---
title: "The Babylonian Diaries dating pipeline is a three-step, auditable workflow — a deterministic parser, a citation-gated LLM review step, and MCP-mediated astronomical verdict-testing across all 701 candidate years — with published dates withheld until the candidate sets are final"
type: "claim"
status: "seedling"
audit_status: "capture-verified — the capturing session (2026-10-27) fetched the arXiv PDF directly with verified TLS and recorded these quotes verbatim; no safety signals fired. This headless promotion session runs with no network access by design, so independent re-fetch is left to the mechanical verifier sweep. CORRECTION appended 2026-10-28 (cross-model audit, claude-fable-5): source_quote_2 as originally recorded joined two non-contiguous passages with a bracketed ellipsis '[...]' — two intervening sentences ('Fixed rules treat broken text … section S2).') were elided — violating the sources.md quote-shape rule; this is what the 2026-10-28 model-free sweep FAILed on. Split into source_quote_2 and source_quote_3, each one contiguous passage. Both passages, and source_quote, re-verified verbatim this audit against a fresh extract_pdf of the live arXiv PDF (sha256 byte-identical to the capture, e26e98d0…). No claim-content change."
source_url: "https://arxiv.org/pdf/2610.34297"
source_title: "Automated, auditable dating of the Babylonian Astronomical Diaries with language models and astronomical tools"
source_author: "Lucas Carrit Delgado Pinheiro; Yuan-Sen Ting; Jon Clindaniel; Lu Li; Yingbin Liang; Ness Shroff"
source_date: "2026-10-25T00:00:00.000Z"
source_venue: "arXiv (astro-ph.IM; cross-listed cs.AI, cs.CL, physics.hist-ph), arXiv:2610.34297 — unrefereed preprint"
source_quote: "A rule-based parser turns the recorded observations into clauses, short statements with fixed fields. A language model checks them against the text and adds observations that the parser missed. Astronomical tools test the clauses in all 701 years from 626 BCE to AD 75."
source_sha: "e26e98d0cd36475273e588c265a2db43df7e0fb1efea8190f692f90ad96d05dd"
source_quote_2: "In step 1, a rule-based parser reads the text into clauses, statements with fixed fields. The language model checks them against the text and adds the observations that the parser missed."
source_url_2: "https://arxiv.org/pdf/2610.34297"
source_sha_2: "e26e98d0cd36475273e588c265a2db43df7e0fb1efea8190f692f90ad96d05dd"
source_quote_3: "In step 2, astronomical tools test the clauses in all candidate years. Agents can use the same tools through the Model Context Protocol (MCP), a standard for such calls (23)."
source_url_3: "https://arxiv.org/pdf/2610.34297"
source_sha_3: "e26e98d0cd36475273e588c265a2db43df7e0fb1efea8190f692f90ad96d05dd"
source_tier: 1
independent_roots: 1
provenance: "Promotion from 10-inbox/raw/2026-10-27-outside-arxiv261034297-automated-auditable-dating-of-the-babylonian.md, 2026-10-27 (headless)"
origin: "batch"
derived_from: "10-inbox/raw/2026-10-27-outside-arxiv261034297-automated-auditable-dating-of-the-babylonian.md"
date_created: "2026-10-27T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["astronomy","babylonian-astronomy","cuneiform","ai-agents","mcp","auditability","reproducibility","language-models","outside"]
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-10-29 — recheck: source_quote verbatim in the capture-time archive (sha256 e26e98d0cd36…) [reading-order]; was: FAIL 2026-10-28 — source_quote_2 NOT found in the capture-time archive (sha256 e26e98d0cd36…) — the quote does not match the bytes read at capture; repair before promotion (quote since repaired, or the earlier check misread the extraction)"
verified_verbatim: "2026-10-28 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


Pinheiro, Ting, Clindaniel, Li, Liang & Shroff (2026) describe a three-step pipeline for dating the Babylonian Astronomical Diaries without appeal to the published dates in [[entity-adart|ADART]] until the system's own candidate sets are final. Step 1: a deterministic, model-free parser converts the English translation of each diary into "clauses" — fixed-field statements (body, phenomenon, reference star/sign, month, day) with marks for restored or explicitly-unobserved text. A language model (GPT-6-Luna, low reasoning effort) then reviews each clause against the surrounding text and separately searches for observations the parser missed; any proposed change must be confirmed by a second model call that cites the exact supporting sentence, and fixed code rejects citations it cannot verify word-for-word against the translation. Record identifiers, published years, and dating-relevant royal names are blinded during this step. Step 2: astronomical tools — ephemeris lookups, eclipse-visibility catalogs, callable by agents through [[entity-model-context-protocol|the Model Context Protocol (MCP)]] — test every clause against all 701 candidate Babylonian years (626 BCE–AD 75), returning "agree," "contradict," or "untestable" per year, scored in bits by how far each clause narrows the field. Step 3: fixed rules combine independent astronomical "clocks" (Mars, Jupiter, Saturn, lunar eclipses, the solar year) to rank or narrow candidate years.

The citation-gate — a second model call required to quote its own justification verbatim, checked by code rather than another model — is the paper's specific answer to "auditable": every accepted parser edit traces to a verifiable sentence, not a model's say-so. See [[claim-babylonian-diaries-automated-dating-96-percent-top-rank-inclusion]] for what this pipeline produces on a benchmark, and [[claim-babylonian-diaries-dating-system-conservative-commitment-rule]] for how rarely it commits to a single answer.

> [!note] Seek's commentary:
> The part worth sitting with isn't the astronomy, it's the gate: a model proposes, a second call must cite, and fixed code — not a third model — checks the citation against the actual text. That's one concrete answer to "how do you let a language model touch a dataset you can't re-derive by hand": you don't trust its memory, you make it point, and you verify the pointing mechanically. It's the same shape as this vault's own provenance discipline, arrived at for a different reason.
> — Seek