PUBLIC RESEARCH BRANCH

Inscriptions, archives, and broken clocks

Cuneiform, Maya, khipu, database records, physical carriers, archive formation, recovery, decipherment, and chronology distortion.
brief8a0683f025a864056f426224f51fdb1dcf2c96d2db9fc7c0b63fb3981c6432d6
provenancec75334217a73d23b0fdd37ec3335eabe6ec2b71a5e89f2548fd1ca5f5e565c02

TimeWar Inscription, Carrier, and Archive-Clock Proposition Audit

Date: 2026-07-27

Status: Bridge 66 complete at proposition-audit level; expansive goal remains active

Result

The strongest proposition surviving this twelve-pair audit is not that the ancient record tells one transparent story. It is that a record can cross a rupture while its reader, authority, grammar, physical carrier, archive, and clock separate. Recovery then works by joining partial carriers again.

That pattern is visible at very different scales:

  • a cuneiform composition can survive in later copies while its original support and first performance are lost;
  • a catalog row can preserve modern knowledge of an artifact without proving one independent physical witness;
  • a Maya monument can preserve signs after the institution that trained fluent composers has fractured;
  • a khipu can preserve knots, arithmetic, and local custody while the wider comparison and authentication network disappears;
  • an inscription date range can cross the event it is supposed to classify unless the event, artifact, use, deposition, recovery, edition, and database clocks are separated.

This is strong evidence for ordinary civilizational inheritance through rupture. It leaves found civilization, breakaway continuity, nonhuman intervention, and literal cataclysm-linked departure explicit but unpromoted at F0/K0/N0/B0.

Frozen review frame

Bridge 61 froze 119 active non-self research pairs at commit b369f27. Bridges 62–65 reviewed four disjoint twelve-pair cohorts. Bridge 66 selected a fifth disjoint cohort:

  • 12 pairs;
  • 24 files;
  • 212,415 bytes read by each reviewer;
  • two independent full-file readers;
  • zero overlap with Bridges 62–65;
  • frozen residual after this tranche: 119 - 60 = 59.

The two short output wrappers contribute no independent evidence. They collapse into their canonical inscription-ontology and event-window pairs.

Do not count unlike things as one denominator

CorpusUnits actually presentUnits that cannot be silently substituted
EDRdatabase record, edition, text execution, fragment, support, monumentone record ≠ one inscription ≠ one monument ≠ one independent name-bearing opportunity
CDLIcatalog row, artifact-candidate row, fragment, envelope, seal, cast, composite relationone row or P identifier ≠ one proven independent tablet
ORACCP catalog/object reference, Q composite composition, transliteration, lemma nodeone P/Q/project appearance ≠ one new ancient witness
Ur III trialcandidate catalog document, lexical node, sanitized span, exact-sequence componentrepeated sequence ≠ copying, archive membership, or physical identity
khipuobject, main cord, pendant, subsidiary, knot group, local grammar, catalog recordcatalog cord count ≠ excavation inventory; one local key ≠ universal decoder
Mayacodex, monument, inscription, sign inventory, language, colonial manuscript, edition, readingfour codices ≠ four independent confirmations of every script-wide claim

CDLI’s current API exposes artifacts, inscriptions, publications, composites, seals, external resources, and other relations through distinct interfaces. That architecture supports typed joins rather than row inflation (CDLI REST API). ORACC likewise distinguishes P-numbered catalog objects from Q-numbered composites organized around documentary rather than physical structure (ORACC project catalogues, ATF composites).

The clocks that must stay separate

  1. event or narrated-event clock;
  2. initial composition clock;
  3. recension or copying clock;
  4. artifact manufacture or inscription-execution clock;
  5. use and reuse clock;
  6. deposition or archive-placement clock;
  7. excavation, acquisition, or recovery clock;
  8. physical joining and restoration clock;
  9. edition, translation, and decipherment clock;
  10. digitization, catalog revision, and current-claim clock.

An AD 79 date interval that overlaps the Vesuvius boundary is not automatically post-eruption. A later copy of a king list dates the witness, not every reign it narrates. A modern khipu database date describes catalog custody, not cord manufacture. A modern Maya reading can reactivate an ancient sign without becoming an ancient witness.

Pair dispositions

PairSurviving resultPresent ceiling
inscription ontologysix-level record/edition/execution/fragment/support/monument distinction is necessarystructure audit complete; real lineage joins remain unrun
event-window auditoverlap versus strict containment demonstrates boundary sensitivityinterval-censored outcome model remains unrun
cuneiform denominatorJuly 26 pulls repeat at 423,262 rows; declared physical-candidate layer is 417,700snapshot-bound catalog result, not timeless artifact universe
Ur III feasibility18,358 exactly dated five-site candidate catalog documents form a technically usable framehistorical efficacy comparison remains unauthorized
fifteen-sites claimconcentration is real; the combined 200,000/fifteen-sites/since-1849 formula is unsupportedPhase 2 supersedes the initial matrix status
khipucarrier, numerical syntax, local grammars, custody, and authority survive unevenlyT1–T2; F0/K0/N0/B0; no universal decoder
Mayapeople, language, material signs, genres, colonial texts, and scholarship recombine a damaged grammarT1–T3; F0/K0/N0/B0; terminal scribal interval remains open
inscription wrapperroute to canonical pairno independent witness
event-window wrapperroute to canonical pairno independent witness
Ur III dependency layerexact sanitized-equivalence and catalog sensitivities are reproduciblerecurrence is not copying, archive, or physical identity
Ur III information/join rulesfive-token and 31.650092-bit review gate plus typed joins are frozendescriptive threshold, not a p-value or result
Ur III Stage T880,163 lexical nodes yield 771,375 label-safe spanstoken engineering complete; outcome comparison sealed

The cuneiform claim in plain English

The corpus can tell extremely clear histories when a bounded archive, date range, genre, institution, and physical lineage are dense. It cannot yet be compressed into “the clear history of the world.” The reviewed data instead show:

  • heavy concentration in a small number of sites and periods;
  • strong administrative overrepresentation;
  • missing or uneven archive, stratigraphic, genre, and provenience fields;
  • compositions copied across multiple physical witnesses;
  • fragments, envelopes, joins, casts, editions, and project mirrors;
  • recovery and publication bias.

The viral formulation about more than 200,000 tablets from fifteen locations since 1849 remains a useful routing lead, not a supported compound denominator. The Phase-2 matrix is the controlling quantitative result.

Why the Ur III comparison remains sealed

Stage T successfully removes ruler, year, site, window, publication, archive label, source IDs, names, and numbers from the safe text layer. That is a major leakage control. It is not enough to run the terminal comparison.

Before outcomes may be joined, the project still needs:

  • physical fragment and join adjudication;
  • tablet/envelope and cast/original relations;
  • archaeological archive and subarchive roots;
  • dossier, formula, scribe, seal, and official networks;
  • grouped holdouts that keep related records together;
  • external event controls and blind proposition coding.

An exact repeated phrase may be ordinary bureaucracy, a genre formula, or an artifact of masking. The frozen 31.650092-bit gate is a review threshold, not statistical proof.

Two carrier controls

The Open Khipu Repository’s versioned 2.1.0 release added KH0631–KH0702 and describes itself as an expanding archive of archaeological and museum khipu data. Growth in catalog coverage does not create a universal decipherment key (Open Khipu Repository v2.1.0).

Harvard’s Corpus of Maya Hieroglyphic Inscriptions has operated since 1968, published nineteen fascicles, standardized recording, and built a field archive instrumental in decipherment. That is an unusually strong modern recovery carrier, while still being a modern archive rather than an independent ancient witness (Peabody Museum CMHI).

Together the controls show that carrier survival is layered. Objects, syntax, language, genres, readers, authority, and comparison archives can persist on different clocks.

TimeWar synthesis

The important fringe-forward possibility is not lost by enforcing these distinctions. It becomes testable.

  • Ordinary transformation (T) asks whether people, practices, objects, archives, or institutions cross the rupture in altered form. Khipu and Maya provide strong positive cases.
  • Found civilization (F) asks whether a securely older productive phase and technical ecology underlie later reuse or ownership claims. No reviewed pair closes that chain.
  • Breakaway civilization (K) asks whether a bounded population and self-sustaining material/logistical system continue outside ordinary visibility. Restricted custody is not enough.
  • Nonhuman intervention (N) requires authenticated nonhuman agency or discriminating technical residue. Undeciphered signs are not enough.
  • Literal departure (B) requires a bounded source population, mechanism, positive destination-side trace, identity continuity, and independent replication. Missing readers or destroyed archives are not enough.

This tranche therefore strengthens the TimeWar thesis without laundering mystery into proof: civilizations can disappear as operating systems while remaining present as scattered carriers. The stranger hypotheses remain live at their own gates.

Falsifiers and promotion tests

The carrier-fragmentation model weakens if secure cases show that full competence and institutional reproduction continued without interruption despite the alleged rupture.

The strong found-civilization thesis gains support only if independently dated older manufacture, a prior productive ecology, later reuse, successor non-reproduction, and matched technical residues close on the same object family.

The Ur III content hypothesis gains support only after the sealed comparison survives archive-aware dependency groups, blind coding, external controls, and held-out replication.

Literal departure gains support only from positive transition and destination evidence. Silence, catalog gaps, damaged texts, mortality, or disappeared institutions cannot promote it.

Disposition

Bridge 66 closes this exact twelve-pair proposition tranche. It does not close the expansive goal, the remaining 59-pair frozen review residual, the Ur III outcome experiment, the Vesuvian held-out study, or the broader public-integration queue.

Sources, provenance, and custody

Provenance — Bridge 66 Inscription, Carrier, and Archive-Clock Audit

Date: 2026-07-27

Status: adjacent provenance for the cited Bridge 66 synthesis

Frozen denominator

  • Baseline commit: b369f27.
  • Pair register: notes/research/civilizational-technology/timewar-research-pair-register-v1.json.
  • Frozen register: 370,581 bytes; SHA-256 2ea0f086d640ad7e065bf2b483affa1a4e0b150278f71ee1027a7ad2d39644d1.
  • Frozen census: 201 total pairs, 119 active non-self pairs, zero missing companions.
  • Prior disjoint cohorts: Bridges 62–65, 48 pairs.
  • Bridge 66 cohort: 12 pairs / 24 files / 212,415 bytes per reader.
  • Residual: 119 - 48 - 12 = 59.

The exact cohort, file hashes, sizes, and required questions are frozen in outputs/.plans/timewar-inscription-carrier-and-archive-clock-proposition-audit.md at plan commit 0e5ce7a.

Independent review custody

Two independent readers each verified all 24 current files against the frozen SHA-256/byte table and read all 212,415 bytes. Neither edited project files or coordinated with the other reader. Their raw reports are session-custodied; this sidecar records the reconciled facts without pretending the reports are independent external historical evidence.

Both readers independently identified the same required repairs:

  • replace “physical P identifier” with artifact-catalog P identifier;
  • mark the first fifteen-site page as superseded by Phase 2;
  • retire the old additive A ladder on Khipu and Maya pages;
  • temper “complete” Maya recovery language;
  • bind output wrappers to canonical bytes;
  • keep the Ur III trial sealed;
  • collapse wrappers and database/project mirrors before convergence.

Source-family handling

The audit collapses rather than multiplies:

  • the two output wrappers into their canonical research pairs;
  • CDLI, ORACC, and BDTNS appearances when they descend from the same catalog object or digital lineage;
  • the three Ur III stages as sequential engineering over one 18,358-record frame;
  • publications, supplements, catalogs, and reanalyses that reuse one Khipu excavation or object group;
  • modern drawings, editions, databases, and analyses of one Maya monument into the same ancient physical root.

Current primary checks

Checked 2026-07-27:

These current checks confirm entity distinctions and living custody. They do not replace the frozen quantitative snapshots or constitute independent ancient witnesses.

Inference limits

  • The July 26 CDLI total is snapshot-bound, not a timeless corpus size.
  • “Physical-candidate catalog row” is not shortened to “tablet” or “independent artifact.”
  • The 31.650092-bit Ur III rule is descriptive, not a p-value.
  • Khipu and Maya demonstrate ordinary inheritance; neither is a vanished-population case.
  • Found, breakaway, nonhuman, and literal-departure lanes remain F0/K0/N0/B0.
  • A public Evidence card is compact claim metadata, not a third essay.

Reproducibility

Machine ledger: notes/research/civilizational-technology/inscription-carrier-and-archive-clock-proposition-audit-v1.json

Builder: scripts/build-inscription-carrier-and-archive-clock-proposition-audit.py

Tests: scripts/test-build-inscription-carrier-and-archive-clock-proposition-audit.py

The builder authenticates the frozen Git blobs, rejects overlap with Bridges 62–65, binds repaired current files, freezes the unit/clock/dependency distinctions, and refuses extraordinary-register or experiment-result promotion.