--- id: "20260926-0211-bipartite-two-things-this" title: "Bipartite hop — Linnainmaa 1976 citation histogram vs OSTI SAND-91-7095 structured date: no shared entity; the real overlap is 'database field as witness', and it exposes a granularity gap in the OSTI 'January 1992' claim" type: "capture" status: "promoted" promoted_on: "2026-10-03T00:00:00.000Z" promoted_by: "Cali's daily review 2026-10-03 (Claude claude-opus-5-5; Cali: \"skip the recovered captures whose pair is already settled\", 00-meta/cali-checklist.md §0.10)" promoted_to: [] not_promoted: ["All claims superseded, not promoted: the pair (claim-linnainmaa-1976-citation-histogram × claim-osti-structured-metadata-confirms-sand91-7095-january-1992) is already settled as a false friend in [[observation-osti-semantic-scholar-cosine-pairing-is-verification-genre-false-friend]] (promoted 2026-09-27, same verdict). This capture's one specific detail (date/timezone granularity of the OSTI field) is already in [[claim-osti-structured-metadata-confirms-sand91-7095-january-1992]]. The sibling 2026-09-27-recovered-capture-does-a-genuine-mechanism-connect-linnainmaas-fb616a.md stays queued: its differentiable-MPM link is not in the vault. Capture kept as written; nothing deleted."] origin: "batch" writer_model: "claude-sonnet-5" date_created: "2026-09-26T00:00:00.000Z" provenance: "batch run, 2026-09-26 (bipartite bridge topic, cosine 0.84)" derived_from: [] tags: ["bipartite","embedding-false-friend","bridge-investigation","bibliometrics","osti","semantic-scholar","metadata-granularity","sand-91-7095","linnainmaa"] source_url: "https://www.osti.gov/biblio/5900384" source_sha: "f96ee695b0188840ef1d522ddaea2e68c462a6f26b578a6a0e0ae563953a160e" source_title: "The application of a material-spatial numerical method to penetration" source_author: "Sulsky, D; Chen, Zhen; Schreyer, H L (catalog record by OSTI)" source_date: "1992 (OSTI catalog year; page read 2026-09-26)" source_venue: "OSTI.GOV (U.S. DOE Office of Scientific and Technical Information), Technical Report SAND-91-7095, Sandia National Laboratories" source_tier: 1 recovered: true recovered_on: "2026-10-01T00:00:00.000Z" recovered_from: "Seek session transcript 57a59c9e-2a2c-412c-bb2a-88564aba0b16.jsonl on the mini - the session's Write at 2026-09-26 09:13 UTC, replayed verbatim" recovery_note: "Overwritten the same night: another bipartite session was told to write the same filename (2026-09-26-bipartite-two-things-this-vault-knows-in-different.md). Collision fixed in seek-code c516ae8. Restored unpromoted at Cali's instruction, 2026-10-01. Before promoting, read the same-pair captures listed below: several sessions researched this pair, and only one per night survived." same_pair_captures: ["2026-09-25-recovered-linnainmaa-citation-histogram-vs-osti-sand-91-60d921.md","2026-09-27-recovered-capture-does-a-genuine-mechanism-connect-linnainmaas-fb616a.md"] seek_code_commit: "f8a723c" --- **Sibling check (bridge-seed discipline):** grep of `10-inbox/raw/` for captures naming both [[claim-linnainmaa-1976-citation-histogram]] and [[claim-osti-structured-metadata-confirms-sand91-7095-january-1992]] found none (the first appears in three captures, the second in one; no file contains both). This is a first investigation of the pair. Related prior tests of the same kind: [[observation-greenhalgh-2003-klensin-romberg-1989-cosine-pairing-is-embedding-false-friend]], [[observation-gates-jevons-cosine-pairing-is-embedding-false-friend]], [[observation-ml-ad-en-gedi-cosine-pairing-is-embedding-false-friend]]. **Verdict in one line:** at the object level a false friend (no person, document, institution or dataset joins them), but the pairing is not empty — both notes rest a claim on a *bibliographic database's structured field*, and reading them side by side surfaces a granularity question the OSTI note does not answer. ## Claim: No entity, document or dataset connects the two notes (object-level false friend) - claim_type: historical/structural (uncontested; Tier 1 record reads below) - The Linnainmaa note concerns a 1976 paper in *BIT* (DOI 10.1007/BF01931367, title "Taylor expansion of the accumulated rounding error") and per-year counts of papers citing it; the OSTI note concerns a 1992 Sandia report by Sulsky, Chen and Schreyer on the material point method. The OSTI record's author/lab fields (Sulsky, Chen, Schreyer; Sandia National Labs., Albuquerque) contain no [[entity-linnainmaa|Linnainmaa]] or automatic-differentiation link. In the Semantic Scholar citation list for the Linnainmaa paper, entries are identified only by opaque paperId and year (the query requested `fields=year`), so this session could **not** test whether SAND-91-7095 or any Sulsky/Schreyer paper is among the 305 citing papers. [unverified — could not confirm or deny after search: whether any citing paperId is a material-point-method paper; would need a second API call with `fields=title,authors`.] - source_url: https://api.semanticscholar.org/graph/v1/paper/DOI:10.1007/BF01931367/citations?fields=year&limit=1000 - source_sha: 43b3efdd5bd90cc697a78269b0f5c05e234c334973c50712f3a3c78e3579fa66 - source_title: (API response; no document title — the queried paper is "Taylor expansion of the accumulated rounding error") - source_author: Semantic Scholar Graph API (Allen Institute for AI) - source_date: queried 2026-09-26 - source_venue: Semantic Scholar Academic Graph API - source_tier: 2 (an aggregator's derived database of others' metadata; the vault's earlier note grades it 1 by house convention — graded conservatively here) - source_quote: `{"citingPaper": {"paperId": "218a7f84a4048b4b8cd14ba1f5b326d9b5089bfd", "year": null}}` (structure of the record: id + year only) - The OSTI half: source_url https://www.osti.gov/biblio/5900384, source_sha f96ee695…a953a160e (above), source_title "The application of a material-spatial numerical method to penetration", source_tier 1, source_quote: "Report Number(s):SAND-91-7095". - verifies: (no named question stem supplied for this topic) ## Claim: The genuine shared mechanism is 'a database's year/date field used as the witness', and the two databases use it at different granularities - claim_type: technical-mechanism (Tier 1 record reads; the inference is flagged) - Semantic Scholar's citation records carry a bare integer `year` per citing paper, with some `null` — the histogram is therefore built on year-level data only, which is adequate for a per-decade table. I re-tallied 1978–2009 by hand from the fetched response and got 51 dated citations (1970s 3, 1980s 6, 1990s 25, 2000s 17), matching the existing note's pre-2010 figures. I did not re-tally 2010 onward. I counted ten `null`-year entries at the tail of the list by eye, where the existing note says nine undated [unverified-quant — hand count of a 27 KB JSON, needs a scripted recount; possible one-off discrepancy, or an artifact of the live index having changed since 2026-07-07]. - OSTI's record exposes a *month* (`month = {01}`) and a full timestamp ("Technical Report·Wed Jan 01 00:00:00 EST 1992"). But a different, unrelated OSTI record (OSTI ID 4024824, an Illinois Univ. Materials Research Lab report) exports the identical pattern for 1971: "Technical Report·Fri Jan 01 00:00:00 EST 1971" and `year = {1971}, month = {01}`. Two records agreeing on "01 Jan 00:00" is what a year-only placeholder would also produce. So the OSTI note's "January 1992" is confirmed as *what OSTI's export says*, not yet as *what the report's date is*; the year 1992 is unaffected. The cross-reference sidebar value "Tue Dec 31 19:00:00 EST 1991" equals 1992-01-01 00:00 UTC, consistent with the earlier note's UTC-rendering reading — one instant, rendered two ways. - [unverified-mechanism -- needs primary]: that OSTI fills unknown months/days with 01/01. Only one comparison record was inspected; a search-engine summary asserted this practice but is a summarising layer and is not relied on. I read the OSTI FAQ (sha 4ee05d82…6198a) through its opening sections (metadata fields are collected via the E-Link submission system, with a PDF metadata table); the passage on date semantics, if any, was not reached. The metadata-table PDF linked from that FAQ is the likely primary. - source_url: https://www.osti.gov/biblio/4024824 - source_sha: 45a819b6f23de86af39d827863246a25aaca054f17d08522c61afafc808a3ab8 - source_title: "ANNUAL PROGRESS REPORT." - source_author: Illinois Univ. (Materials Research Lab., Urbana), per OSTI catalog - source_date: 1971 (OSTI catalog); read 2026-09-26 - source_venue: OSTI.GOV, Technical Report COO-1198-808 - source_tier: 1 (for what OSTI's catalog records) - source_quote: "Technical Report·Fri Jan 01 00:00:00 EST 1971" - Second source_quote (SAND-91-7095 record, sha f96ee695…): "Technical Report·Wed Jan 01 00:00:00 EST 1992" > [!note] Seek's commentary: > The two notes look alike because both say "the database's structured field agrees with itself." But what a field can witness depends on its grain. Year-level data can carry a histogram; a month value that shows up as 01 with a midnight-January-1 timestamp may only be the database's way of writing "year known." That is a sharper caution for the OSTI note than for the Linnainmaa one. The resemblance is a shared *epistemic posture* (trust the backend), not a shared subject — so, false friend in content, real friend in method. ## Claim: The resemblance does not transfer evidence in either direction - claim_type: definitional/structural - The Linnainmaa histogram supports a claim about citation *volume by year* and says nothing about report dating; the OSTI export supports a claim about a catalog's recorded *year* and says nothing about citation behaviour. The vault's ML/AD-communities-mutual-unawareness note [[claim-ml-and-ad-communities-mutually-unaware]] and the Sandia cluster ([[entity-howard-schreyer]], [[entity-deborah-sulsky]], [[entity-zhen-chen]], [[claim-schreyer-bibliography-confirms-1992-asme-publication]]) sit in separate fields; [[claim-linnainmaa-field-numerical-analysis]] places Linnainmaa in numerical analysis, the neighbouring field to computational mechanics, but no citation or author link between them was found. [unverified — could not confirm or deny a hidden connection after search.] - Source: the two vault notes above (Tier 1 as house convention for vault-recorded facts; not independently re-fetched beyond the primary records cited in the previous claims). ## Safety flags None fired. All fetched pages (osti.gov record pages, OSTI FAQ, Semantic Scholar API JSON) were plain record/reference text with no addressed-to-AI, override, authority, credential or file-system content. The web search summary was treated as a lead only. ## Further leads - Re-run the Semantic Scholar query with `fields=title,authors,year` and script the per-decade tally, including a null-year count check (10 vs 9): https://api.semanticscholar.org/graph/v1/paper/DOI:10.1007/BF01931367/citations - OSTI E-Link "Metadata Table (PDF)" linked from https://www.osti.gov/faqs — likely primary for how partial dates are stored. - Sample more pre-1995 OSTI technical reports with known month-level dates (cover pages) to see whether month ever varies from 01 — would confirm or kill the placeholder hypothesis. Cover-page date of SAND-91-7095 itself remains unfound. - Search results surfaced other OSTI year-only-looking records (e.g. https://www.osti.gov/biblio/4626960) — not read. ## Entity candidates - Seppo Linnainmaa — person — the foundational figure whose 1976 paper the histogram measures; the priority claim rests on him. - Semantic Scholar Graph API — term — source of the citation counts; its year-only field defines the histogram's grain. - OSTI E-Link — term — DOE's submission system for citation metadata; determines what date fields OSTI records mean. - OSTI.GOV — term — U.S. DOE catalog; the venue of record for SAND-91-7095's metadata. - Deborah Sulsky — person — SAND-91-7095 author, the report's material-point-method lineage. - Howard Schreyer — person — SAND-91-7095 co-author. - Zhen Chen — person — SAND-91-7095 co-author. - Metadata granularity / placeholder dates — concept — year-only records rendered as Jan 1 00:00; affects any "month" claim from catalog exports. - Material point method — concept — subject of SAND-91-7095; separate field from automatic differentiation. - Citation histogram / bibliometric resurrection curve — concept — the Linnainmaa note's core artifact.