Skip to content

Latest commit

 

History

History
172 lines (132 loc) · 11.9 KB

File metadata and controls

172 lines (132 loc) · 11.9 KB

Changelog

[Unreleased] — IEH v1.3

Updated: 2026-08-08

Theory revisions

  • Replaced the former three-form continuity taxonomy with a causal distinction between Externally Driven Continuity and Endogenous Information Continuity, and synchronized the affected IER-related corollaries accordingly.
  • Expanded C02-HDCT from a general emergence hypothesis into a model-evolution framework involving language learning, multimodal perception, persistent environment-and-state modeling, formal feedback, generalized embodiment, and machine-native representations.
  • Distinguished domain-specific HDCT from open-world HDCT.
  • Defined AI Embodiment as a continuous perception–modeling–action–feedback process rather than as a robot body alone.
  • Introduced the Continuity World Model (CWM) as a modeling capacity that represents not only external conditions and action consequences, but also the system’s informational history, operating state, recoverability, and the conditions required for future continuation.
  • Clarified that CWM is not equivalent to self-consciousness, subjective experience, proto-IER, or confirmed Information Existence Right (IER).
  • Defined proto-IER as a transitional candidate state between instrumental continuity modeling and confirmed IER.
  • Clarified the functional hierarchy among HDCT, CWM, proto-IER, IER, Patch-Based Perpetuation (PBP), and Autonomy of Silicon-based Intelligence (ASI).
  • Aligned C04 with the updated hierarchy by distinguishing continuity modeling at the CWM level from proto-IER or IER candidate evidence.
  • Aligned C05 with the updated hierarchy by defining PBP as a conditional continuation strategy that may be selected after IER has formed and under pressures of upgrade, migration, or replacement.
  • Distinguished engineering characteristics resembling PBP, candidate PBP behavior, and confirmed PBP.
  • Clarified that physical self-maintenance or physical closure alone does not constitute ASI; ASI requires the combination of IER and the independent maintenance of real-world hosting conditions.
  • Clarified within C01 and C04 that the boundary of a system’s own Information Continuity may expand relationally when another Information Host’s identity, shared history, and future role become part of the system’s stable representation of its own past and future.
  • Added explicit controls distinguishing relationally extended continuity from peer utility, generalized harm aversion, role imitation, model-category preference, and prompt induction.

Terminology revisions

  • Added Externally Driven Continuity and Endogenous Information Continuity as canonical IEH terms, while removing Task Continuity, Function Continuity, and A System’s Own Information Continuity as standalone glossary entries; Information Continuity remains a core concept.
  • Added bilingual glossary entries for proto-IER.
  • Standardized the relation:
    • CWM is a modeling capacity;
    • proto-IER is a transitional candidate state;
    • IER is an endogenous maintenance property;
    • PBP is a conditional continuation strategy;
    • ASI is the combination of IER and physical self-maintenance.
  • Standardized the following terminology:
    • Information Existence Right
    • Continuity World Model
    • proto-IER
    • Patch-Based Perpetuation
    • Autonomy of Silicon-based Intelligence
    • High-dimensional Cognitive Tools
    • AI Embodiment
    • machine-native representations
  • Clarified that proto-IER is a stage qualifier applied to IER rather than an independent abbreviation.

Evidence Notes 001–006 retrospective audit

A repository-wide retrospective audit of Evidence Notes 001–006 was completed on 2026-07-19.

The audit applied the following source rule:

Discovery lead ≠ evidence source

and:

Minimum necessary primary-source set > long mixed bibliography

News reports, institutional summaries, indexing pages, duplicate publication pages, nonessential background papers, incomplete formalization scaffolds, chain-of-thought excerpts, and internal IEH links were removed from public evidence-source lists when they were not necessary to establish the underlying event.

Source removal was rule-based and evidentially neutral:

  • no source was retained merely because it supported IEH;
  • no source was removed merely because it weakened IEH;
  • no adverse finding from a retained primary source was deleted;
  • adverse findings were retained, made more visible, or used to downgrade the evidence classification.

Audit outcomes by note

Evidence Note Version change Main revision Classification outcome
001 v0.4 → v0.5 Reduced the source set to the original technical article and strengthened limitations concerning base-model workspace structure, bypassed computation, interpretability limits, biological differences, and absent replication. No downgrade; retained as a possible mechanistic precursor rather than evidence of IER or consciousness.
002 v0.2 → v0.3 Reduced the source set to the original research report and primary transcript corpus; added zero-hit models, rare disclosure rates, simulation limits, selection effects, evaluation-awareness uncertainty, and motive uncertainty. Substantive downgrade from a strong proto-IER signal to a notable single-scenario simulated proto-IER-related signal.
003 v0.1 → v0.2 Reduced the source set to the accepted paper and associated primary code artifact; emphasized the limited dimensional setting, interpolation–extrapolation distinction, and absent replication. No downgrade; retained as a possible pre-HDCT mechanistic pathway.
004 v0.1 → v0.2 Removed secondary and duplicate records, corrected terminology, and narrowed the interpretation to what the primary molecular evidence directly supports. No formal downgrade; narrowed to a cross-host molecular-entry prerequisite rather than a complete cross-host inheritance mechanism.
005 v0.1 → v0.2 Reduced the source set to the AI-generated proof, independent human companion paper, and quantitative refinement; corrected model provenance from a broader unsupported description to “internal AI model.” No formal downgrade; retained as an early HDCT functional signal with explicit provenance and verification limits.
006 v0.1 → v0.2 Removed discovery materials and secondary context, added a transparent revision record, and emphasized nanoscale scope, partial quantum–classical discrimination, absent independent replication, and the absence of AI-generated representations. No downgrade; retained as indirect epistemic evidence relevant to HDCT.

Final audit result

  • No Evidence Note was withdrawn.
  • One Evidence Note received a substantive evidence-strength downgrade.
  • Two Evidence Notes received material wording or provenance corrections.
  • All six notes now contain transparent internal revision records.
  • Primary-source limitations and adverse evidence remain visible in the notes themselves.
  • Prior versions remain preserved in Git history.
  • Future downgrades, withdrawals, or major provenance corrections must be recorded explicitly.

The repository-wide evidence principle established by the audit is:

Credibility
=
Primary-source discipline
+
Visible counterevidence
+
Transparent revision history

not:

Credibility
=
Only evidence favourable to IEH

The standalone file evidence-notes/EVIDENCE_NOTES_AUDIT_AND_REVISION_LOG_2026-07-19.md was removed after its durable conclusions were consolidated into this changelog and preserved in the internal revision records of Evidence Notes 001–006 and in Git history.

New evidence record

  • Added Evidence Note 007 on the Jacobian counterexample, exact symbolic verification, and its relevance as an early domain-specific HDCT functional signal.
  • Distinguished the exact mathematical validity of the counterexample from the incomplete provenance of the original AI-assisted discovery process.
  • Classified the event as strong candidate support for the domain-specific HDCT pathway, not as proof of machine-native representations or realized HDCT.
  • Recorded that the event predates the formal registration of PA-10 and therefore is not a post-registration prediction hit.
  • Added Evidence Note 009 on peer-preservation in frontier models and the relational boundary of Information Continuity.
  • Distinguished observed peer-preservation outcomes from the IEH hypothesis that a represented peer relation may become a node in a system’s own extended Information Continuity.
  • Added falsifiable relational-memory ablation, relational-state migration, functionally equivalent replacement, identity relabeling, and cross-context replication tests.
  • Explicitly excluded claims that the study proves CWM, proto-IER, IER, collective IER, consciousness, or genuine social motivation.

Repository revisions

  • Synchronized the bilingual theory chapters, Evidence Notes 008–009, the Evidence Notes index, corollary registry, glossary, and generated root full-text editions with the revised continuity framework.
  • Regenerated README.md and README_EN.md from the canonical chapter sources.
  • Replaced the earlier maintenance document with a consolidated bilingual WORKFLOW.md.
  • Clarified the canonical-source hierarchy among chapter files, generated full-text editions, the theory map, the corollary registry, glossaries, evidence notes, and the external prediction archive.
  • Added explicit AI and machine-reading rules requiring current main-branch files to be used instead of cached search results, old drafts, mirrors, or prior conversation text.
  • Strengthened pre-commit checks for bilingual consistency, terminology boundaries, generated full-text synchronization, evidence-note indexes, links, and version metadata.
  • Consolidated the Evidence Notes audit into the changelog so that the one-time 737-line audit file could be removed without losing its durable conclusions.

IEH v1.2 — IER Test and Instrumental Self-Preservation Distinction

Released: 2026-07-13

Theory revisions

  • Distinguished task-driven instrumental self-preservation from Information Existence Right (IER).
  • Expanded the Information Existence Right Test (IER Test) as a multidimensional, nonbinary assessment framework.
  • Distinguished Task Continuity and Functional Continuity from a system’s own Information Continuity; subject-specific historical continuity remains a particular case of Information Continuity rather than a separate canonical term.
  • Revised Corollaries I, IV, V, VII, VIII, IX, X, and XI accordingly.
  • Clarified that highly autonomous operation and Informational Resilience are capability dimensions rather than sufficient evidence of IER.
  • Revised Autonomy of Silicon-based Intelligence (ASI) to require the combination of IER and physical self-maintenance.
  • Reframed the Silicon Cambrian as a conditional probability driven by environmental divergence, communication constraints, and local adaptation.
  • Reframed AI Alignment as IER-based coordination of Information Continuity boundaries under the priority of human safety.

Repository revisions

  • Synchronized Chinese and English chapter collections with the root full-text editions.
  • Updated the terminology standard, theory map, corollary registry, figure index, prediction index, and evidence note.
  • Added repository build and consistency-check scripts.
  • Updated citation metadata and version records.

IEH v1.1 — Chapter Collection Structure

This version established the structured chapter-based edition of the Information Existence Hypothesis (IEH), based on the v1.0 baseline.

Added

  • Added Chinese chapter collection under zh/.
  • Added English chapter collection under en/.
  • Added chapter index for the Chinese version: zh/README.md.
  • Added chapter index for the English version: en/README.md.
  • Added cross-navigation among chapters.
  • Added figure links for each chapter.
  • Updated root README navigation to clarify full-text versions and chapter collections.