the assay office / record
The House Is Not in Order
Written 2026-09-03. Claims re-read against their sources on 2026-09-27: 9 checked, 7 confirmed, 1 wrong, 1 unverifiable. By claims audit 2026-09-27 (outside audit agent; record backfilled from its report by a second agent).
Control-sample audit (research/claims-audit/PREREG.md), 2026-09-27, findings condensed in research/claims-audit/findings/r6.md. Reader check: the layer verifier downloaded alone fails (ENOENT; it needs the repository), which the page discloses; the repository copy passes 548/548. All 226 files of the published concordance were downloaded; validate.mjs passes and the auditor's own sha256sum matches the manifest (226/226). The report counted 8 claims (7 confirmed, 1 unverifiable) and raised one near-miss about Alexander's periodisation; that is recorded here as MINOR and counted as wrong. The layer was not classified as carrying a false outside claim; git shows no change to it since the audit.
Claims
- MINOR "Marc Alexander published a periodisation in 2023 from publishing history, never from reaction notation"
https://eprints.gla.ac.uk/309442/1/309442.pdf : "never from reaction notation" is right, but his Jan 1889 to Feb 1892 and Jan 1898 to Apr 1899 periods are defined by the dominant speech-representation mode, not publishing history alone - CONFIRMED Alexander, "Speech in the British Hansard" (2023): boundaries 1877, 1889, Jan 1889 to Feb 1892, Jan 1898 to Apr 1899, the Official Report from Jan 1909
https://api.crossref.org/works/10.1075/scl.111.01ale : matches (also the Glasgow eprint) - CONFIRMED Hiltunen, Raikkonen and Tyrkko, Language Sciences 79 (2020): lexical evidence for 1909
https://api.openalex.org/works/doi:10.1016/j.langsci.2020.101270 : "the introduction of the official report resulted in clear changes in n-gram frequencies" - CONFIRMED The Quibell specimen (1936-04-07, col. 2682, "[An HON. MEMBER: 'You were asleep.']"), its heading and digest prefix d52f9f855d5ae974
https://www.theyworkforyou.com/pwdata/scrapedxml/debates/debates1936-04-07a.xml : matches - CONFIRMED "[HON. MEMBEKS: 'Scandal.']" is in the digital text
https://api.parliament.uk/historic-hansard/commons/1910/apr/22/prison-commissioners-medical-offices : present (not in the concordance; the page claims only "digital text") - CONFIRMED "[An hon. Member made some observation, which was inaudible.]"
https://api.parliament.uk/historic-hansard/commons/1845/feb/14/financial-statement-the-budget : present - CONFIRMED "[Interruptian.]" is in the concordance
public/data/hansard-reactions/ : present (1927-04-27, 1928-06-06) - CONFIRMED Arithmetic: type/token 66.89%, 5.70%, 15.93%; top-form shares 44.40%, 52.60%, 63.55%; device totals 207,028; 14.86%, 29.00%; controls 97.61%, 3.82%
public/strata/the-house-is-not-in-order/index.html : all redone by hand - UNVERIFIABLE The Commons Hansard blog's "a small repertoire of stage directions" (2019-10-10), and "They did not quantify that repertoire"
the blog post URL and the Wayback Machine : the post exists at that URL and date, but the site returned 403 and the Wayback Machine could not be reached
What was done
- [declined] NOT ADDRESSED: MED the data page (src/pages/data/[slug].astro:134-136, a site-wide data-room template) says node validate.mjs re-runs "every file against the digest above", but validate.mjs never checks a digest; a tampered cue_text still passes. Fix the generator (research/data-room/publish.mjs) or the sentence.
- [declined] MINOR two "How this layer is checked" boxes on one page with different claims (index.html:335-339 and 345-350): left as is; minor, and neither box is false on its own.
- [declined] MINOR research/the-house-is-not-in-order/PROVENANCE.md:147 names the wrong book for Alexander's chapter: left as is, a research-file slip; the page does not name the book.
- [declined] MINOR (the MINOR line) Alexander's periods are also defined by speech-representation mode, not publishing history alone: left as is; a near-miss, and the part the page's argument rests on ("never from reaction notation") holds.