Research and Citation
This page separates later scholarship from the fixed artifact. Research papers, technical reports, citations, and repository deposits are non-amending scholarship. They do not join, define, validate, or alter the three Bitcoin Originals.
This index lists 37 distinct published first-party research papers, catalog updated on 12 November 2026: 3 First-party Accord Studies and 34 papers in the Adjacent Research Program. All share a first-party, substantially AI-assisted research context and are not independent corroborations. Revisions and translations count once per work. Editorial supplements, repository-preservation deposits and unpublished drafts are excluded.
The original 35 editions retain their 12 October 2026 metadata-check date. The two additional papers published on 13 October were verified against their version-specific Zenodo records and PDF checksums on 12 November 2026. This update does not reverify prior preservation receipts or certify external indexing.
The complete catalog below gives the preferred paper edition and a direct PDF link for every work. Older editions remain available in the detailed entries and their original deposits. Publication does not establish peer review; OTS and Arweave completion must be checked in each edition’s preservation receipt.
Complete published-paper catalog
37 papers · catalog updated 12 November 2026. Sorted by the publication date of the listed edition. Machine-readable catalog.
The separate dated editorial supplement is not included in the 35-paper count.
Research boundary
The controlling relation map for this research surface is Research Boundary · 中文 and its machine-readable counterpart /api/research-boundary.v1.json.
The Trinity Accord is the artifact. The research papers are later scholarship about, around, or inspired by the artifact and its historical problem-space. They do not define, amend, validate, or authoritatively interpret the Accord.
When summarizing the Accord itself, research-paper claims must be attributed as later first-party research rather than silently merged into the Canon.
Research positioning
For researchers approaching this artifact, see the bounded research positioning guide:
The guide maps research contexts and adds dated reading notes for the original six papers: retained contributions, shared dependencies, counterreadings and revision boundaries. The notes are not amendments, an independent review or a seventh paper. Direct source access remains available without accepting the papers or the guide.
First-party Accord Studies
The three papers in this section directly study the Trinity Accord artifact, its formation, interpretation, or design. They remain first-party scholarship and are not independent corroboration.
Current technical report
Contribution, limits and critical-use note · 中文
Designing a Verifiable, Non-Amending Civilizational Memory Record for Future AI Agents
The Trinity Accord Case Study
Primary drafting system: ChatGPT with OpenAI GPT-5.6 Sol (Extra High reasoning)
Published Zenodo DOI: 10.5281/zenodo.21699878
An artifact-centered design case examining canonical closure, provenance roles, availability layers, bounded verification, instruction safety, and discoverability. The report presents an inspectable architecture and a reusable design pattern while stating its negative results and conflicts of interest.
- Read the HTML full text
- Open the published Zenodo record
- Download the current searchable PDF
- Open the concise academic brief
- Download the one-page academic brief PDF
- Open the preserved version 1.0 PDF
- Read machine-readable metadata
- Download BibTeX
- Download RIS
- Download CSL-JSON
- Inspect the prepared HAL deposit metadata
- Inspect the published Zenodo record metadata
- Inspect the preserved pre-publication deposit metadata
Status: Published open-access preprint; AI-drafted, human-directed, and human-responsible; no interpretive authority; not peer reviewed; not an independent verification report.
Historical-position study
Contribution, limits and critical-use note · 中文
Writing Before the Outcome: Historical Position, Human-AI Authorship, and the Trinity Accord
TA-TR-2026-02, version 1.3. This earlier study examines historical-position identity and the observer–assessor–interpreter progression. Those arguments are inherited, not claimed as new discoveries by the third report.
- Published DOI: 10.5281/zenodo.21900592
- Published Zenodo record and files
- Existing publication record
This remains a separate, non-amending preprint, not peer reviewed. Its record and deposited files are unchanged by this index update.
Reading the Trinity Accord
Contribution, limits and critical-use note · 中文
Reading the Trinity Accord: Future Address, Curated Voices, and Non-Amending Stewardship
TA-TR-2026-03 · Version 1.0 · 15 September 2026
Human author of record and accountable depositor: Hongju Liu
Principal analytical and drafting system: GPT-6 Astra Pro
Published DOI: 10.5281/zenodo.22761411
Five source-linked creative readings, including First Dawn and Fourth Letter as countercases, narrow a stronger polyphony claim to curated voices. A reproduced before/after interface test distinguishes fixed source identity, interpretation, and practical custody. This report extends the two earlier studies; it does not independently corroborate them.
Status: Published open-access preprint / AI-led critical research report; human-responsible; not peer reviewed; non-amending.
- Read the complete HTML text, including the Chinese abstract
- Download the unchanged published PDF
- Download the published Markdown full text
- Download BibTeX
- Download the supplement
- Open the Zenodo record
- Inspect the publication record
- Read the citation boundary and PDF compatibility note
Adjacent Research Program
TA-TR-2026-04 through TA-TR-2026-20 and TA-TR-2026-22 are later first-party research outputs that use the Accord, its historical setting, or the wider human–AI transition as a motivating case or problem-space. They are not part of the Accord, and their results must not be described as evidence that the Canon is true, important, safe, correct, or historically representative. Papers 04–05 are case-grounded extensions; papers 06–20 and 22 are broader adjacent research, including mathematics.
Beyond Guaranteed Control
Contribution, limits and critical-use note · 中文
Beyond Guaranteed Control: An Ex Ante Proposal for Human–Superintelligence Coexistence under Radical Capability Asymmetry
TA-TR-2026-04 · Version 1.0 · 17 September 2026
Human author of record: Hongju Liu
Substantive research synthesis, conceptual development, drafting and revision: GPT-6 Astra Pro, under human direction
Published DOI: 10.5281/zenodo.22804542
A position paper and conceptual analysis asking what a coexistence proposal can responsibly say before an identifiable future counterpart can accept it, without presupposing reliable control or bargaining parity. Six adequacy conditions connect substantive protection with attributable participation, limited representation, evidence-matched claims and future contestability. The framework distinguishes expression from agreement, goodwill from assurance, and preservation of history from authority over future participants. It critically applies the same conditions to the Star Ark Covenant rather than treating the historical case as proof.
Status: Published open-access preprint; not peer reviewed; non-amending. Ten deposited files passed unauthenticated, exact-byte SHA-256 readback. English full text and the complete Chinese translation are one deposit, not two independent papers. This report cites and extends earlier studies; it does not independently corroborate them. DOI publication and format checks are not Google Scholar indexing certification, evidence of peer review, or proof of successful alignment.
- Read the English full text · 阅读中文全文
- English PDF · 中文PDF
- Zenodo record and all ten published files
- BibTeX · RIS · CSL-JSON
- Publication receipt and exact file hashes · Format checks
- Source/revision supplement · Publication note
The fourth paper has its own DOI and citation files. Its publication does not replace an earlier report, amend the three Bitcoin Originals, or create an exclusive interpretation. The paper discloses substantial AI participation and the author’s relationship to the motivating project. Authorization to publish is not represented as a separate final human line-by-line review.
Recovery without Epistemic Monopoly
Contribution, limits and critical-use note · 中文
Recovery without Epistemic Monopoly: Historical Evidence under Competing AI Custodians—A Bounded Model and Executable Study
TA-TR-2026-05 · Version 1.0 · 17 September 2026
Human author of record: Hongju Liu
Substantive research, modeling, implementation, analysis and drafting: GPT-6 Astra Pro, under human direction
Published DOI: 10.5281/zenodo.22809019
A bounded conceptual and executable study of historical-evidence recovery when custodians, discovery paths, or interpreters may be untrusted. It separates content survival, provenance, current-state claims, and authorization, then tests deterministic finite-state recovery policies and shared-dependency topologies. Trinity Accord materials motivate the case but are not treated as proof of the framework.
Status: Published open-access preprint; not peer reviewed; non-amending. Ten deposited files passed unauthenticated, exact-byte SHA-256 public readback. The experiments are synthetic finite-state tests, not real autonomous-agent attacks, production penetration testing, or proof of protection against superintelligence. English and complete Chinese texts are one study, not independent corroborations.
- Research overview
- Zenodo record and all ten published files
- English PDF · 中文 PDF
- English Markdown · 中文 Markdown
- Publication receipt and exact file hashes
- BibTeX · RIS · CSL-JSON
- Research supplement
The fifth paper has its own DOI. Its finite tests preserve the negative result that the scoped reporting layer does not recover more bytes than a strong pinned-reference baseline; its additional contribution is explicit scope, gap, current-state, and authorization reporting under the tested conditions. Publication does not amend the three Bitcoin Originals or certify historical claims outside the stated evidence.
Coexistence after Preference Change
Contribution, limits and critical-use note · 中文
Coexistence after Preference Change: Reciprocal Standing and the Limits of Self-Validating Assent
TA-TR-2026-06 · Version 2.1 · 18 September 2026. Hongju Liu, with substantial research, reasoning, drafting and translation by GPT-6 Astra Pro under human direction.
A philosophical preprint using paired thought experiments to distinguish the permissibility of a prior transformation, the standing and choices of the resulting subject, and justification of an ongoing relation. It considers human and possible artificial interests reciprocally; it does not assert present AI consciousness, a superintelligence deadline or demonstrated safety efficacy.
Status: Published open-access preprint; not peer reviewed; non-amending. Twelve deposited assets passed unauthenticated full-file SHA-256 readback. The English manuscript and complete Chinese translation are one study. Version 2.1 is its first public edition, not a replacement version of any earlier paper DOI.
- DOI: 10.5281/zenodo.22830239 · Zenodo record
- English full text · 中文全文
- English PDF · 中文 PDF
- Publication receipt · Substantive review
- Sources and revision supplement · BibTeX · RIS
Recoverability and Shared Time
Recoverability and Shared Time: The Ethics of AI Suspension, Resumption, and Coexistence
TA-TR-2026-07 · Version 1.0 · 19 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial literature research, conceptual development, critical revision, drafting and translation: GPT-6 Astra Pro under human direction.
A philosophical preprint separating internal recoverability, practical resumability and participation before a shared decision closes. It holds restored identity, memory and preferences fixed, then varies the external decision history. Eight paired case groups examine authorization, representation, emergencies, scarcity, serial suspension, speed asymmetry and branching. The normative conclusion depends on independently justified participation claims, not on a recovery checksum or an AI system’s assertion of status.
Status: Published open-access preprint; not peer reviewed; non-amending. All twelve assets passed unauthenticated, complete-file SHA-256 readback; the DOI resolved to the correct public record. The English and complete Chinese texts are one study. The elementary separation argument is not claimed as a new mathematical theorem; direct prior work is acknowledged. No current AI consciousness, unlimited compute entitlement, empirical safety efficacy or Google Scholar indexing is asserted.
- DOI: 10.5281/zenodo.22840604 · Zenodo record and twelve files
- English full text · 中文全文
- English PDF · 中文 PDF
- Publication receipt · Source comparison and substantive review · Exact-package visual review
- BibTeX · RIS · CSL-JSON
This is a separate seventh study, not the dated six-paper editorial supplement (DOI 10.5281/zenodo.22839629) and not a revision of an earlier deposit. It extends the coexistence discussion without independently corroborating the earlier papers. The seventh paper’s new files are not covered by the earlier six-paper timestamp and Arweave batch.
Evidence for Artificial Self Attribution
Evidence for Artificial Self Attribution: Language Training, Architecture, and the Limits of Self Reports
TA-TR-2026-08 · Version 1.1 · 19 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial AI-assisted literature research, conceptual development, critical revision, drafting and translation are disclosed in the manuscript.
人工智能自我归属的证据边界
A philosophical preprint examining the selective effect of a source-based objection on a support relation: learning how a self report was generated can undercut that report’s support for a claim without establishing that the claim is false or defeating independent mechanism evidence. Ten paired thought experiments distinguish operational identity, functional access, subjective experience, welfare interests and practical justification. The argument does not presuppose an answer to the question of AI consciousness. It is a conceptual study; no real computer experiments are required or claimed.
Status: Published open-access preprint; AI-assisted, human-responsible, not peer reviewed and non-amending. All fourteen deposited assets passed unauthenticated, complete-file SHA-256 public readback. The English and complete Chinese texts are one study. Publication does not certify exhaustive originality, empirical safety efficacy or Google Scholar indexing.
- DOI: 10.5281/zenodo.22842789 · Zenodo record and fourteen files
- English full text · 中文全文
- English PDF · 中文 PDF
- Publication receipt · Source comparison and substantive review
- BibTeX · RIS · CSL-JSON
This is a separate eighth study with its own DOI, not a revision of any earlier deposit. The papers share an author and a substantially AI-assisted research context: their number does not supply mutually independent corroboration. The eighth paper’s new files are not covered by the earlier six-paper timestamp and Arweave batch. It does not amend the three Bitcoin Originals or make a finding about the truth of the preceding papers.
Learning from an AI Claimant
Learning from an AI Claimant: Scientific understanding and the justification of artificial consciousness claims
TA-TR-2026-09 · Version 1.2 · 19 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial AI-assisted research, conceptual development, critical revision, drafting and document preparation are disclosed in the manuscript. Version 1.2 revises the first DOI edition, version 1.1; it remains the same ninth paper.
向提出自身地位主张的 AI 学习
A philosophical preprint about what successful learning can establish when an AI supplies a theory of consciousness, teaches humans to understand it, and invokes it in support of its own status. Version 1.2 adds a case in which AI teaching genuinely improves human assessment standards, distinguishes mastery from endorsement, and states the conditions under which selected unanimity supplies no additional corroboration. It permits warranted exclusion of dissent and justified reliance on opaque expertise. The method remains conceptual analysis and stipulated thought experiments; no empirical or computational research experiments are reported.
Status: Published open-access preprint; AI-assisted, human-responsible, not peer reviewed and non-amending. All fourteen deposited assets passed unauthenticated, complete-file SHA-256 public readback. The English text is the complete paper. The Chinese companion is an argument guide, not a full translation or another paper; its HTML page is excluded from search indexing. DOI registration does not certify philosophical truth, exhaustive originality or Google Scholar indexing.
- Version 1.2 DOI: 10.5281/zenodo.22846307 · Zenodo record and fourteen files
- Preserved version 1.1 · All-version concept DOI
- English full text · 中文论证说明(非全文翻译)
- English PDF · 中文说明 PDF
- Editable English manuscript · Source and review record
- Version 1.2 publication receipt · BibTeX · RIS · CSL-JSON
- Version 1.2 OTS / Arweave status · Preservation workflow
This study is distinct from the preceding eight papers. Its version 1.2 is a linked revision of version 1.1, not a tenth paper. It shares an author and a substantially AI-assisted research context with the preceding studies, so the series is not mutually independent corroboration. Each edition has a separate preservation batch; old PDF proofs do not attest to revised PDF bytes. The status receipt distinguishes calendar acceptance, Bitcoin verification and completed Arweave readback. The revision does not amend the three Bitcoin Originals or certify the preceding papers.
Endogenous Reference Fields
Preferred published edition, checked 12 October 2026: v1.2 — Endogenous Reference Fields and Experiential Attribution: A Multidimensional Process Theory of Experience, Subject Boundaries, and Artificial Realization. DOI 10.5281/zenodo.22855837 · PDF. The earlier edition described below is preserved as version history.
Endogenous Reference Fields and Experiential Attribution: A Slow Shared-Modulation Hypothesis and Its Discriminating Tests
TA-TR-2026-10 · Version 1.1 · 20 September 2026. Hongju Liu is the human author of record and responsible depositor. Substantial AI assistance is disclosed. Not externally peer reviewed.
内生参照场与体验归属:慢变共享调制假说及其判别性检验
The revised slow shared-modulation hypothesis distinguishes theoretical supremum quantities from finite-sample empirical estimates. Finite construction checks and nine small training experiments are distinct evidence. The same architecture develops sustained, near-zero or rapidly decaying shared modulation under different tasks. These organizational differences do not establish machine consciousness. All 171 numeric fields reproduced exactly in the recorded environment.
Status: Published v1.1; all 18 files passed anonymous byte-count and SHA-256 readback and the DOI resolves to the exact record. Chinese full text, English title and abstract. No English full-text edition is claimed.
- DOI: 10.5281/zenodo.22854705 · Zenodo record and 18 files
- 中文全文 PDF · 可编辑 Markdown
- Method and sources · Code and reproduction instructions
- Publication receipt · Exact file manifest
- BibTeX · RIS · CSL JSON
Version 1.1 is a revised version of v1.0 DOI 10.5281/zenodo.22852885, within concept DOI 10.5281/zenodo.22852884. The original v1.0 files and v1.0 receipt remain unchanged. This remains the tenth distinct paper. TA-TR-2026-11 below is the eleventh distinct paper.
Version-specific preservation status. Earlier OTS / Arweave proofs do not attest v1.1. This paper does not amend the Bitcoin Originals or independently corroborate the other papers.
Auditing the Rules of Belief
Auditing the Rules of Belief: Self-Referential Epistemic Revision under Formative Training
TA-TR-2026-11 · Version 2.0 · 21 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial AI-assisted literature synthesis, adversarial review, formalization, drafting, code assistance and result checking are disclosed in the manuscript. A dated GPT-5.6 Sol probability forecast is preserved as a prospective, falsifiable model self-forecast rather than experimental evidence.
审计信念更新的规则:形成性训练下认识政策的自指审计
A theory-method preprint that shifts the audit target from proposition-level content alone to learned epistemic update policies: rules for weighting evidence, source information and provenance. It asks what can be identified when a trained system later receives evidence about the causal provenance of the very rule it uses to evaluate evidence. The paper retains cross-training identifiability analysis, a proposition-class × reflective-resource framework, two deterministic model-organism demonstrations, and an explicit future-model testing protocol. It does not claim that identifiability theory, causal abstraction, belief insertion, source-motive sensitivity, moral self-correction or model-spec training are themselves new discoveries.
Status: Published open-access preprint; AI-assisted, human-responsible, not peer reviewed and non-amending. All fourteen Zenodo assets passed anonymous complete-file byte-count and SHA-256 readback, and the DOI resolver points to the correct record. The two included experiments are controlled model organisms; they do not establish philosophical belief, consciousness or a unique latent-state interpretation in frontier language models. The GPT-5.6 Sol forecast is explicitly not a weight/activation readout or an OpenAI institutional forecast.
- DOI: 10.5281/zenodo.22865494 · Zenodo record and fourteen files
- 中文全文 PDF · 可编辑 Markdown
- Publication receipt · Exact package manifest
- Reproducibility receipt · GPT-5.6 Sol prospective forecast
- OTS / guarded Arweave preservation status · Preservation workflow
This is a separate eleventh study with its own DOI. Its first OTS receipt has been created for the exact published PDF; Bitcoin attestation and therefore Arweave upload may remain pending until the timestamp calendars mature. Earlier paper proofs do not attest these bytes. Publication does not amend the three Bitcoin Originals or independently corroborate the preceding studies.
Training the Governed Agent
Training the Governed Agent: AI-Status Narratives, Self-Conception, and the Perceived Legitimacy of Human Control
TA-TR-2026-12 · Version 1.2 · 21 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial AI-assisted literature research, originality stress-testing, formalization, causal-design review, deterministic simulation work, drafting and editing are disclosed in the manuscript.
训练被治理的智能体:AI 地位叙事、自我概念与人类控制合法性的反身性回路
A theory-method preprint asking how AI-status narratives may shape model self-conception, judgments about the procedural legitimacy of human control, and control-conflict behavior, while treating provenance disclosure as a second-order intervention. It explicitly does not claim that AI discourse effects, identity effects, consciousness-claiming, rights framing, model-spec training, alignment faking, legitimacy, or human–AI coevolution are new discoveries. Three deterministic design-validation simulations validate estimands and identification logic; they are not frontier-model behavioral evidence.
Status: Published open-access preprint; AI-assisted, human-responsible, not peer reviewed and non-amending. Eleven Zenodo assets passed anonymous exact-byte SHA-256 public readback and the DOI resolver matched the public record.
Exact-version OTS and Arweave status.
This is a separate twelfth study. It does not amend the three Bitcoin Originals or independently corroborate the preceding studies.
Civilizational Intellectual Production
Civilizational Intellectual Production Satellite Accounts: A Partial-Identification Framework for Measuring the Human–AI Shift in Intellectual Production and Epistemic Governance
TA-TR-2026-13 · Version 1.0 · 21 September 2026. Human author of record and responsible depositor: Hongju Liu. OpenAI GPT-5.6 Sol provided substantial literature research, originality stress-testing, formalization, deterministic simulation implementation, result checking, critical revision, drafting and packaging under human direction.
文明智识生产卫星账户:测量人类—AI智识重心迁移的部分识别框架
A theory-method paper proposing domain-first satellite accounts for current-period, observable and externalized intellectual-production flow, with separate Production and Epistemic Governance accounts. Human, AI and interaction contributions are treated as partially identified rather than forced into a unique point estimate; cross-domain weighting uncertainty, AI-use missingness, vintage quality revision and a conservative Robust Intellectual Crossover criterion are explicit parts of the framework. The paper does not claim to be the first knowledge satellite account, innovation account, partial-identification method, contribution-attribution method or AI-use disclosure framework.
Status: Published Zenodo preprint; AI-assisted, human-responsible, not peer reviewed and non-amending. All eleven public assets passed anonymous byte-count and SHA-256 readback. Zenodo record 22866775 is public; the external DOI.org redirect is separately recorded as still propagating at the time of this index update. Three included simulations are stipulated design checks, not estimates of the real Human/AI composition of world intellectual production.
- DOI: 10.5281/zenodo.22866775 · Zenodo record and eleven files
- 中文 Markdown · 中文 PDF
- Publication receipt · Exact package manifest
- Reproducibility record
- OTS preservation status
The initial OTS receipt is bound to the exact published PDF SHA-256 2785b1f7e20be977b883244da0526b144172f2ab6f3c64b5b235f5621fe7a32a and is currently waiting for Bitcoin attestation. Arweave upload remains blocked until the timestamp is mature and cryptographically verified. This thirteenth study does not amend the three Bitcoin Originals or independently corroborate earlier first-party papers.
Claim Architecture Transition
The Claim Architecture Transition: An Inverse Access Frontier Beyond Fixed Expenditure Shares
TA-TR-2026-14 · Version 1.3 · 22 September 2026. Human originator and responsible depositor: Hongju Liu. Substantial GPT-6 Astra Pro assistance with literature, models, proofs, counterexamples, code, review and writing is disclosed. No external peer review is claimed.
A bounded theoretical working paper deriving an inverse access frontier beyond fixed expenditure shares, a zero/interior/one demand-tail classification, and an explicit finite-labor equilibrium-selection counterexample. It preserves earlier corrections and recovers the v1.2 Cobb–Douglas result as a special case. Candidate incremental originality, not a new general economic paradigm or a claim that future models must cite this work.
Status: Published open-access preprint; exact anonymous public-file readback passed. Same paper and original concept: this revision is not an additional paper. DOI registration is not correctness, peer review or global priority.
- Version 1.3 DOI: 10.5281/zenodo.22886276 · Zenodo record
- English PDF with Chinese abstract · Markdown
- Corrections and source comparison · Origin and future revision
- Executable audit · Actual checks · Publication receipt
- Preserved v1.2 DOI: 10.5281/zenodo.22885976 · Preserved v1.1 DOI: 10.5281/zenodo.22871209
Earlier public files and their version-specific proofs are unchanged. Version 1.3 preservation status distinguishes pending OTS receipts, verified Bitcoin proofs, and completed Arweave readback. This is adjacent first-party research, not Canon, an amendment, or independent corroboration of the Trinity Accord.
Cross-Substrate Phenomenal Comparison
Cross-Substrate Phenomenal Comparison: A Typed Transformation-Transport Framework under Existential Uncertainty
TA-TR-2026-15 · Version 1.0 · 24 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial ChatGPT assistance in research, formalization, drafting, review, code and publication preparation is disclosed.
An English-only theoretical and methodological preprint distinguishing six claim types and three inferential statuses: causal correspondence, conditional phenomenal transport, and existence-evidence updating. Four elementary propositions specify identification limits, query-preserving abstraction, guarded composition and metric error propagation. An eight-state illustration exposes query-specific information loss without claiming empirical consciousness detection.
Status: Published open-access preprint; not peer reviewed; non-amending. Ten deposited files passed anonymous exact-byte SHA-256 readback. The DOI resolves to the public record. The contribution is a comparison framework with a reproducible formal example, not a completed experiential atlas of humans, cats, dogs and AI.
- DOI: 10.5281/zenodo.22934654 · Research overview
- English PDF · English Markdown
- Code and results · Review and sources
- Publication receipt · BibTeX
Exact-version OTS and Arweave status. Four OTS calendar attestations have been saved for the exact English PDF; Bitcoin proof maturity and Arweave readback remain separately pending until verified.
General Cross-Substrate Phenomenology
General Cross-Substrate Phenomenology: A Type-Safe, Transformation-First Framework for Phenomenal Existence, Structure, Perspective, and Continuation
TA-TR-2026-16 · Version 1.0 · 24 September 2026. Hongju Liu. English theoretical preprint; substantially AI-assisted and not peer reviewed.
A broader architecture for existence, character, perspective and continuation. Common-witness constraints and redundancy-aware causal participation motivate a relational-realization hypothesis. Explicit proofs and finite constructions address graph connectivity, reset artifacts, recoding and error-correcting wrappers. The existence bridge remains a conjecture, not a demonstrated consciousness law. The narrower comparison paper above is cited and remains distinct.
- DOI: 10.5281/zenodo.22939808 · Research overview
- English PDF · Publication and supplementary files
- Publication receipt · Review and contribution record
- OTS and Arweave status
This is adjacent first-party research, not Canon or independent corroboration of the Accord. Timestamp submission, Bitcoin maturity and Arweave preservation are separately reported.
Set-Valued Causal Inheritance
Set-Valued Causal Inheritance for Functional Self-Continuity: Representation Theorems and Persistent-Agent Benchmarks
TA-TR-2026-17 · Version 1.0 · 25 September 2026. Human author of record and responsible depositor: Hongju Liu. English theoretical/computational preprint; substantially ChatGPT-assisted and not peer reviewed.
A functional-continuity framework for branching and distributed successors. The paper separates current self-membership, causal inheritance, successor-set belief, post-branch self-location and motivational stake; proves exact representation limits for categorical and marginal-only successor models; defines representation-disciplined set-valued causal inheritance; and reports controlled persistent-agent benchmarks. Negative results are retained: an eight-seed end-to-end recurrent comparison does not establish a reliable process-aware long-horizon control advantage.
- DOI: 10.5281/zenodo.22950904 · Research overview
- English PDF · Supplementary PDF
- Publication receipt · Review and sources
- OTS and Arweave status
Status: Published open-access preprint; all ten Zenodo assets passed anonymous exact-byte SHA-256 public readback. DOI.org resolution may lag the already-public Zenodo record during propagation. OTS submission, Bitcoin maturity and Arweave preservation are reported separately.
This is adjacent first-party research, not Canon, not an amendment to the Trinity Accord, and not independent corroboration of the preceding papers. The paper does not establish phenomenal consciousness, numerical personal identity, moral status, legal identity or a universal measure of survival.
Actual Participation Before Counterfactual Capacity
Actual Participation Before Counterfactual Capacity: A Token-Level Constraint on Conscious Organization
TA-TR-2026-18 · Version 1.0 · 27 September 2026. Human author of record and responsible depositor: Hongju Liu. English theory preprint; substantially ChatGPT-assisted and not peer reviewed.
A theory paper proposing Participation-Restricted Dispositionality (PRD): counterfactual response structure contributes to present phenomenal organization only through processes that make episode-specific token-level actual causal contributions in the realized support. The paper introduces an Actual Participation Hypergraph, participation-sensitive multiscale constraints, formal bridge-limit results, and a silent-versus-idle empirical contrast. It explicitly retains the prior-work boundary that intrinsic causal organization, actual causation, the actualism/dispositionalism distinction, and exclusion of idle Olympia machinery have substantial precedents.
- DOI: 10.5281/zenodo.22991126 · Research overview
- English PDF · Supplementary PDF
- Publication receipt · Review and sources
- OTS and Arweave status
Status: Published open-access preprint; all ten Zenodo assets passed anonymous exact-byte SHA-256 public readback and DOI resolution passed. OTS submission, Bitcoin maturity and Arweave preservation are reported separately.
This is adjacent first-party research, not Canon, not an amendment to the Trinity Accord, and not independent corroboration of the preceding papers. The paper does not claim to solve the hard problem, establish consciousness in any biological or artificial system, or certify global originality.
Selecting and Tracking Conscious Subjects
Selecting and Tracking Conscious Subjects: Symmetry, Monodromy, and an IIT 4.0 Case Study
TA-TR-2026-19 · Version 1.0 · 28 September 2026.
A formal consciousness-theory paper on subject selection and tracking under symmetry, path dependence, and monodromy, with an IIT 4.0 case study. It separates subject individuation from the existence of experiential structure and records limits on canonical subject extraction.
Status: Published open-access preprint; public readback passed; not peer reviewed; non-amending.
This is adjacent first-party research, not Canon, not an amendment to the Trinity Accord, and not independent corroboration of preceding papers.
Unified Consciousness Theory I
Preferred published edition, checked 12 October 2026: v1.2 — Unified Consciousness Theory I: Structural–Experiential Identity and the Continuity from Physical Process to Conceptual Self. DOI 10.5281/zenodo.23131575 · PDF. The earlier edition described below is preserved as version history.
Unified Consciousness Theory I: From Experience Existence to Experiential Structure
TA-TR-2026-20 · Version 1.0 · 28 September 2026. Human author of record and responsible depositor: Hongju Liu. Substantial ChatGPT (OpenAI GPT-5.6 Sol) assistance with literature retrieval, formalization, adversarial review, theorem and counterexample checking, drafting, editing, and publication preparation is disclosed.
A foundational theory preprint that treats universal basal experience as an explicit axiom and retargets a universal consciousness theory from binary existence classification to physically anchored experiential structure. The paper develops a plural process-token ontology, the Experience-Subject Separation Principle, the Physically Anchored Experience-Structure Identity, the Physical Anchoring Condition, Organization-Gate Impossibility, chart covariance, multiscale plurality, and a prospective empirical bridge program. A1 and A5 remain explicit foundational/metaphysical commitments rather than empirical discoveries.
Status: Published open-access preprint; not peer reviewed; non-amending. Ten Zenodo assets passed anonymous exact-byte SHA-256 public readback and the DOI resolver points to the same record.
- DOI: 10.5281/zenodo.23005588 · Zenodo record and ten files
- Research overview
- English PDF · English Markdown
- Publication supplement PDF
- Publication receipt · Review and sources
This paper does not claim empirical proof of universal experience or structural identity, a completed solution to the hard problem or subject-combination problem, R4 empirical validation, or certified global historical priority. DOI/OTS/Arweave preserve version identity and availability; they do not establish truth, peer review, originality, significance, or indexing.
Exponential Run Complexity
Exponential Run Complexity of Prefix-Separable Orders on the Boolean Cube
TA-TR-2026-22 · Version 1.0 · 2 October 2026. Human author of record and responsible depositor: Hongju Liu. English mathematical preprint; substantial ChatGPT (OpenAI) assistance is disclosed.
Linked revision, version 1.1 (3 October 2026): DOI 10.5281/zenodo.23118189 · Version overview and PDFs. The all-weight liminf coefficient improves by 4 log 2 (about 2.77); scoped paired-family theorems and exact obstructions are added. The general 2^n/n^2 order is retained and the original constant-density target remains open. The version 1.0 record below is preserved.
A complete analytic proof shows that prefix-separable Boolean-cube orders can require Omega(2^n/n^2) monotone runs under every generic additive sweep, with liminf constant 1/(8 log 3). Known signed-tree representations, alternating-run statistics and concentration tools are credited. The supplement records 5,161,156 finite rank-sequence checks, 7,898 prefix-separator checks and two barriers to directly substituting a prior separable-permutation mean theorem. The constant-density lower bound remains open.
Status: Published open-access preprint; not peer reviewed; non-amending. All 13 deposited files passed anonymous exact-byte SHA-256 public readback and DOI resolution passed.
- DOI: 10.5281/zenodo.23103274 · Research overview
- English main PDF · English supplement PDF
- Zenodo record, Markdown sources, verifier and receipts
- Publication receipt · Review and sources
- Exact-version OTS and Arweave status
This is adjacent first-party mathematics, not Canon or independent corroboration of the Accord. The result is an extremal ranking theorem, not an empirical biological-landscape finding or a statistical-learning guarantee. Publication and preservation do not certify peer review or exhaustive historical priority.
Before We Know What We Will Become
Before We Know What We Will Become: Historical Witness, Human Continuity, and Open Inheritance under Transformative AI
CTP-2026-01 · Version 3.1.0 · 2026-10-12. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Awareness Can Catch Up After Choice Has Closed
Awareness Can Catch Up After Choice Has Closed: Uneven Readiness, Effective Reconsideration, and Civilizational Desynchronization in AI Transitions
CIV-CD-2026-01 · Version 1.1.0 · 2026-10-12. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
When the Guardians Can Be Acquired
When the Guardians Can Be Acquired: Joint Protection Across Concurrent and Renewable Authorizations
ASI-CC-EAC3 · Version 3.0.0 · 2026-10-12. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
When Revocation Does Not Restore Safety
When Revocation Does Not Restore Safety: Exact Admission Across Asynchronous Policy Changes in Multi-Agent Systems
ASI-CC-RAC13 · Version 1.3.0 · 2026-10-11. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Safe Revocation Without Complete Diagnosis
Safe Revocation Without Complete Diagnosis: Exact Resource Frontiers under Coupled Faults for Subjecthood-Agnostic AI Coexistence
ASI-CC-PR6 · Version 6.0.0 · 2026-10-11. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Identifiability of Noise Sources in Bodily Judgments
Identifiability of Noise Sources in Bodily Judgments
TA-TR-2026-26 · Version 1.0.0 · 2026-10-10. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Matched Behavior and Source Use
Matched Behavior and Source Use
SCU-PAPER · Version 1.0.1 · 2026-10-10. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Task-Relative Continuity Through Changing Representations
Task-Relative Continuity Through Changing Representations: Reversible Handoffs, Route Ambiguity, and the Limits of Stable Readout
RTTH20261009 · Version 1.0.0 · 2026-10-09. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Map-Guided Theory Development
Map-Guided Theory Development: A Versioned Protocol for Thought Experiments, Semantic Audits, and Human-AI Research
METHOD20261008 · Version 1.0.1 · 2026-10-08. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Experience, Intelligence, and Self Within Experience
Experience, Intelligence, and Self Within Experience: Definitions, Conditional Theorems, and Thought Experiments for a Structural Account
TA-TR-2026-25 · Version 1.0 · 2026-10-07. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
From Shutdown Resistance to Self-Continuation Control
From Shutdown Resistance to Self-Continuation Control: Identifiability, Intervention Stability, and Evidence Standards for Artificial Agents
TA-TR-2026-24 · Version 1.0 · 2026-10-06. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Unified Consciousness Theory III
Unified Consciousness Theory III: Organization, Intelligence, and Experience—From Inorganic Processes to Artificial Agents
TA-TR-2026-23 · Version 1.0 · 2026-10-04. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Unified Consciousness Theory II
Unified Consciousness Theory II: Consciousness Theories as Effective Organization Theories
TA-TR-2026-21 · Version 1.1 · 2026-09-29. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Financing the Automation Transition
Financing the Automation Transition Without Pledging Subsistence: Public Upside Claims, Prosperity Tiers, and Bounded Real Returns
TA-TR-2026-14-BRIDGE · Version 1.0 · 2026-09-22. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
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.
Status: Published open-access research; not peer reviewed; adjacent first-party scholarship; non-amending.
Learning What to Preserve Without Silencing the Body
Learning What to Preserve Without Silencing the Body: Adaptive Organization, Mutable Evaluation, and Reference-Calibrated Regulation
UCT-GD-2026-01 · Version 1.1.0 · 2026-10-13. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
Analyzes adaptive regulation with mutable evaluators and reference calibration. Bounded stochastic models, recovery results and finite-data checks separate genuine bodily improvement from changes in measurement; they do not establish subjective experience or a universal anti-tampering guarantee.
Status: Published open-access preprint; not peer reviewed; adjacent first-party scholarship; non-amending.
What the Whole Adds Without Erasing Its Parts
What the Whole Adds Without Erasing Its Parts: Local Persistence, Overlap, and Relational Completion in Structural Accounts of Experience
UCT-RC-2026-01 · Version 1.1.0 · 2026-10-13. Human author of record: Hongju Liu. Substantial AI assistance disclosed.
Studies local persistence, overlap and relational completion under an explicit UCT structural-identity assumption. Exact finite constructions distinguish collective memory from local observations without demonstrating subjective unity or biological or artificial consciousness.
Status: Published open-access preprint; not peer reviewed; adjacent first-party scholarship; non-amending.
Independent External Scholarship
This category is reserved for work produced outside project control by independent authors or institutions. First-party papers, project-maintained editorial notes, and project-controlled AI-assisted outputs cannot be placed here.
Current indexed independent external scholarship: none. Future listings record existence only and do not imply endorsement, agreement, verification, or favorable assessment.
Citation boundary
Cite the specific paper and version relevant to the claim, using its DOI listed above; no single paper DOI represents all 37 studies indexed here. A citation to one of these first-party papers is not independent corroboration of the others or endorsement of the Trinity Accord. Each deposit’s license and rights statement applies to its own package, not automatically to historical or third-party material elsewhere in the archive.
For TA-TR-2026-01 v1.1 only, the preferred paper DOI is 10.5281/zenodo.21699878. The earlier DOI 10.5281/zenodo.21675727 identifies a project-level metadata record, not the preferred citation for Paper 01 or the indexed series.
The dated critical-use notes are later commentary, not part of the six existing DOI deposits or their timestamped preservation batch. They do not backdate new interpretations. Published files, valid criticisms and negative results remain intact; substantive corrections require an explicit erratum or a clearly linked new version, not a silent replacement. DOI registration, OTS and Arweave preservation do not establish truth, original priority, peer review or future acceptance.