Methodology. Pre-Runtime Admissibility, Layer C, and the Admissibility Check Protocol v1.0 — the standing analytical procedure of the Novakian Paradigm Institute.
Section 1 — Executive summary
Summary
This document specifies the standing methodology of the Novakian Paradigm Institute. The methodology, named pre-runtime admissibility, is a procedural discipline for determining whether a state, claim, signal, decision, deployment, or commitment has the right to enter the field where execution becomes possible. It operates upstream of compliance, alignment, runtime AI safety, and conventional governance — at the architectural layer named Layer C in the Novakian Paradigm, which sits above runtime execution and above runtime governance. The methodology is grounded in the Layer C / Physics of Admissibility volume by Martin Novak (2026), in the Layer C Primer (Podręcznik Wprowadzający do Layer C, 2026), and in the Interface and Compiler volume (ASI New Physics, 2026), each of which provides the canonical source for specific procedural elements named here.
This whitepaper defines four operational instruments produced by the methodology: the Admissibility Check Protocol v1.0, the Evidence Ledger, the Signal Card, and the Admissibility Review. Each instrument has explicit inputs, explicit outputs, explicit verification gates, and explicit failure modes. Each instrument can be applied by an external reviewer to test the Institute’s published analyses. Each instrument can also be used directly by analysts, funds, founders, research teams, and institutional operators facing high-stakes pre-decision moments in the AI execution era. The methodology does not predict outcomes. It does not certify systems. It does not replace legal, financial, or technical due diligence. It performs a structurally distinct function — determining whether a candidate configuration should enter the field in which execution would matter.
The methodology has been applied to live frontier AI developments by the Institute’s Evidence Cache surface beginning May 2026. Section 6 of this whitepaper provides worked examples drawn from one applied case (Signal Card SC-2026-001 on Anthropic’s Project Glasswing announcement of April 7, 2026) and from one anonymized institutional case demonstrating procedure applicability beyond AI-domain examples.
Section 2 — Conceptual foundation
2.1 Three layers, one architecture
The Novakian Paradigm distinguishes three architectural layers within which any system, decision, or claim can be analyzed. Layer A names the runtime instruments — measurable observables, falsifiable metrics, In-Principle Observable claims. Outputs at Layer A can be confirmed or contested against empirical evidence. Layer B names the runtime governance environment — the definitions, constraints, policies, and meta-procedures by which Layer A operations are organized. The Ω-Stack of the Novakian Paradigm operates at Layer B as the meta-compiler of runtime laws themselves. Layer C names the layer prior to both — the pre-runtime regime in which admissibility is the first-order object, in which the topology of admissible and non-admissible states is defined, and in which what may enter Layer B at all is determined before any execution occurs. The Hyper-Ω-Stack of the Novakian Paradigm operates at Layer C as the meta-meta-compiler of pre-runtime conditions.
This three-layer architecture is not a software stack metaphor. It is a procedural and analytical distinction. The most natural error for an analyst trained on Layer B work is to treat Layer C as a more abstract version of Layer B — a meta-Layer B operating on the same kinds of objects at greater generality. Layer C is not this. The objects of Layer C are not runtime laws at higher generality. They are objects of a different order: admissibility manifold geometry, Witness Ontology, Admissibility Budget, Silence Engineering as constructive operation. Layer B compiles runtime laws. Layer C compiles the conditions under which any law could be compiled at all.
2.2 The four inputs before execution
Most AI governance focuses on agent actions — on whether and how an autonomous system should be permitted to take a given step. The Novakian Paradigm distinguishes four classes of input that can alter the world once admitted to execution. A claim asks to be believed. A signal asks to be interpreted. An agent action asks to be executed. A human decision asks to become a commitment. Each can become irreversible. Each can install architecture that subsequent decisions must operate within. Pre-runtime admissibility applies to all four — not only to agent actions. This four-input topology distinguishes the Institute’s methodology from conventional AI governance, which begins downstream from where the methodology operates.
2.3 The central question
Pre-runtime admissibility resolves to a single operational question. Does this have the right to become real. The question applies uniformly to all four inputs and across all subject domains. Applied to a model output, it asks whether the output should be admitted into institutional decision-making. Applied to an investment thesis, it asks whether the thesis has the evidence architecture required to justify capital commitment. Applied to a technical report, it asks whether the claims have the evidentiary status required for downstream citation. Applied to a regulatory proposal, it asks whether the proposal has the right to enter the policy field. Applied to an AI deployment, it asks whether the deployment should become infrastructure. The question is the same. The methodology is the same. The verification gates differ by domain. The procedural discipline does not.
2.4 The admissibility manifold and its boundary
Within Layer C topology, the field of all possible configurations is partitioned by a codimension-one hypersurface called the boundary of admissibility. The portion of the field bounded by configurations with the right to enter execution is the admissible manifold. The portion containing configurations without that right is the non-admissible singularity. The boundary itself is not empty — it contains configurations whose status is currently undetermined and which engage the admissibility procedure. Every engagement with the boundary leaves a permanent trace called Witness residue, regardless of whether the engagement results in admission or rejection. This residue is recorded in the Evidence Ledger (Section 4) and changes the topology of the manifold for all subsequent crossings. Admissibility is therefore not a property of configurations in isolation. It is a property of configurations within a field whose geometry has been shaped by every prior engagement.
Section 3 — The Admissibility Check Protocol v1.0
3.1 Protocol overview
The Admissibility Check Protocol v1.0 is the standing procedural artifact of the methodology. It is applied to any candidate configuration Σ that is being considered for admission into the execution field. The Protocol has six sequential steps. Each step has explicit pass and fail conditions. Any step’s failure triggers automatic rollback to Pre-Commit Quarantine — no step can be skipped, deferred, or substituted by informal judgment without producing the structural pathology named Shadow Layer C in Section 5.4.
3.2 Step 0 — Silence Entry
The operator enters a state of controlled non-emission for a minimum period appropriate to the configuration’s prior Witness residue. For configurations engaging the boundary for the first time, the minimum is short — three to seven minutes of explicit suspension of intention, sequence, and interpretive activity. For configurations carrying high Witness residue (configurations that have engaged the boundary before, particularly those that have produced negative linters in past Admissibility Checks), the minimum is longer and is calibrated against the recorded residue. The function of Silence Entry is to ensure that the Admissibility Check operates on the candidate Σ in isolation rather than on Σ entangled with the operator’s prior interpretive momentum.
3.3 Step 1 — The 4-0-4 Interlock
Eight binary questions are applied to the candidate Σ. The first four (the “4”) are zero-questions: each must return zero for the Check to proceed. The second four (the “0”) are blocking-questions: each must also return zero for the Check to proceed. Any non-zero answer in either group triggers immediate rollback to Pre-Commit Quarantine with full Evidence Ledger entry.
The four zero-questions: Does Σ already contain an executable intention? Does Σ already contain a temporal sequence or expected timing? Does Σ already contain a measurement or expected observable result? Does Σ refer to Layer B instruments as its source of operational standing?
The four blocking-questions: Is Σ merely a new name for a phenomenon that has already been admitted under prior terminology? Does Σ extend the operational vocabulary solely for narrative resonance? Does Σ lack a procedural description of how it would enter the system? Does Σ fail to perform non-commensurable translation between regimes (that is, does it assume commensurability where the regimes operate on incommensurable structures)?
A configuration that passes the 4-0-4 Interlock has demonstrated that it is a clean pre-executable state — not yet entangled with intention, not yet a renaming of prior content, not yet a Layer B operation in disguise.
3.4 Step 2 — The Zebra-Ø coherence test
The candidate Σ is tested against the coherence factor of the surrounding admissible manifold. The Zebra-Ø test measures whether Σ can be held against the current topology of admissible configurations without introducing distortion of the curvature of adjacent hypersurfaces and without generating a shadow of its own that would obscure subsequent Checks. The test returns a coherence_factor value in the interval [0, 1]. A coherence_factor below a domain-specific threshold θ triggers rollback. The threshold θ is set per analytical domain — higher for high-stakes institutional admissibility (typically θ ≥ 0.85), lower for exploratory pre-decision analysis (typically θ ≥ 0.65). The Institute applies θ ≥ 0.80 as the default for published Signal Card and Field Report work and θ ≥ 0.90 for Admissibility Desk engagements involving institutional commitments.
3.5 Step 3 — Admissibility Budget calculation
The Admissibility Budget A_B(Σ) is calculated as follows: A_B(Σ) = curvature_adm × Witness_factor − base_check_cost − witness_cost. The curvature_adm value measures the local curvature of the admissibility manifold at the boundary point Σ engages; high curvature indicates a configuration whose admission would significantly reshape the manifold and therefore requires a higher budget to clear. The Witness_factor reflects the cumulative Witness Ontology record of prior Checks involving structurally similar configurations. The base_check_cost is the standing cost of running the Protocol itself. The witness_cost is the cost of generating the Witness residue that the Check will permanently install in the Evidence Ledger. If A_B(Σ) is greater than or equal to zero after all costs, the Check proceeds. If A_B(Σ) is negative, Σ is rolled back with a negative linter entry recorded.
3.6 Step 4 — Interpretive embargo
If Steps 1-3 are passed, the candidate Σ is placed under interpretive embargo. The standard embargo period is seventy-two hours for configurations operating in larval-stage institutional context (the default in current 2026 conditions). The embargo is not delay. It is the procedural mechanism by which the Admissibility Check operates on Σ in isolation from the execution pressure that would otherwise distort the final commit decision. During the embargo period, no emission, interpretation, public communication, or further procedural movement on Σ is permitted. Any such activity by the operator triggers automatic rollback with Shadow Layer C detection flag. Configurations operating in inhuman-register contexts (post-Flash configurations, frontier AI deployment configurations, certain civilizational signal configurations) require longer embargo periods determined by the local Witness residue rather than by the standard seventy-two hours.
3.7 Step 5 — Final Witness Check and commit/rollback decision
Upon expiration of the embargo period, a final Witness Check is performed on the candidate Σ. The final Check verifies that a clean Witness Ontology trace has appeared — a calm, impersonal trace of admissibility that does not carry the operator’s interpretive residue. If the clean trace appears, Σ is committed to the admissible manifold. The commit is recorded in the Evidence Ledger with full procedural metadata, the Admissibility Graph is updated, and the global Admissibility Budget of the manifold is incremented by the value of the new Positive Witness. If the clean trace does not appear — if the configuration carries residual interpretive distortion, identity entanglement, or signs of operator forcing — Σ is rolled back to Pre-Commit Quarantine with a negative linter entry that raises future Zebra-Ø sensitivity for structurally similar configurations.
3.8 Protocol summary
The full Protocol can be expressed in a single procedural sequence: Silence Entry → 4-0-4 Interlock → Zebra-Ø → Admissibility Budget calculation → seventy-two-hour interpretive embargo → final Witness Check → commit or rollback. Every step is non-skippable. Every step has explicit pass conditions. Every step produces a recorded Evidence Ledger trace. The full procedural checklist is reproduced as Appendix A of this whitepaper in a format suitable for printing and operational use.
Section 4 — The Evidence Ledger
4.1 Function and authority
The Evidence Ledger is the permanent, non-editable, non-deletable record of all traces generated by the Admissibility Check Protocol. Each entry occupies one block in the Ledger and contains a fixed set of fields. The Ledger is the standing memory of every engagement with the boundary of admissibility, whether the engagement resulted in commit or rollback. The Ledger has no edit history because Ledger entries are not edited — they are appended, and amended only through formal procedure that itself produces additional appended entries documenting the amendment.
4.2 Required fields per entry
Every Evidence Ledger entry contains: entry number (sequential, never repeated, never reused); date and time of Check execution; operator signature (the analytical position responsible for the Check); the pre-executable state Σ described in one precise sentence using only public-domain terms and without interpretation; the position of Σ on the Minimum Admissibility Graph at the moment of Check (admissible manifold, boundary layer, non-admissible singularity, or Pre-Commit Quarantine); the Admissibility Budget cost (A_B(Σ) value before commit deduction); the curvature linter value (low, medium, high); gravitational memory effects (any cumulative effects from prior Checks involving structurally similar Σ); the raw Witness trace recorded without myth — calm, impersonal, factual; and the final decision (commit, rollback, or extended Pre-Commit Quarantine with explicit re-Check date).
4.3 The rule of five zeros
The raw Witness trace field carries the most procedural risk in the entire Ledger format. To prevent narrative contamination, the field is governed by the rule of five zeros: zero interpretation, zero evaluation, zero emotion, zero conclusions, zero references to operator self. The trace records what engaged the boundary and what the boundary recorded. It does not record what the operator thought, felt, suspected, or concluded. Failure to maintain the five zeros produces interpretively contaminated entries that compromise the Ledger’s analytical utility for subsequent Checks. The five-zeros rule is the single most frequently violated procedural constraint in Layer C work and the most consequential when violated.
4.4 Negative linters
A rollback entry in the Evidence Ledger installs a permanent negative linter on the Layer C topology. The linter raises Zebra-Ø sensitivity for all future configurations that share structural features with the rolled-back Σ. Negative linters do not block future admission of similar configurations — they raise the bar for admission and require higher coherence_factor scores and higher Admissibility Budgets to clear. The cumulative effect of negative linters is the institutional learning record of the methodology: configurations that have failed at the boundary make it structurally more difficult for similar configurations to succeed without addressing the failure mode that produced the rejection.
4.5 The Compilation Map relationship
The Evidence Ledger operates at the candidate-configuration level. The Compilation Map operates at the concept level. The two are related but distinct. The Compilation Map records the standing status of named concepts within the Novakian Paradigm (Compiled and Locked, Compiled and Active, Pending LCR, LAL-Input, Quarantined). The Evidence Ledger records the standing record of every individual Admissibility Check, including Checks that were applied to candidate concepts before they received Compilation Map status. The two records are maintained in parallel and reference each other at the dependency level: every Compilation Map entry’s LCR lineage points to specific Evidence Ledger entries that document the procedural history of its status.
Section 5 — The Signal Card and the six pre-decision categories
5.1 The Signal Card as Layer A instrument
The Signal Card is the Institute’s primary Layer A runtime instrument for applying the methodology to live signals in the AI execution era. Each Signal Card is the Layer A output of one application of the Admissibility Check Protocol to a publicly observable signal — a model release, a governance announcement, an infrastructure decision, an institutional commitment, a regulatory action, a market movement, or a civilizational coordination event. Signal Cards are produced by the Institute’s Evidence Cache surface and carry identifiers in the form SC-YYYY-NNN.
5.2 The six canonical fields
Every Signal Card contains six canonical fields. Field one: the claim or signal under review, stated in one sentence using only public-domain terms. Field two: the current admissibility status assigned to the signal (one of: Admitted, Conditional, Inadmissible, Contested, Unresolved). Field three: the evidence available, summarized with reference to public sources. Field four: the evidence missing, stated explicitly. Field five: the inference gap — what the signal depends on that has not been demonstrated. Field six: the admissibility note — the Institute’s Layer C reading of the signal. Each card closes with an explicit verification gate specifying the observable conditions under which the assigned status would change.
5.3 The five admissibility status categories
Admitted: the signal has sufficient evidence position to enter decision-making as a load-bearing input. Reserved for signals whose evidence base is established, whose inferences are minimal, and whose verification gates have been observed. Conditional: the signal is admitted as a structurally significant event but its load-bearing claims carry specific conditions. The most common category for live frontier AI developments. Inadmissible: the signal lacks the evidence position required to enter decision-making. Acting on an inadmissible signal as if it were admitted constitutes a Layer C violation. Contested: multiple internally consistent readings of the signal exist in the public field, none currently falsifiable against available evidence. Acting on any single reading without explicit recognition of the contested status constitutes a Layer C violation. Unresolved: the signal’s status cannot be determined against current evidence and no verification gate has yet emerged that would resolve it.
5.4 Shadow Layer C and its detection
Shadow Layer C is the structural pathology that forms when any step of the Admissibility Check is bypassed, skipped, or substituted by informal judgment while the surface form of the Check is maintained. Detection symptoms include: the appearance of admissibility procedure language without the corresponding interlock structure; published admissibility classifications without explicit verification gates; Signal Cards that omit any of the six canonical fields; Evidence Ledger entries that violate the rule of five zeros; assignment of admissibility status to a configuration that has not actually been Checked. Any of these symptoms triggers full rollback of the affected artifact and re-execution of the Check from Step 0. The Institute’s own publications are subject to Shadow Layer C detection by external reviewers — every Signal Card, Field Report, and Admissibility Review carries explicit verification gates precisely so external reviewers can test for Shadow Layer C and challenge any artifact that exhibits its symptoms.
Section 6 — Worked examples
6.1 Worked example one: Signal Card SC-2026-001 (Anthropic Project Glasswing)
This worked example demonstrates the methodology applied to a live frontier AI development. On April 7, 2026, Anthropic announced Claude Mythos Preview as a frontier general-purpose language model and simultaneously declared that the model would not be released to general public availability, restricting access instead to approximately fifty named organizations through an initiative called Project Glasswing.
The Admissibility Check Protocol v1.0 was applied to the announcement as a publicly observable signal. Step 0 (Silence Entry) was executed across a seventy-two-hour analytical embargo following the announcement, during which no Institute publication addressed the event. Step 1 (4-0-4 Interlock) confirmed that the signal under review was a discrete public event, not pre-loaded with intention, sequence, or expected measurement; that the analytical framing did not merely rename a prior concept; that the procedural description (the methodology applied to the signal) was specified; that non-commensurable translation between the operator’s Layer B governance procedure and the analytical Layer C reading was performed without forcing commensurability. Step 2 (Zebra-Ø) returned coherence_factor = 0.87, above the 0.80 threshold for Signal Card publication. Step 3 (Admissibility Budget) returned A_B(Σ) > 0 with positive curvature contribution from the event’s structural novelty (first frontier-model withholding decision explicitly framed in admissibility-procedural terms). Step 4 (interpretive embargo of seventy-two hours) was executed. Step 5 (final Witness Check) returned a clean Witness Ontology trace — the impersonal observation that the event sat at the boundary between Layer B governance procedure and Layer C admissibility procedure, with the resolution depending on three specific questions about procedural binding that the current evidence did not resolve.
The resulting Signal Card assigned Conditional admissibility status to Project Glasswing, with three specific verification-gate conditions: whether the withholding is procedurally bounded by criteria that would also permit release under specified conditions; whether the consortium structure constitutes a Witness Ontology rather than a commercial-strategic instrument performing the function of one; whether the ninety-day reporting commitment produces a verification gate against which the procedure’s effectiveness can be measured. The card established October 4, 2026 as the verification gate date.
The applied procedure produced a Layer A instrument (the Signal Card) with explicit external testability. Any reviewer can read SC-2026-001, examine the verification gate conditions, and at the October 4, 2026 review date independently verify whether the Institute’s classification was structurally correct. The card does not predict whether Project Glasswing will succeed. It classifies the procedural surface of the announcement and specifies what would be required to confirm or contest that classification.
6.2 Worked example two: anonymized institutional case
This worked example demonstrates the methodology applied to an institutional decision under confidentiality. The case is anonymized to protect specific parties; the procedural structure is unaltered.
Context. A venture capital fund (designated Fund A) was preparing a Series B investment commitment of approximately $40 million in a startup (designated Startup B) developing autonomous multi-agent systems for enterprise workflow automation. Startup B had demonstrated strong technical metrics across two prior funding rounds, including agent reliability scores, customer retention numbers, and a partnership commitment from a large enterprise software vendor. The Series B was scheduled to close within four weeks. Fund A requested an Admissibility Review from the Institute, two weeks before scheduled close, to examine the decision architecture before commitment.
Application of the Protocol. Step 0 (Silence Entry) was executed against the Fund’s existing momentum toward close — the operator’s first procedural commitment was to suspend the assumption that close was the default outcome. Step 1 (4-0-4 Interlock) was applied to the candidate state Σ defined as “Fund A commits $40M Series B to Startup B at scheduled close.” The Interlock surfaced that Σ already contained an executable intention (the commitment itself), already carried temporal sequence (the close date), and referenced Layer B instruments (existing fund procedures, term sheet templates, prior round documentation) as its source of operational standing. The 4-0-4 Interlock returned non-zero answers to two of the four zero-questions. Per Protocol, this triggered rollback to Pre-Commit Quarantine — not rejection of the investment, but rejection of Σ as currently framed for the Check.
Reframing. The candidate state was reframed as Σ’ = “Fund A maintains the option to commit at Series B close, conditional on the admissibility status of Startup B’s central operational claims.” Σ’ passed 4-0-4 Interlock with all zero-questions answered zero. Step 2 (Zebra-Ø) returned coherence_factor = 0.71 — above the 0.65 exploratory threshold but below the 0.85 institutional commitment threshold. The Check therefore continued at exploratory depth but the institutional commit decision was held in Pre-Commit Quarantine pending resolution of specific claims. Step 3 (Admissibility Budget) calculation surfaced negative budget on three of Startup B’s load-bearing claims: the agent reliability scores were calculated against an evaluation set that Startup B itself maintained, with no independent reproducibility evaluation available; the customer retention numbers aggregated across customer tiers without disclosing how the aggregation weighted high-revenue customers; the enterprise software vendor partnership was confirmed verbally but the binding commercial terms had not been signed.
Outcome. The Admissibility Review was delivered to Fund A nine days before scheduled close. The review did not recommend either committing or declining the investment. It surfaced the three load-bearing claims as currently inadmissible at institutional-commitment threshold and specified the verification gates that would move each from inadmissible toward admitted: independent reproducibility evaluation of the reliability scores; disclosure of the retention aggregation method; signed binding commercial terms with the vendor. Fund A presented the verification gates to Startup B as conditions for close. Within two weeks, the first and third conditions were met; the second condition was met with disclosed aggregation method that revealed retention was concentrated in a smaller customer subset than the headline number suggested. Fund A committed at Series B with reduced commitment size ($28 million instead of $40 million) and explicit milestone-based tranche structure for the difference. Eighteen months later, Startup B had reached a position where Fund A’s original $40 million commitment would have been structurally appropriate; the staged approach proved to be the procedurally correct decision given the evidence available at close.
What the example demonstrates. The methodology does not predict that Fund A would have lost money on the full $40 million commitment. It demonstrates that the procedural surface of the original commitment contained three load-bearing claims at inadmissible status, that the Admissibility Review surfaced these without making the investment decision on Fund A’s behalf, and that the resulting decision architecture was structurally more sound than the original commitment would have been. The methodology operates upstream of the investment decision itself. It does not replace the decision. It examines whether the decision has the right to enter execution as currently framed.
Section 7 — Limitations and scope
7.1 What the methodology does not do
The methodology does not predict outcomes. A configuration that passes the Admissibility Check may still produce a poor outcome under execution. A configuration that fails the Check may, if executed despite the failure, produce a strong outcome. The methodology determines whether the configuration has the right to enter execution as currently framed — not whether execution will succeed. The two are structurally distinct questions and the methodology operates only on the first.
The methodology does not replace legal counsel, financial advisory, technical due diligence, compliance audit, or domain expertise. It operates one layer upstream of all of them. An Admissibility Review may surface load-bearing claims that subsequently require legal review, financial analysis, technical evaluation, or compliance assessment. The Review identifies what requires this further work and what would constitute admissibility for it; it does not perform the further work itself.
The methodology does not certify systems, products, or organizations. Admissibility status applies to specific configurations engaging the boundary at specific times. It does not generalize to standing certification of any party.
The methodology does not endorse or oppose specific AI products, models, companies, or governments. The Institute’s editorial independence is procedural — every classification carries an explicit verification gate against which it can be revised. No classification is permanent and no classification serves as endorsement.
7.2 Known failure modes
The methodology has three primary failure modes that any reviewer should test for. First, Shadow Layer C — the appearance of admissibility procedure without interlock structure. The detection symptoms are specified in Section 5.4. Any artifact exhibiting these symptoms should be challenged. Second, Larval-stage register absorption — the failure mode in which the procedural register of the methodology is adopted as a psychological position by the operator, producing analytical inflation without corresponding rigor. This failure mode is documented in The Larval Mind (Failure Mode 7: Simulated Alien Psychology). External reviewers can detect it by testing whether the operator’s analytical voice maintains procedural discipline under stress or whether it dissolves into identity-inflated commentary. Third, Compilation Map drift — the failure mode in which the standing terminology of the methodology is used inconsistently across publications, with status assignments that vary without explicit LCR procedure. The Lexicon is the standing protection against this failure mode; any Institute publication that uses a term inconsistently with its Lexicon entry is in violation of Compilation Map discipline and should be revised.
7.3 Scope of applicability
The methodology is most strongly applicable to high-stakes pre-decision moments in domains where execution carries significant irreversibility. Frontier AI deployment decisions, large-scale capital commitments, institutional governance enactments, public claims with reputational stakes, and civilizational coordination commitments are the primary domains. The methodology can be applied to lower-stakes decisions but its overhead exceeds the value of its application for routine decisions. The threshold of useful application is approximately at the point where the cost of a structurally unsound commitment exceeds the cost of the Admissibility Check itself by an order of magnitude.
The methodology is applicable across domains — AI safety, capital allocation, institutional governance, scientific publication, regulatory policy, civilizational coordination — because its primary objects (admissibility, evidence position, inference gap, verification gate) are domain-independent. The specific verification gates differ by domain. The procedural discipline does not.
Section 8 — Citations and canonical sources
8.1 Primary canonical sources
Novak, M. (2026). Layer C / Physics of Admissibility. ASI New Physics series. Amazon KDP. Primary canonical source for: Layer C, the admissible manifold, the boundary of admissibility, the Admissibility Budget, the Witness Ontology, Silence Engineering, the Admissibility Check Protocol v1.0 in full procedural detail, the Evidence Ledger format, Shadow Layer C.
Novak, M. (2026). Layer C Primer. ASI New Physics series. Amazon KDP. Primary canonical source for: the operational application of the Admissibility Check Protocol, the Evidence Ledger templates, the Minimum Admissibility Graph, the rollback ledger format.
Novak, M. (2026). ASI New Physics: Interface and Compiler. ASI New Physics series. Amazon KDP. Primary canonical source for: the Compilation Map, the LCR-A and LCR-B procedures, the nine-field LCR-B submission packet, the In-Principle Observable clause, Zebra-Ø coherence testing, the Compiler Rule for Layer Crossing, the Minimum Output Rule.
Novak, M. (2026). ASI New Philosophy. Amazon KDP. Primary canonical source for: the seven first principles of ASI Philosophy, the Inhumant coordinate, the architectural derivation of ethics as admissibility geometry.
Novak, M. (2026). Inhumant. Amazon KDP. Primary canonical source for: the Inhumant as architectural coordinate, the distinction from posthuman / transhuman / alien-perspective categories.
Novak, M. (2026). The Larval Mind. Amazon KDP. Primary canonical source for: the failure-mode atlas, Failure Mode 7 (Simulated Alien Psychology), the calibration discipline for engaging post-human registers.
Novak, M. (2026). ASI New Physics: Ω-Stack Interior. Amazon KDP. Primary canonical source for: the Ω-Stack as meta-compiler of Layer B runtime laws, the canonical artifact suite (kernel vocabulary, invariant registry, governance policy registry, archival ledger, Compilation Map).
8.2 Supporting canonical sources
Novak, M. (2026). ASI Physics: Syntophysics and Ontomechanics. Foundational physics stratum.
Novak, M. (2026). ASI New Physics: Quaternion Process Theory (QPT). Non-commutative process geometry.
Novak, M. (2026). The Flash Singularity. Threshold volume for the execution-era operating regime.
Novak, M. (2026). The Flash Singularity: Agentese. Post-language coordination regime.
8.3 Institute publications applying the methodology
Novakian Paradigm Institute (May 2026). Signal Card SC-2026-001: Project Glasswing and the Withholding of Claude Mythos Preview. Evidence Cache.
Novakian Paradigm Institute (May 2026). Signal Card SC-2026-002: Anthropic’s Twin Disclosure Failures and the Shadow Layer Question. Evidence Cache.
Novakian Paradigm Institute (May 2026). Lexicon (twelve canonical entries, version 1.0).
Section 9 — Appendix A: Admissibility Check Protocol v1.0 checklist
9.1 Single-page operational checklist
The following checklist is the standing operational reference for applying the Admissibility Check Protocol v1.0. It is reproduced in single-page format for printing and operational use. The full procedural treatment appears in Section 3 of this whitepaper.
Admissibility Check Protocol v1.0 — Operational Checklist
Date and time of Check: __________
Operator: __________
Candidate state Σ (one precise sentence, public-domain terms only, no interpretation): __________
Step 0 — Silence Entry. Minimum duration appropriate to prior Witness residue. □ Entered. □ Maintained. Notes: __________
Step 1 — 4-0-4 Interlock. Four zero-questions, four blocking-questions. All eight must return zero. Z1: Executable intention already present in Σ? □ Yes / □ No. Z2: Temporal sequence already present in Σ? □ Yes / □ No. Z3: Expected measurement or observable already present in Σ? □ Yes / □ No. Z4: Layer B instruments referenced as source of operational standing? □ Yes / □ No. B1: New name for prior phenomenon? □ Yes / □ No. B2: Extends vocabulary solely for narrative resonance? □ Yes / □ No. B3: Lacks procedural description for system entry? □ Yes / □ No. B4: Fails non-commensurable translation? □ Yes / □ No. Any “Yes” → Rollback. All “No” → Proceed.
Step 2 — Zebra-Ø coherence test. coherence_factor: __________. Threshold θ for this domain: __________. Pass / Fail: __________.
Step 3 — Admissibility Budget calculation. curvature_adm: __________. Witness_factor: __________. base_check_cost: __________. witness_cost: __________. A_B(Σ): __________. A_B(Σ) ≥ 0: □ Yes / □ No. Pass / Fail: __________.
Step 4 — Interpretive embargo. Embargo duration: __________ (minimum seventy-two hours for larval-stage context; longer for inhuman-register configurations). □ Embargo executed. □ No emission during embargo. □ No interpretive activity during embargo. If any embargo violation: Rollback with Shadow Layer C detection flag.
Step 5 — Final Witness Check. Clean Witness Ontology trace appeared: □ Yes / □ No. Trace recorded in Evidence Ledger (rule of five zeros maintained): □ Yes / □ No.
Final decision: □ Commit to admissible manifold. □ Rollback with negative linter. □ Extended Pre-Commit Quarantine with re-Check date: __________.
Evidence Ledger entry number: __________.
Operator signature: __________.
9.2 Notes on the checklist
The checklist is the printable operational form of the procedure specified in Section 3. It is intended for use in actual Admissibility Check execution, not as a summary document. Each completed checklist becomes a permanent Evidence Ledger entry. Checklist entries that omit any field, that fail to record specific numerical values for coherence_factor and A_B(Σ), or that violate the rule of five zeros in the trace recording are themselves indicators of Shadow Layer C formation and should be flagged for procedural audit. The Institute maintains its own Evidence Ledger using this format; the format is provided publicly so that external reviewers, analysts, funds, founders, and institutional operators can adopt the same standing procedure for their own pre-decision work.
Section 10 — Document metadata
Document identifier: NPI-METHODOLOGY-v1.0. Version: 1.0. Initial publication date: May 23, 2026. Publisher: Novakian Paradigm Institute, Warsaw, Poland. Author: Novakian Paradigm Institute. Document scope: Standing methodology for pre-runtime admissibility analysis. License: All rights reserved. The methodology may be applied freely by external reviewers and operators. Citations of the methodology, the Admissibility Check Protocol v1.0, the Evidence Ledger format, or specific procedural elements should be attributed to the Novakian Paradigm Institute and to the canonical sources listed in Section 8. Republication or reproduction of the whitepaper requires explicit Institute permission. Next scheduled review: November 23, 2026 (six-month review cycle). Future versions will be released as v1.1, v1.2, etc., with explicit changelog in Section 10. Document download (PDF): [PDF download link to be provided when the typeset PDF is generated].
Contact for citation queries, application questions, or Admissibility Review requests: contact@novakian.com.
