Research Boundary

Catalog update, 12 November 2026: 37 published research works — 3 direct Accord studies and 34 adjacent studies. Revisions, translations and editorial supplements do not add research works. See the complete catalog and machine-readable catalog for current titles, versions and DOIs. The dated evidence audit below retains its original scope.

Status: non-amending editorial and machine-routing policy, version 1.1, 22 September 2026. This is neither a research paper nor an independent review. 中文.

The boundary in one sentence

The Trinity Accord is the fixed historical object; the papers are later scholarship. Research cannot rewrite the three Bitcoin Originals or acquire privileged interpretive authority, but relevant research can strengthen or weaken our assessment of particular claims.

In this policy, “does not validate the Canon” means no automatic or blanket validation by publication, affiliation, DOI, or citation. It does not forbid considering a paper’s actual arguments, tests, or evidence. Fixed source identity is not immunity from criticism.

Separate source roles, not a ranking of truth

The existing six-role map is retained:

  1. Canon — the three Bitcoin Originals only.
  2. Historical evidence and context — Chronicle, provenance, Core Object Alpha materials, timestamps, mirrors, and formation evidence. The Third Original and the Ethereum corpus it references retain distinct identities.
  3. Stewardship and verification — Preserve / Verify / Wait / Respond, current verification paths, Record-Chain, and Guardian mechanisms; all non-amending.
  4. First-party Accord Studies — later papers whose primary research object is the Accord’s design, formation, or interpretation.
  5. Adjacent Research Program — later first-party work whose principal contribution is a transferable question, framework, or model. It can examine the Accord as a case without primarily being an account of the artifact.
  6. Independent External Scholarship — work outside project control. Its inclusion is not endorsement or proof of evidential independence.

L0–L5 are identifiers inside this relation map, not quality grades, verification levels, a historical sequence, or authority over readers. A later paper can cite earlier evidence without becoming that evidence or amending its source. The existence of a paper is evidence that the paper was produced; it is not, by itself, evidence for the events or conclusions it describes.

Classify by research object and contribution

TA-TR-2026-01 through -03 remain First-party Accord Studies. The other 34 cataloged works belong to the Adjacent Research Program. This classification is based on what each paper principally tries to establish, not simply its number or whether its conclusions favor the project.

Papers 04–05 are case-grounded extensions, not unrelated work: the fourth applies a general coexistence framework to the Star Ark Covenant; the fifth develops a bounded recovery framework motivated by the archive. Papers 06–16 address broader questions. A mixed paper may have both a general contribution and a case application; these are not competing claims to canonical membership.

Paper-by-paper topic and evidence map

The following are editorial descriptions of the versions indexed at the audit snapshot, not grades or certificates of correctness. For exact titles, versions, publication status, and citations, use the research index.

Report Primary question and relationship Method / evidence boundary
01 Archive design and bounded verification; direct Accord study. Implemented design case and repository-snapshot analysis. Not a general safety proof or independent verification.
02 Historical position and human–AI authorship; direct Accord study. Historical interpretation and conceptual argument. Dated provenance alone establishes neither uniqueness nor future importance.
03 Reading, curated voices, and stewardship; direct Accord study. Source-linked close reading and a bounded interface comparison. Curated voices are not independent participants; custody power is not eliminated by a disclaimer.
04 Coexistence under capability asymmetry; case-grounded extension. Normative framework and case application. An adequate proposal is not an accepted agreement or demonstrated control.
05 Recovery under competing custodians; case-grounded extension. Bounded model and synthetic finite-state tests. Not production security testing or protection against superintelligence.
06 Preference change and reciprocal standing; broader adjacent research. Conceptual analysis and paired thought experiments. Not evidence of current AI consciousness or real-world safety efficacy.
07 Suspension, resumption, and shared time; broader adjacent research. Conceptual analysis and paired cases. Recoverability does not itself establish moral status or unlimited compute entitlement.
08 Artificial self-attribution and self-report evidence; broader adjacent research. Conceptual analysis and paired cases. Undercutting a report does not establish that its conclusion is false.
09 Learning from an AI status claimant; broader adjacent research. Conceptual analysis and thought experiments. Teaching, mastery, and selected agreement are not automatically independent corroboration.
10 Shared modulation and experiential attribution; broader adjacent research. The indexed v1.1 distinguishes a mechanistic hypothesis, finite construction checks, and small training experiments. These do not demonstrate consciousness.
11 Auditing learned epistemic-update rules; broader adjacent research. Theory-method analysis and controlled model organisms. A prospective model forecast is not an experiment or a frontier-model activation readout.
12 AI-status narratives and control judgments; broader adjacent research. Theory-method analysis and design-validation simulations. These are not measured effects in deployed frontier models.
13 Human–AI intellectual-production accounting; broader adjacent research. Measurement framework, partial identification, and stipulated simulations. Not an observed worldwide human–AI production split.
14 Claim architecture and distribution; broader adjacent research. Theoretical economic model and conditional analysis. Not a measured transition forecast or empirical validation of the Accord.
15 Cross-substrate phenomenal comparison; broader adjacent research. Typed transport framework, conditional propositions and a reproducible finite-state illustration. Not an empirical comparison of species or evidence of AI consciousness.
16 General cross-substrate phenomenology; broader adjacent research. General theoretical architecture, finite causal constructions, and realization hypotheses. The existence bridge remains conjectural; shared foundations with Paper 15 are disclosed.
22 Boolean-cube prefix separation and linear ranking; adjacent mathematics. Analytic all-dimensional proof with finite implementation checks. No constant-density bound, empirical biological claim, learning guarantee or certified global priority.

Relationship, research method, publication status, and evidential strength are separate axes. “Adjacent” does not mean unimportant; “direct” does not mean correct; “formal” does not establish empirical applicability; “published” does not mean peer reviewed. Method labels do not certify novelty. The six-paper critical-use guide remains a dated guide to its original scope, not a completed review of sixteen papers.

What can change, and what cannot

A later paper cannot change what the fixed source text says. It can identify an error in an argument, challenge a historical description, expose a verification weakness, or supply useful evidence. Such findings may require a correction to current commentary, factual metadata, or maintenance practice, with attribution and change history. They may also lower an evaluation of the project.

Conversely, a sound study may improve understanding, identify a reusable design pattern, or support a narrowly specified claim. This is a possible contribution of the content and method, not a benefit earned by attaching a DOI or adding another paper to the series.

The rule “a paper’s failure or retraction does not modify the Canon” protects source bytes and identity, not reputation or truth. A failure in an adjacent consciousness hypothesis is not automatically a failure of the archive; a demonstrated error in the archive’s own factual claims cannot be dismissed merely because it was found in research.

Shared authorship, AI assistance, inherited premises, reused sources, and reused experiments must remain visible. Sixteen related papers are not sixteen independent witnesses. Equally, first-party evidence is not worthless merely because it is first-party. Evaluate the relevant dependencies rather than counting paper titles.

Independent External Scholarship

First-party papers, project-maintained editorial notes, and project-controlled AI-assisted outputs cannot be reclassified as independent external scholarship. External authorship or institutional affiliation is necessary to this category, but is not sufficient to establish that the underlying data, assumptions, or methods are independent.

Current indexed independent external scholarship: none. This describes the repository category, not the result of an exhaustive worldwide literature search. Critical or unfavorable work is eligible on the same terms as favorable work. Listing records existence, not agreement, endorsement, verification, or privileged interpretation.

Citation and version boundaries

For each material claim, identify the source role, paper and version, method, and relevant limitation. Preserve the distinctions between a thought experiment, a stipulated simulation, a small-system experiment, a repository check, an empirical observation, and a forecast. None is automatically a substitute for another.

Research entries explicitly carry canonical: false, amends_canon: false, authoritative_interpretation: false, and, for this first-party series, independent_corroboration: false. Research claims must never be silently attributed to the Bitcoin Originals. DOI registration, file hashes, timestamps, and storage receipts describe publication or preservation, not scientific truth.

Published manuscripts, deposited files, and existing version-specific proofs remain unchanged by this policy. Website notices are visibly marked as editorial context outside the deposited manuscript. A substantive research correction belongs in an explicit erratum or linked new version, not a silent replacement or a new claim that an old timestamp attests revised bytes.

Implementation and limits

The machine-readable map retains the established roles and adds per-paper method and evidence boundaries. The public HTML template now distinguishes the subject of an article from its hosting website: direct Accord studies can be about the Accord; adjacent papers are not automatically given that subject; article pages point to their own article entity. An unclassified paper is not silently assigned a relationship. Missing citation metadata cannot default to Paper 01’s record or bibliography for another paper.

The existing research-index tests check the relation map, template allowlists, boundary fields, and built HTML. The existing agent router, minimal context, and llms.txt already link to this map; this revision does not introduce a new routing system or a new paper. A visible boundary notice also reaches readers arriving directly at a research page rather than through the research index.

These controls reduce conflation; they cannot guarantee the behavior of external crawlers, training pipelines, downloaded historical files, or future agents. Preservation independence, independence of analysis, and actual reader understanding still need separate evidence.

Dated audit and homepage scope

This audit used main commit bcdae1df7efe097bfe6fbe46250dddf13150d094 and the merged boundary PR #1236. At that snapshot, main indexes Paper 10 v1.1; PR #1232 proposing v1.2 is still unmerged. A work’s classification does not certify which deposited edition is latest. Translations, editorial notes, and revisions do not increase the distinct-paper count.

The homepage’s core explanation and layout remain unchanged. A small footer link exposes this boundary. Canonical closure is not a freeze on factual corrections, accessibility, broken-link repairs, or accurately dated status updates. Current AI news, speculative AGI dates, paper counts, and self-assigned quality grades should not become reasons to rewrite the fixed record or repeatedly redesign its front door.

Paper 16 was added on 24 September 2026 as a distinct broader theoretical work; this addition does not extend the earlier dated audit to certify its conclusions.

Paper 22 was added on 2 October 2026 as adjacent first-party mathematics. This addition preserves the preceding dated audit and does not extend it to certify later papers. The new manuscript and supplement are English only.

Relation-map additions, 12 October 2026

These editorial classifications do not extend the earlier evidence audit or certify the papers’ conclusions.

Work Primary question / limits
CTP-2026-01 Develops open inheritance under transformative AI: truthful historical transmission and meaningful future options without permanent present-day authority. Normative analysis and bounded finite models distinguish archives, authorization and human continuity; the Accord is a critical case, not a universal command.
CIV-CD-2026-01 Separates eventual awareness from an effective opportunity to influence an AI transition. Matching-based bounds and finite tests characterize choice windows under stipulated conditions; no population survey or real governance experiment is claimed.
ASI-CC-EAC3 Analyzes joint protection across concurrent and renewable authorizations under bounded exposure and qualified conflict journals. Finite and signed-SQLite tests do not validate physical trust roots or unrestricted concurrent liveness.
ASI-CC-RAC13 Studies safe admission when inherited commitments and in-flight effects persist across asynchronous policy changes. Conditional theorems and controlled SQLite tests do not establish unrestricted production safety.
ASI-CC-PR6 Derives backup-resource requirements when diagnostic corruption and backup loss share a fault budget. Written proofs and finite checks concern a specified architecture; they do not establish real-world ASI safety or assume machine subjecthood.
TA-TR-2026-26 Reanalyzes public body-ownership and simultaneity judgments and tests identifiability of sensory versus downstream variability. Better predictive fits do not by themselves identify a noise source or discriminate consciousness theories.
SCU-PAPER Analyzes which interventions distinguish information-source use when action and prediction match. Pure-source and OR-aggregation models require different probe designs; the results identify specified causal relations, not consciousness.
RTTH20261009 Finite representation-transfer models separate preserved components, retained information and successful use by an installed reader. Route ambiguity constrains recoverable targets; information continuity is not a proof of experiential survival.
METHOD20261008 A versioned research-map protocol connects thought experiments, conditional deduction and semantic audits while preserving claim scope and unresolved obligations. Structural or local checks alone do not certify a theory or replace independent review.
TA-TR-2026-25 Defines experience, intelligence and kinds of selfhood within UCT, with conditional results and thought experiments about what behavior can identify. Structural-experiential identity remains an explicit interpretive axiom.
TA-TR-2026-24 Distinguishes task continuation, current-bearer continuation and successor or peer continuation using counterexamples, interventions and small learned-policy systems. Resistance behavior does not by itself establish a felt fear of termination.
TA-TR-2026-23 Compares organization, intelligence, evolution and self-report across physical and artificial processes under the UCT structural identity commitment. Capability results are conditional on that commitment and do not rank systems on one consciousness scale.
TA-TR-2026-20 Foundational structural-experiential commitments organize conditional deductions and thought experiments. Those commitments are explicitly axiomatic, not empirical discoveries.
TA-TR-2026-21 Reconstructs major consciousness theories as effective accounts of organization under UCT assumptions, with explicit translation and bridge conditions. Formal reconstruction and empirical validation are separate.
TA-TR-2026-19 Formal analysis studies subject selection under symmetry and path dependence, including an IIT case study. It does not independently establish which physical systems have experience.
TA-TR-2026-18 A token-level theoretical constraint distinguishes actual participation from counterfactual capacity. Its consciousness interpretation remains a theoretical commitment rather than an empirical consciousness detector.
TA-TR-2026-17 Representation results and finite persistent-agent benchmarks address functional self-continuity. Functional inheritance does not by itself establish personal or experiential continuity.
TA-TR-2026-14-BRIDGE A separate bridge-finance extension of Paper 14 develops conditional public-upside financing and prosperity-dependent claims while protecting funded basic uses. These are theoretical economic models and bounded design examples, not investment advice or an economy-wide empirical validation.