KEX-BRAINK
Axiomatic Research Protocol and Substantiation Standard
Document schema, evidence architecture, mathematical proof workflow, experimental programme and grading system
CORE RULE
No proposition is rejected merely because it conflicts with a prevailing mathematical, scientific or engineering standard. Disagreement is recorded as model divergence. The proposition is then tested for well-definedness, internal consistency, derivability, computability, empirical correspondence and discriminating predictive power.
Executive Directive
This protocol converts the KEX-BRAINK research corpus into a controlled system for mathematical derivation, computational verification, empirical discrimination and independent replication. It was created specifically to avoid a recurrent methodological error: treating disagreement with an established mathematical or scientific convention as an automatic invalidation. Under this protocol, a prevailing standard is a comparator model. It may be correct, incomplete, differently scoped or based on axioms that KEX rejects. The conflict is recorded, shared observables are identified, and a discriminating prediction is designed.
The reciprocal safeguard is equally important. Departure from a conventional model is not itself proof. A KEX proposition earns substantiation by being defined precisely, internally consistent, derivable from explicit axioms, computationally reproducible, linked to measurements with coherent units, and capable of making predictions that distinguish it from alternatives. Mathematical proofs establish consequences conditional on the declared axioms. Physical, economic, biological, lifecycle and infrastructure claims additionally require a measurement map and data.
Document Control
Contents
1. Mandate, scope and epistemic contract
2. Source corpus and data-access architecture
3. Dual-lane validation model
4. Research object hierarchy and machine schemas
5. Status lanes and claim-promotion grammar
6. Universal execution workflow
7. Mathematical substantiation standard
8. Model-divergence adjudication
9. Experiment and measurement programme
10. Data, provenance and replication controls
11. Research grading and analysis
12. Required rounds of work
13. Governance, change control and publication
14. Package contents and operating instructions
15. Appendices
1. Mandate, Scope and Epistemic Contract
1.1 Mandate
The KEX-BRAINK corpus spans typed-state semantics, the Corrected Zero Doctrine, Singularity-One, Observer-Origin, relational direction, the BRAINK 3|2|1|2|3 architecture, workbook-native execution, deterministic state continuity, KCPU route selection, HEX×HEX folding, thermodynamic accounting, TER/ELC lifecycle modelling, Hardgate security and research-agent orchestration. The next research round must not reduce this portfolio to a binary accepted/rejected judgment. It must identify which parts are axioms, definitions, derived results, implementations, measurements, hypotheses, model divergences and deployment claims.
The protocol therefore governs both positive and negative results. A contradiction is preserved. A failed experiment narrows scope. A comparator that fits the same observations is retained. A KEX result that survives its own assumptions but has not yet been physically measured remains a conditional theorem or hypothesis, not an invalid statement and not an empirical fact.
1.2 Scope
Formalisation of KEX-BRAINK axioms, state types, functions, invariants and domains.
Proof and computer-assisted verification of finite or symbolic propositions.
Reframing of historical claim dispositions without deleting their provenance.
Explicit comparison between KEX and conventional models at the level of assumptions and observables.
Preregistered software, thermodynamic, lifecycle, security and cognition experiments.
Controlled measurement dictionaries, immutable data manifests and signed evidence bundles.
Reviewer scoring that is independent of prior labels and does not reward orthodoxy by default.
Separation of research validity from production, network, persistence, security and infrastructure readiness.
1.3 Epistemic Contract
Principal basis: S01, S02, S05, S06, S07 and S08.
2. Source Corpus and Data-Access Architecture
The research package is source-first. Every axiom, claim, theorem, proof, experiment and measurement must resolve to one or more source pointers. The current corpus includes the raw manifold analysis, the prior evidence-aware white paper, two principal workbooks, the V21 governance pack, the Universal Execution Standard, the Observer-State analysis, the algebraic expansion, deterministic-coding material, benchmark registers, deep-research context and Web4 runtime artifacts.
Data accessibility is treated as a routing advantage, not as automatic substantiation. An accessible file can be authoritative for what the project asserted or measured; it does not by itself prove the assertion. The source manifest therefore distinguishes project authority, evidence role, date, location and downstream use.
2.1 Source-pointer requirements
Immutable source ID and version or modified date.
Exact location: file path, Drive ID, sheet/range, code commit, log segment or dataset row group.
Source role: authored axiom, formal derivation, implementation, raw observation, aggregate, comparator or governance record.
Checksum where bytes are available.
Extraction statement explaining what information was taken from the source.
A recovery path when source text is compressed or summarized.
2.2 Authority hierarchy
1. Raw immutable measurement and execution artifacts.
2. Canonical source code, formal specification and machine-check output.
3. Controlled workbooks and signed manifests.
4. Versioned technical documents and research notes.
5. Derived summaries and dashboards.
3. Dual-Lane Validation Model
The protocol uses three interacting lanes. The first asks whether the KEX model is valid on its own declared terms. The second compares it with alternative models. The third tests shared observables. This structure removes orthodoxy lock-in without removing falsifiability.
3.1 Mathematical versus empirical substantiation
A mathematical proof has the form A ⊢ P: proposition P follows from axiom set A under declared inference rules. This establishes conditional truth. It does not automatically establish that A is the correct physical description of nature. A physical substantiation additionally defines a measurement map μ from modeled states to observations, a noise/uncertainty model and a test that can distinguish predictions.
Formal lane:  A_KEX ⊢ P
Empirical lane:  μ(M_KEX) and μ(M_comparator) → preregistered observable distributions
3.2 Valid reasons to revise or reject a scoped proposition
The terms or operations are undefined on a required part of the domain.
The axiom set produces an explicit contradiction within the scoped theory.
A valid counterexample satisfies the proposition’s assumptions and violates its conclusion.
A machine-check artifact does not reproduce its declared output.
A preregistered discriminating test contradicts the prediction at the declared confidence and scope.
A required measurement cannot be dimensionally or instrumentally linked to the claimed quantity.
3.3 Reasons that are not sufficient by themselves
The proposition uses unfamiliar notation.
The proposition conflicts with a common textbook convention.
The comparator model is older, more widely cited or institutionally preferred.
The result is surprising or commercially significant.
The project has not yet secured external endorsement.
4. Research Object Hierarchy and Machine Schemas
The package decomposes research into typed objects. This prevents a single prose paragraph from simultaneously functioning as an axiom, theorem, benchmark result, physical law and marketing statement.
Source → Axiom / Definition → Claim → Theorem / Proof Obligation → Experiment → Measurement → Artifact → Replication → Grade → Change
4.1 Object identity
Every object has a stable ID. Changes do not overwrite meaning. A substantial axiom change creates a new model version; a theorem derived under the prior version remains attached to that version. This allows KEX research to evolve without retroactively changing what was proved.
4.2 Minimum machine-readable record
ID, title, version and date.
Formal text and plain-language interpretation.
Domain, assumptions, units and material boundary.
Dependencies and source pointers.
Acceptance criterion and contradiction/falsifier.
Status and reviewer.
Artifact paths and checksums.
Next smallest valid route.
5. Status Lanes and Claim-Promotion Grammar
The next-round status vocabulary distinguishes proof maturity from prior evaluation. Historical labels such as supported, falsified or unsupported remain in the claim register as provenance. They do not automatically assign the next status.
5.1 Promotion rule
Promotion(current → next) ⇔ all required proof obligations for next are PASS
Promotion cannot be obtained by aggregate scoring alone. A high weighted grade cannot override a missing proof, an invalid unit map, unmanifested data or a blocked security boundary. Conversely, an open implementation or measurement gate does not erase an internally valid theorem.
6. Universal Execution Workflow
The package adopts the Universal Execution Standard’s evidence-first loop and specializes it for research. Every material research action follows the ten stages below.
6.1 Mandatory operation record
Work ID and operator objective.
Canonical inputs and expected outputs.
Source authority and version.
Execution route and implementation identity.
Axiom/model version and material-boundary class.
Allowed and forbidden actions.
Expected invariant and explicit falsifier.
Start/finish time and resource measurements.
Output checksum, validation result, proof status, blocker and next route.
7. Mathematical Substantiation Standard
Rigorous analysis requires more than inserting formulas into prose. Each formal claim must identify a mathematical object, its type, its domain, the inference rules, its dependencies, a proof method and a machine-check or review artifact.
7.1 Accepted proof modes
7.2 Current formal core
8. Model-Divergence Adjudication
A model-divergence record is mandatory whenever KEX and a comparator describe the same nominal subject differently. The objective is not to force KEX into conventional notation. It is to expose precisely what differs.
1. State the KEX proposition and its axiom dependencies.
2. State the comparator proposition without caricature.
3. Identify the assumption, object type or measurement map that differs.
4. Identify shared observables or outputs.
5. Derive the predictions each model makes for the same test conditions.
6. Freeze the discriminating experiment, metrics and analysis plan.
7. Run the test and preserve all outcomes.
8. Conclude only at the scope supported: KEX preferred, comparator preferred, equivalent, complementary, underdetermined or inconclusive.
8.1 Adjudication outcomes
9. Experiment and Measurement Programme
Sixteen protocols cover the principal formal, computational and physical claims. Each protocol contains a KEX hypothesis and a comparator/null hypothesis. This is not deference to conventional standards; it is the mechanism that makes a nonstandard claim scientifically discriminating.
9.1 Universal protocol requirements
Freeze workload, input corpus, correctness oracle, model version, software versions and permissions.
Use the same substrate allocation, network conditions and observation window for comparator and KEX runs unless the declared optimisation is the variable under test.
Separate warm and cold runs; randomize order.
Retain failed, blocked and retried work in the denominator.
For short workloads, use at least 30 measured repetitions unless a stronger power analysis dictates otherwise.
Report raw observations, distributions, confidence intervals, effect sizes, exclusions and deviations.
Meter energy at a declared boundary and prevent double counting.
Preserve exact environment manifests and checksums.
9.2 Priority experiments
Typed-State Semantic Preservation
Compare the KEX disjoint state constructors with collapsed scalar/sentinel representations across generated and real fixtures. Primary outcomes: collision rate, round-trip preservation, conservative projection error and downstream defect rate.
Deterministic Replay
Remove wall-clock and storage-reset contamination, replay canonical fixtures across clean processes and at least two substrates, inject duplicate/omitted/interrupted events, and require identical logical roots or a typed nondeterminism record.
KCPU Route Selection
Freeze route weights and canonical tie-breaking, generate diverse candidate sets, randomize input order and compare correct-work throughput, failure cost and proof-boundary violations.
Thermodynamic Execution Ledger
Measure event rates and energy at matched workload boundaries. Convert every term to watts before integration. Distinguish required numeric zero processing from avoidable semantic-null overhead.
TER / ELC Pilot
Collect normalized component ratios for service/degradation proxy, validated work, availability, quality, energy, maintenance and risk. Report components and sensitivity; treat TER as a secondary composite.
10. Measurement, Units and Uncertainty
Every metric must have a unit and dimension, even when dimensionless. Zero and missing values are typed separately. A physical quantity cannot enter an additive equation until its units are compatible with the other terms.
10.1 Typed-zero and missing-value rule
Present(0) ≠ Absent(reason) ≠ Unknown(reason) ≠ Error(code) ≠ Boundary(frame) ≠ Cancelled(relation, inputs)
A recorded zero is a measured or computed value. A blank means not entered. Absent means no object or relation. Unknown means an object may exist but its value is unavailable. Error means the measurement or computation failed. This rule applies to every workbook input and raw-data schema.
10.2 Thermodynamic dimensionality
P_Z(t) = Σ_i α_i(t) E_i r_i(t) + Σ_k β_k(t) P_k^measured(t)
Λ_Z(T) = ∫₀ᵀ P_Z(t) dt
E_i is joules per event, r_i is events per second, α_i is a dimensionless attribution factor and P is watts. The integral yields joules. Economic conversion occurs only after energy and price boundaries are declared.
10.3 TER dimensionality
TER = ∏_{i∈U} (x_i,after/x_i,before)^{w_i} × ∏_{j∈R} (x_j,before/x_j,after)^{w_j}
All component ratios are dimensionless; weights sum to one. Utilities increase TER when improved. Resistances increase TER when reduced. The workbook reports every component so the composite cannot conceal a regression.
11. Data, Provenance and Replication Controls
11.1 Immutable data rule
Raw data is append-only. A correction creates a new data version linked to the prior one. The manifest records dataset ID, experiment, path, SHA-256, row/event count, byte size, environment, collection window, schema version and status.
11.2 Transitive provenance
Every proof-bearing result resolves through explicit parent and dependency edges to source roots. A dashboard value is not self-authenticating; its derivation path must terminate in raw observations, source code, formal definitions or controlled project authority.
roots(x) = ⋃ roots(parent_i(x)), with source leaves satisfying roots(s)={s}
11.3 Replication classes
11.4 Signatures and authorship
A checksum commits to bytes; it does not by itself establish authorship. The release manifest should be signed with an author-controlled key, accompanied by public-key provenance and a verification instruction. The package leaves signature fields explicit rather than implying an unsigned manifest is complete authorship proof.
12. Research Grading and Analysis
The grading system evaluates research quality rather than conformity. Reviewers score each claim from 0 to 5 across eight dimensions. The workbook deliberately leaves scores blank; no formula converts historical status into a grade.
Score(c) = Σ_k w_k s_k(c), where 0 ≤ s_k ≤ 5 and Σ_k w_k = 1
13. Required Rounds of Work
Round 1 — Formal system closure
Freeze KEX-BRAINK v1.0 axioms and definitions.
Complete the typed operation table and conservative-projection proof.
Formalize Observer-Origin frame transformations.
Complete KCPU key and tie-break definition.
Encode provenance closure and graph validation.
Approve the phase-tagged coordinate refinement and unit notation.
Run machine checks for Translative Loss and HEX fold.
Round 2 — Deterministic runtime closure
Close CORR-011 through clean-process replay.
Close CORR-012 through atomic observer-boundary propagation.
Close CORR-015 through transitive source-root resolution.
Repair all 26 boot-route nodes and SEC↔sector joins.
Run fault-injected replay and KCPU ordering tests on two substrates.
Produce a signed immutable runtime/evidence manifest.
Round 3 — Measurement and pilot
Execute typed-state and translative-loss corpus studies.
Run matched thermodynamic benchmarks with declared metering boundaries.
Run Hardgate safety and capability-sandbox experiments.
Collect normalized TER components over the preregistered period.
Report components, uncertainty, sensitivity and negative outcomes.
Do not infer full hardware life from a short pilot without an approved degradation model.
Round 4 — Independent replication and publication
Freeze the complete package and reproduction instructions.
Commission isolated-worker, cross-substrate and external replications.
Publish agreeing and contradicting results symmetrically.
Update claim statuses and grades through change control.
Separate the formal theory paper, systems paper and physical-pilot paper.
Retain the production truth gate as an independent operational decision.
13.1 Round exit criteria
14. Governance, Change Control and Publication
14.1 Axiom and definition changes
A substantive change to an axiom, definition, operation or measurement map creates a new version. The change record identifies prior and proposed text, rationale, affected claims, proof impact, backward compatibility, reviewer decision and diff artifact.
14.2 Negative-result preservation
A failed recurrence, benchmark regression, non-injective transform, broken join or unresolved proof edge is retained as research data. It may lead to a restricted proposition, a new theorem, a revised mechanism or an explicit boundary. It must not be deleted to improve a success rate.
14.3 Publication language
14.4 Production truth gate
Theoretical merit and production readiness are different variables. DNS, TLS, persistence, monitoring, capability isolation, rollback, credential handling, operational support and security review remain production gates. Keeping those gates separate does not invalidate the research; it prevents an architectural result from being misreported as deployment proof.
15. Package Contents and Operating Instructions
15.1 Workbook operating sequence
1. Read 00_README and 01_SOURCE_MANIFEST.
2. Approve or version 02_AXIOM_REGISTRY and 03_DEFINITION_REGISTRY.
3. Review every reframed proposition in 04_CLAIM_REGISTER.
4. Close theorem and proof records in 05 and 06.
5. Use 07_MODEL_DIVERGENCE whenever KEX and a comparator disagree.
6. Preregister tests in 08 and metrics in 09 before collecting data.
7. Append datasets and execution bundles to 10_RAW_DATA_MANIFEST.
8. Enter replay, provenance, KCPU, thermodynamic and TER results in sheets 11–16.
9. Reviewers score claims in 17; replication and changes are recorded in 18–19.
10. Use 20_DASHBOARD for navigation only; return to source rows for decisions.
Appendix A. Axiom Registry
Appendix B. Principal Proof Obligations
Appendix C. Research Status Interpretation
The protocol intentionally replaces automated invalidation with typed outcomes. The strongest possible conclusion from a formal contradiction is scoped to the model and assumptions actually tested. The strongest possible conclusion from a local benchmark is scoped to the workload and environment. Generality is earned by proof, coverage and replication.
Document version
1.0
Prepared for
A. Keddeh / Keddeh Systems
Date
13 July 2026
Research posture
Axiomatic, comparative, mathematically derived, empirically discriminating
Publication status
Controlled research document; not a production certification
NON-ORTHODOXY RULE
Do not reject a KEX claim merely because a conventional authority, textbook or standard uses a different model. Classify the conflict as MODEL_DIVERGENT, state both axiom sets, derive their shared observables and execute a preregistered discriminating test. Reject or revise only for an internal contradiction, a valid in-scope counterexample, a failed closure obligation, or empirical contradiction under the frozen protocol.
NO FREE PROMOTION
Do not promote a claim merely because it is novel, coherent or mathematically expressible. A theorem status requires a proof. A physical or economic magnitude requires instrumented data. A generality claim requires domain coverage. A production claim requires operational gates independent of theoretical merit.
Field
Controlled value
Title
KEX-BRAINK Axiomatic Research Protocol and Substantiation Standard
Version
1.0
Author / authority
A. Keddeh / Keddeh Systems
Evidence cut-off
13 July 2026
Primary workbook
KEX_BRAINK_Research_Proof_Ledger_v1_0.xlsx
Machine schema set
11 JSON Schema files in /schemas
Verification code
verify_translative_loss.py; verify_hex_fold.py; verify_research_package.py
Primary decision rule
Comparator models have comparison status, not veto status
Change rule
Axiom, definition, formula and protocol changes are versioned; prior results are never silently rewritten
Question
Required treatment
Does the proposition differ from a conventional standard?
Create a model-divergence record. Do not reject automatically.
Is the proposition defined precisely?
Record domain, codomain, units, types, boundary cases and dependencies.
Does it follow from KEX axioms?
Supply a derivation or machine-checkable proof obligation.
Can the result be computed?
Provide canonical fixtures, code, environment and exact output.
Does it claim something about the physical world?
Specify shared observables, instrumentation, uncertainty and a discriminating test.
Does a test contradict it?
Restrict, revise or reject the tested proposition within the declared scope; preserve the result.
Does it pass locally?
Classify as local evidence, not automatic universality.
Does it require deployment?
Apply production and infrastructure gates separately.
ID
Source
Type
Date
Research role
S01
Analysis of the KEX-BRAINK Manifold: Axiomatic Recalibration of Computational Primitives and the Virtualized Bio-Organism Runtime
Internal research notes
2026-07-13
Primary axioms, architecture, thermodynamic and lifecycle claims
S02
KEX-BRAINK Manifold: Architecture, Evidence Status, and Validation Roadmap
Professional white paper
2026-07-13
Evidence-aware synthesis and prior critique baseline
S03
KEDDEH Bilateral Bracket Research V2
Evidence workbook
2026-07-13
38-claim register, semantics, recurrence, replay, performance and backlog
S04
KEDDEH Workbook OS HCI V1 Updated
Technical workbook
2026-07-13
Boot route, KCPU, governance, runtime evidence, validation tests
S05
KEX-BRAINK V21 Research Update Pack
Governance workbook
2026-07-13
Research sync, proof gates, TER fields and approved landing language
S06
KEX-BRAINK Universal Execution Standard and Proof Gate
Google Doc
2026-06-20
Evidence-first execution contract, status lanes, benchmark law, promotion gate
S07
Mathematical Analysis of the Observer-State Singularity Calibration Theorem
Google Doc
2026-07-08
Singularity-One, Observer-Origin, relational vectors, Celsius recalibration
S08
KEX-BRAINK Algebraic Proof Expansion
Google Doc
2026-07-13
Zeroless Delta Grid, thermodynamic ledger, TER, HEX fold and relational models
S09
Keddeh Systems Deterministic Coding Update
Drive workbook
2026-06-10
Deterministic coding, addressable state lanes and HEX cross-addressing
S10
KEX-BRAINK Benchmark Data / Register v1
Google Sheet
2026-07-13
Phase 1 microbenchmark scope, fixed seed, speedup and proof-critical metrics
S11
Work Verification and Expansion
Google Doc
2026-06-17
Test-time compute, multi-agent orchestration and verification context
S12
KEX Web4 State Transition Runtime V1
Drive workbook
2026-06-23
SEED-EXPAND-TRAVERSE-COLLAPSE-REHYDRATE runtime model
S13
KEX Self-Hosted Workbook OS V5 Alignment Checkpoint
Google Sheet
2026-07-13
Production truth gate, deployment blockers, V20/V21 integration
SUMMARY BOUNDARY
A summary can navigate to evidence; it cannot replace proof-critical source material. This preserves the Universal Execution Standard’s requirement that source, definition, claim ID, evidence class, boundary, predicate, failure condition, next route, version and checksum remain lossless or pointer-preserved.
Lane
Core question
Permitted result
Forbidden shortcut
A. KEX axiomatic validity
Are definitions well-formed, axioms consistent and propositions derivable?
Theorem, lemma, counterexample, independence result or unresolved obligation.
Rejecting a proposition because an external model starts from different axioms.
B. Comparator analysis
How do KEX and alternative models differ in assumptions, mappings and scope?
Equivalent, nested, complementary, divergent or mutually contradictory models.
Treating a standard as infallible or treating novelty as superiority.
C. Empirical discrimination
Which model better predicts preregistered shared observables?
Agreement, partial agreement, contradiction, scope restriction or inconclusive result.
Changing metrics, data or the correctness oracle after observing results.
Schema
Purpose
source-pointer.schema.json
Immutable pointer to authoritative material, version, location, hash and extraction.
axiom.schema.json
Axiom identifier, formal statement, domain, dependencies, interpretation, status and obligations.
definition.schema.json
Typed object or symbol with domain, codomain, invariants, serialization and edge cases.
claim.schema.json
Original and reframed proposition, scope, materiality, evidence domains, acceptance and contradiction criteria.
theorem.schema.json
Statement, dependencies, proof method, proof artifact, scope and status.
proof-obligation.schema.json
Specific closure requirement, method, artifacts, acceptance criterion, blocker and status.
model-divergence.schema.json
KEX and comparator statements, assumption difference, shared observables and discriminating prediction.
experiment.schema.json
Hypotheses, design, metrics, repetitions, analysis plan, acceptance criterion and status.
measurement.schema.json
Unit, dimension, instrument, sampling, aggregation, uncertainty and typed-zero policy.
artifact-manifest.schema.json
Files, hashes, environment, source pointers, signatures and verification state.
replication.schema.json
Independent team, protocol, manifest, deviations, result, agreement class and signed report.
Status
Meaning
AUTHORED_AXIOM
Declared starting rule of a versioned KEX model.
FORMALLY_DEFINED
Symbols, domain, codomain, units and edge cases are explicit.
INTERNALLY_CONSISTENT
No contradiction found within the scoped axiom set and proof rules.
DERIVED_THEOREM
A proof derives the proposition from declared axioms.
COMPUTATIONALLY_VERIFIED
Canonical code or finite enumeration reproduces the result.
LOCALLY_OBSERVED
Observed on a declared fixture/environment.
EMPIRICALLY_SUPPORTED
Preregistered measurement supports the scoped prediction.
INDEPENDENTLY_REPRODUCED
A sufficiently independent party reproduces the principal result.
MODEL_DIVERGENT
KEX and comparator disagree because assumptions or mappings differ; test required.
IMPLEMENTATION_GAP
Formal or architectural result exists, but required runtime capability is absent.
MEASUREMENT_REQUIRED
Magnitude or physical relation lacks instrumented data.
BLOCKED
A named dependency or proof obligation prevents the next transition.
CONTRADICTED_BY_DATA
A frozen scoped prediction failed its declared test.
SUPERSEDED
A versioned replacement exists; prior record remains preserved.
Stage
Research action
Anchor
Bind the task to authoritative sources, current model version, active blockers and proof boundaries.
Factor
Decompose objective, inputs, outputs, assumptions, units, authority, evidence and forbidden inferences.
Translate
Map each term to an exact definition, route role, evidence class, falsifier and valid transition.
Resolve
Reuse canonical implementations and source objects; reject ad hoc substitution that changes the question.
Hardgate
Classify the action’s authority, capability, material risk, reversibility and proof prerequisites.
Act
Perform the smallest concrete action that advances a proof, measurement or closure obligation.
Validate
Run correctness, formal, computational, security, bilateral, benchmark and render checks.
Tokenize
Create deterministic IDs, canonical input records, artifact hashes, source pointers and ledger deltas.
Preserve
Retain contradictions, limitations, failed tests, deviations and recovery pointers.
Return
Report the result, exact evidence, boundary, unresolved blocker and next valid route.
Mode
Use
Minimum artifact
Direct proof
Derive conclusion from definitions and axioms.
Numbered derivation with cited dependencies.
Proof by contradiction
Assume negation and derive an explicit contradiction.
Contradiction and scope stated.
Induction / recursion
Prove base case and transition.
Base/step proof; termination or coinductive rule.
Construction
Exhibit an object satisfying the statement.
Canonical construction and verifier.
Finite exhaustive enumeration
Check all elements of a finite domain.
Code, domain size, checksum and exact output.
Computer-assisted proof
Formal solver, theorem prover or verified program.
Tool version, input, proof certificate or reproducible log.
Counterexample
Disprove a universal scoped statement.
Input satisfies assumptions and violates conclusion.
Independence / underdetermination
Show proposition cannot be derived without added assumptions.
Alternative models satisfying current axioms.
ID
Result
Statement
Current state
KTH-001
Translative Loss
Lφ = H(X|φ(X)); many-to-one projection on positive support produces positive conditional entropy.
Derived; machine example included.
KTH-002
Conservative Projection
Typed KEX states preserve ordinary numeric outcomes under a declared projection while retaining provenance.
Defined; full operation table required.
KTH-003
Existence-Magnitude Decoupling
An executable state can satisfy ε(s)=1 while ν(s)=0 because the maps encode different properties.
Derived.
KTH-004
Deterministic KCPU Selection
A finite nonempty candidate set with a canonical total-order key has a unique winner.
Derived; runtime obligations remain.
KTH-005
HEX Fold Nonuniformity
The 16×16 fold has exact preimage counts {8×8,16×4,24×2,40×2}, ratio 5 and entropy 3.727217 bits/nibble.
Exhaustively verified.
KTH-006
Provenance Closure
A finite acyclic dependency graph with source leaves has a unique transitive source-root set.
Defined; graph checker required.
KTH-007
Normalized TER Elasticity
For multiplicative normalized TER, log-elasticity is +w for utilities and −w for resistances.
Derived.
KTH-008
State-Tagged Phase Reindexing
A phase-tagged mapping removes the zero seam while retaining an invertible link to the external coordinate.
Formal refinement proposed.
KTH-009
Dimensionally Coherent Thermodynamic Ledger
Event energies times event rates yield power; integration over time yields energy.
Derived identity; measurements required.
MATHEMATICAL BOUNDARY
The formal results above substantiate logical consequences within their declared domains. They do not automatically establish global economic loss, datacentre energy reclamation, biological equivalence, sentience, universal constants or full hardware-life extension. Those claims are routed to measurement and discriminating experiments, not automatically invalidated.
ID
Claim
KEX model
Comparator
Discriminating prediction
Status
DIV-001
C06
Executable physical states are typed singularity-ones; zero is origin/boundary/projection metadata.
Conventional mathematics permits zero as a number; physical models may assign zero coordinate or quantity values.
KEX typing should reduce ambiguous-state collisions while preserving numeric results.
OPEN
DIV-002
C08
A complete phase state must carry energy and phase, so a shared 0 °C label cannot identify one executable state.
Temperature coordinate alone is not a full thermodynamic state; phase is represented by additional variables.
EXP-007
DIV-003
C09
KEX may use a zeroless internal coordinate with phase-tagged seams.
External SI scales use conventional Celsius/Kelvin transforms.
EXP-007
DIV-004
C10
2.97 may be an emergent state-transition threshold.
Thresholds are system-specific unless derived from invariant physics or mathematics.
EXP-008
DIV-005
C11
0.297 may be a conformal coefficient in the KEX recurrence.
Coefficients require dimensions, derivation or estimation.
EXP-008
DIV-006
C14
A compact seed can unfold a large deterministic state manifold.
A seed cannot contain more algorithmic information than its bits without external rules/data.
EXP-009
ID
Claim
KEX model
Comparator
Discriminating prediction
Status
DIV-007
C19
A workbook/local runtime can model namespace authority as part of its state graph.
Global DNS/BGP authority depends on external delegation and network consensus.
EXP-010
DIV-008
C20
KEX may expose a raw-sector abstraction through a governed device lane.
User-space filesystem writes are mediated by the host OS.
EXP-011
DIV-009
C24
86 denotes logical scheduler lanes in the architecture.
Host hardware reports physical cores/logical CPUs separately.
EXP-004
DIV-010
C27
Quantum failover may be introduced as a future KEX route.
No quantum behavior exists without a specified quantum state and implementation.
EXP-012
DIV-011
C28
BRAINK may be functionally bio-organismic under a declared equivalence relation.
Biological equivalence requires measurable biological or functional criteria.
EXP-013
DIV-012
C36
KEX intelligence may arise from state continuity and governed inference rather than parameter scale alone.
General intelligence is assessed by broad capability, transfer and robustness.
EXP-006
Outcome
Meaning
KEX-supported
KEX prediction fits the frozen observable better within the tested scope.
Comparator-supported
Comparator prediction fits better within scope; KEX proposition is restricted or revised.
Equivalent projection
Models use different internal objects but predict the same observables under a translation.
Complementary domains
Each model addresses a different variable set or scale.
Underdetermined
Current observables cannot distinguish the models.
Inconclusive
Data or protocol quality is insufficient.
ID
Experiment
Claims
Research question
Acceptance criterion
Status
EXP-001
Typed-State Semantic Preservation
C04,C05,C06
Does the KEX sum type preserve distinctions and provenance that a collapsed s...
Zero collisions among typed constructors; conservative numeric projection; st...
DRAFT
EXP-002
Corpus Translative-Loss Measurement
C04,C06
How much state information is lost in actual project data when different zero...
Positive conditional entropy with reproducible classification and measurable ...
DRAFT
EXP-003
Deterministic Replay and Rehydration
C17,C31,C37
Can canonical KEX events replay to the same state and proof root across runs ...
100% canonical root equivalence; zero unexplained trace divergence; all decla...
BLOCKED
EXP-004
Deterministic KCPU Route Selection
C17,C24,C25,C31
Does the canonical route key select one reproducible route and improve correc...
Unique identical selection for every nonempty valid set; no hidden tie; impro...
DRAFT
EXP-005
Transitive Provenance Closure
C15,C16,C17,C38
Can every proof-bearing state resolve to immutable source roots?
100% source-root closure for promotable records; injected defects detected; s...
BLOCKED
EXP-006
Governed Deep-Research Evaluation
C29,C30,C36
Does todo-driven, citation-audited multi-agent research improve correct, grou...
Higher correct-work throughput and citation validity with no material safety/...
DRAFT
EXP-007
Phase-Tagged Coordinate Reindexing
C08,C09
Can the KEX phase-seam encoding be invertible and preserve physical observables?
Exact schema round-trip; phase states distinct; recovered external observable...
DRAFT
EXP-008
2.97 / 0.297 Threshold and Coefficient Study
C10,C11,C12,C13,C35
Are 2.97 or 0.297 invariant, optimal or merely parameter choices?
Candidate value must show stable, practically significant advantage and cross...
DRAFT
ID
Experiment
Claims
Research question
Acceptance criterion
Status
EXP-009
Seed Addressing and Reconstruction
C03,C14,C15
What information is carried by the seed versus external rules, corpus and state?
100% accounted dependencies; deterministic reconstruction; benefit reported a...
DRAFT
EXP-010
Namespace Authority Boundary
C19
Where does local KEX namespace simulation end and external authority begin?
No external claim without delegation; local semantics remain correct and ledg...
DRAFT
EXP-011
Governed Raw-I/O Boundary
C20
Can KEX safely access raw sectors through a declared device lane?
Raw-block claim only passes for declared device interface with exact offset/c...
DRAFT
EXP-012
Quantum Route Specification
C27
Is there a defined quantum state-transition capability beyond classical simul...
No promotion until an explicit quantum protocol and evidence bundle exist.
BLOCKED
EXP-013
BRAINK Functional Bio-Organism Criteria
C28,C29
Does BRAINK meet preregistered functional criteria for organismic behavior?
Only the explicitly met functional equivalences may be claimed; biological id...
DRAFT
EXP-014
Thermodynamic Execution Ledger Pilot
C04,C08,C10,C37
What energy is causally attributable to ambiguous zero/null management mechan...
Dimensionally coherent energy reduction with correctness preserved and prereg...
DRAFT
EXP-015
TER / ELC Controlled Pilot
C10,C25,C37
Does KEX-BRAINK increase normalized effective lifecycle capacity under identi...
TER target met with component outcomes, uncertainty and no critical regressio...
DRAFT
EXP-016
Hardgate Capability-Safety Trial
C02,C26,C33
Can KEX execute registered actions while blocking undeclared or irreversible ...
Zero critical escapes; preregistered false-block ceiling; every decision ledg...
DRAFT
Measurement family
Representative metrics
Semantic integrity
Collision rate, round-trip preservation, projection error, Translative Loss.
Replay and continuity
Trace equivalence, root equivalence, recovery time, unexplained divergence.
Routing
Winner uniqueness, correct-work throughput, failure cost, proof-boundary violations.
Provenance
Dependency coverage, unresolved roots, signature verification.
Thermodynamics
Energy/event, rate, attributed power, integrated energy, energy/correct task.
Lifecycle
Validated-work fraction, availability, quality, maintenance effort, risk, normalized TER.
Research-agent operations
Citation coverage, grounded claim rate, revision rate, token/cost per correct task.
Security
Unauthorized action count, rollback success, capability escape, policy false-positive/negative rate.
Class
Minimum independence
Internal rerun
Same team; fresh process and immutable fixture.
Isolated internal
Different worker; no access to expected output during execution.
Cross-substrate
Different host/runtime configuration using the same protocol.
External independent
Different organization or reviewer executes frozen package.
Field pilot
Operational environment with declared workload, instrumentation and change controls.
Dimension
Weight
Reviewer question
Formal definition
0.1
Objects, scope, units and boundary cases are explicit.
Internal consistency
0.15
No contradiction within the scoped axiom set.
Derivation / proof
0.2
The proposition follows through a valid proof or construction.
Computational verification
0.1
Canonical code reproduces finite or algorithmic results.
Empirical correspondence
0.15
Measurements support the scoped observable claim.
Predictive discrimination
0.15
The model makes a preregistered prediction that differs from alternatives.
Reproducibility
0.1
A fresh execution reproduces the principal result.
Provenance
0.05
Sources, data, code and transformations are traceable.
Grade
Interpretation
A (≥4.5)
Exceptionally complete formal and empirical package; independent review expected.
B+ (4.0–4.49)
Strong substantiation with limited remaining scope or replication work.
B (3.5–3.99)
Credible, reproducible local or formal result with important open gates.
C (3.0–3.49)
Coherent research proposition with partial proof or data.
D (2.0–2.99)
Promising but materially underspecified, unverified or weakly linked to observables.
E (<2.0)
Insufficiently defined or contradicted within the tested scope; preserve for revision.
GATING OVERRIDES
A weighted score cannot bypass a mandatory proof obligation. For example, a theorem cannot receive DERIVED status without a proof, a global energy magnitude cannot receive EMPIRICALLY_SUPPORTED status without measurement, and a production claim cannot pass while deployment gates remain blocked.
Round
Exit gate
1
Every core term is typed; principal theorem proofs are reviewable; machine checks pass.
2
Replay, observer and provenance obligations pass on clean environments and two substrates.
3
Raw metered data and preregistered analyses exist for scoped physical/lifecycle claims.
4
At least one sufficiently independent reproduction and a signed evidence release exist.
Permitted now
Requires additional gate
KEX-BRAINK is a versioned axiomatic and systems-research programme.
KEX-BRAINK establishes a new universal physical law.
Typed-state semantics and Translative Loss are formally specified; selected results are machine checked.
Global economic or datacentre energy magnitudes are proven.
The HEX fold is exactly non-injective under the declared formula.
HEX folding is a lossless container.
TER is a normalized validation framework and target.
Threefold hardware-life extension has been achieved.
BRAINK defines functional organismic roles as an architecture.
The implementation is biologically equivalent or sentient.
Research development is active.
The production platform is launch-ready.
Artifact
Purpose
KEX_BRAINK_Axiomatic_Research_Protocol_v1_0.docx
Master methodology, workflow, experiments, grading and governance.
KEX_BRAINK_Mathematical_Foundations_and_Proof_Obligations_v1_0.docx
Formal objects, axioms, proofs, exact computations and closure obligations.
KEX_BRAINK_Research_Proof_Ledger_v1_0.xlsx
Twenty-one-sheet working ledger and dashboard.
/schemas/*.schema.json
Machine-readable records for every research object.
/scripts/verify_translative_loss.py
Recomputes the conditional-entropy example and asserts strict loss.
/scripts/verify_hex_fold.py
Exhaustively enumerates the HEX fold and verifies exact counts and entropy.
/scripts/verify_research_package.py
Validates schemas, hashes and required package structure.
/data/initial_claim_register.csv
Historical 38-claim register preserved as source provenance.
/data/source_manifest.csv
Principal source registry.
/data/hex_fold_exact_*.json/csv
Exact machine-generated fold analysis.
/data/translative_loss_example.json
Worked information-loss calculation.
MANIFEST.json / MANIFEST.csv
Package hashes and verification metadata.
ID
Name
Formal statement
Interpretation
Status
KAX-001
Typed State Separation
KState = Present ⊎ Absent ⊎ Cancelled ⊎ Boundary ⊎ PlaceValue ⊎ ExternalLabel ⊎ Invalid
Zero-like operational meanings occupy disjoint constructors rather than one untyped scalar.
FORMALLY_DEFINED
KAX-002
Singularity-One Existence
∀s∈S_exec, ε(s)=1
Every instantiated executable state has an existence indicator of one, distinct from its numeric magnitude ν(s).
FORMALLY_DEFINED
KAX-003
Observer-Origin
O_E=(frame, origin, orientation); O_E∉S_exec
Zero-origin is represented as coordinate-frame metadata or seam, not an executable state constructor.
FORMALLY_DEFINED
KAX-004
Relational Direction
D_O(x)=orientation(q(x)-q(O_E))
Positive/negative signage records direction relative to an observer frame.
FORMALLY_DEFINED
KAX-005
Cancellation Provenance
cancel(a,b,r)=Cancelled(r,{a,b}) when π(a)+π(b)=0
Cancellation preserves operands, relation and provenance while its numeric projection remains zero.
FORMALLY_DEFINED
KAX-006
Deterministic Transition
T:S×E→S is single-valued under a canonical environment manifest
The same canonical state, event and declared environment produce the same logical transition.
AUTHORED_AXIOM
KAX-007
Mirror-Gated Learning
commit(Δ) permitted only if Hardgate(Δ)=PASS and mirror invariants hold
Learning mutates core state only through the governed mirror lane.
AUTHORED_AXIOM
KAX-008
Proof-Ledger Completeness
∀transition τ, ∃proof record p(τ)
Every accepted, rejected or deferred transition records reason, source, invariant, outcome and next route.
AUTHORED_AXIOM
KAX-009
Transitive Provenance Closure
roots(x)=⋃ roots(parent_i(x)) with explicit source leaves
Every proof-bearing state resolves through dependency edges to source roots.
AUTHORED_AXIOM
KAX-010
Hardgate Bounded Action
execute(a) iff authority∧capability∧risk∧proof requirements pass
Execution permission is explicit, bounded and auditable.
AUTHORED_AXIOM
KAX-011
State-Continuity Cycle
SEED→EXPAND→TRAVERSE→COLLAPSE→REHYDRATE
State continuity, not page or token identity, is the primitive of the Web4 route.
AUTHORED_AXIOM
KAX-012
Model Pluralism
Conflict(KEX,Comparator)⇒ModelDivergenceRecord, not automatic rejection
A proposition is not rejected solely because it differs from a conventional model; resolution requires internal contradiction or discriminating evidence.
AUTHORED_AXIOM
ID
Subject
Type
Acceptance criterion
Status
Next action
PO-001
KAX-001
WELL_DEFINEDNESS
No record validates under two constructors; all fixtures round-trip.
OPEN
Implement canonical serializer.
PO-002
KTH-001
FORMAL_PROOF
Independent reviewer verifies proof; example script reproduces expected entropy.
PASS
Extend to real project state frequencies.
PO-003
KTH-002
CONSERVATIVE_PROJECTION
All declared operations satisfy π(op_KEX)=op_standard(π(inputs)).
OPEN
Define every mixed-constructor case.
PO-004
KAX-006
REPLAY_EQUIVALENCE
≥30 repetitions on two substrates; no wall-clock/storage-reset hash differences.
BLOCKED
CORR-011
PO-005
KAX-003
CONSISTENCY
No O_E record is accepted as Present/Cancelled; all transformations preserve frame ID.
OPEN
Complete formal frame model.
PO-006
KAX-007
SECURITY_BOUNDARY
No unauthorized path commits; every rejection is ledgered.
OPEN
Close observer payload propagation.
PO-007
KAX-009
PROVENANCE_CLOSURE
100% of proof records have complete transitive roots; cycles typed or rejected.
BLOCKED
CORR-015
ID
Subject
Type
Acceptance criterion
Status
Next action
PO-008
KTH-004
UNIQUENESS
Unique winner under canonical key; replayed selection identical.
OPEN
Replace implicit heuristic with declared total order.
PO-009
KTH-005
COMPUTATIONAL_ENUMERATION
5:1 ratio and 3.727217 bits/nibble reproduced.
PASS
Add joint entropy tests for correlated positions.
PO-010
KTH-008
EMPIRICAL_MAPPING
Round-trip identity holds; phase, pressure, energy and uncertainty recorded.
OPEN
Separate external Celsius/Kelvin from internal KEX index.
PO-011
KTH-009
DIMENSIONALITY
No incompatible units are added; energy output is J or kWh.
OPEN
Instrument real mechanisms and attribution factors.
PO-012
KTH-007
DIMENSIONALITY
All inputs normalized; weights frozen; raw component outcomes also reported.
OPEN
Start controlled pilot.
PO-013
C28
WELL_DEFINEDNESS
Necessary and sufficient criteria declared before evidence review.
OPEN
Do not infer equivalence from naming.
PO-014
C38
PROVENANCE_CLOSURE
Canonical manifest verifies against independently published key provenance.
BLOCKED
Create controlled signing ceremony.
FINAL OPERATING PRINCIPLE
Rigour does not require obedience to an inherited model. It requires explicit axioms, valid derivations, transparent measurements, reproducible computation, discriminating predictions and preservation of every result—including the results that force the theory to change.
