itself generate a measurable response. A selection rule becomes physics-bearing only when it is implemented as a monitored recovery instrument — syndrome bits, a recovery map, and a Stinespring environment whose orthogonal labels are the copyable records later physics can condition on. The central theorem is a register-access rule: a monitored constraint can write records only in the registers its syndrome and recovery actually read — so a colour rule cannot supply a generation phase, a sterile repair cannot generate a generation covariance, and a chirality lock cannot repair a colour-counting deficit. The worked example is the framework's Dynamic-R1 generation mechanism: the forbidden fourth-generation corner G0G1 = 11 is not merely absent — under monitored recovery it is an active boundary whose allowed generation order ideal {00, 01, 10} writes Hasse-edge environment records; the symmetric second moment of those records gives the Koide covariance block K_R1 = BBᵀ, while the closed oriented cochain Ω_R1 carries the Majorana/CP orientation sign. Composed with the sector defect inventory, the CP-even Koide row closes in its type-correct form — the absolute contact ledger (e, ν, d, u) = (2, 3, 3, 2) over N_eff = (9, 9, 27, 27). The calculus is deliberately disciplinary as well as constructive: it explicitly forbids using this closure to rescue the baryogenesis magnitude η = (3/14)α₀⁴ — the numerator 3 remains an ideal-code count, not a physical B−L source count. The slogan: static predicates select states; monitored recovery writes records — and a monitored rule can only write what it reads. Every claim is implemented by a named executable gate, and the cleanest experimental consequence is inherited by construction: leptonic Dirac CP null in long-baseline oscillations, with CP living in the Majorana/recovery orientation.
Working note
The Born Rule as a Closed Record Pair: Measurement, Objectivity, and the Arrow of Time
A measurement is not primitive — it is the substrate recording a syndrome. Revised 30 August 2026, and deliberately more careful than the previous version: two complementary but *non-equivalent* arguments constrain the Born rule inside the accepted complex-QEC substrate, rather than jointly closing it. Repeatable syndrome readout gives orthogonal, non-contextual projectors, so Gleason-type uniqueness fixes pᵢ = Tr(ρPᵢ) — but only given an additive frame function, a premise the revision now names explicitly. The Naimark/Stinespring record map makes measurement an isometric copying of syndrome facts into orthogonal record sectors, whose closed forward/backward history leaves the surviving diagonal AA* = |A|²; that route earns the quadratic form from strictly weaker, non-probabilistic premises — no measure or extensive valuation is introduced — so it is no second uniqueness theorem, but it does avoid assuming measure-level additivity. The reconstruction frontier is stated as two problems, not one: get uniqueness without importing that additive premise, and then the deeper floor — why nature is a complex, locally tomographic, record-writing QEC system with the monitored service alphabet at all. Also fixes the byte: the unique minimal distance-4 record cell in the binary balanced Type-II CSS class is the self-dual doubly-even [8,4,4] code, and α₀ = 1/137 is the uniform Born weight of one label in Sym²(16)+1.
Working note
Quantum Darwinism as Syndrome Broadcast: A Finite-QEC Reconstruction of Objective Records
Why does the everyday world look objective — the same for every observer? Quantum Darwinism already answers that classical facts are information about einselected pointer states copied redundantly into many fragments of the environment; this note reconstructs that mechanism from the single premise of stable local records. Repeatable records force orthogonal sectors; finite noise forces an error-correcting code; syndrome extraction selects the pointer basis; syndrome fan-out supplies the redundant environmental copies; and finite, reusable registers make the irreversible part of measurement a Landauer reset. In the ideal syndrome-broadcast limit the characteristic Darwinist plateau is immediate — any non-empty proper fragment of the environment carries the full classical syndrome entropy, while the extra quantum phase information lives only in the whole. Deliberately modest: it does not replace decoherence theory, derive the Born rule, or solve the measurement problem; it closes by mapping the programme's older geometry-first language (cells, strain, defects) onto the finite-QEC vocabulary (code, syndrome, ledger).
