Resolution: Supported
Caveats: SUPPORTED on the registered primary clause: political verse is enriched 5.8× among the book epigrams of Synopsis-Chronike manuscripts relative to the DBBE corpus (15.8% vs 2.7%), clearing the staked 3× threshold, and the dodecasyllable does lose its corpus majority within the subset (42% vs 68%, the secondary claim). BUT the result is honestly small and its causal story is weaker than the headline, and this must be read alongside the verdict: (1) n = 19 metre-bearing epigrams clears the registered ≥15 floor by only 4, and just THREE political-verse occurrences drive the numerator — a single misclassification would move it. (2) Critically, none of those three is one of the Chronicle's OWN dedicatory epigrams (to Eirene Sebastokratorissa); two of the three sit in a single manuscript (13784). So what is demonstrated is that political verse is over-represented among the epigrams that SHARE A CODEX with the Chronicle — not, more strongly, that the Chronicle's own paratext was composed in its host's metre, which is the mechanism the conjecture pictures. (3) DBBE tags 11 manuscripts; the Chronicle's real tradition runs to dozens of witnesses, so this is DBBE's catalogued slice, not a codicological census. The verdict stands on the criteria as registered; the enrichment is real but the sample is thin and the paratextual-'catching' reading is suggestive rather than established. Full compute + independent re-derivation in docs/generated/instrument_builds/dbbe-w02006/.
Registered against the in-house DBBE relational dump (dbbe.sql, on disk since 2026-06-17) BEFORE any figure was computed. Claim under test (the conjecture's own primary clause): among book-epigram occurrences carried in manuscripts that also contain Constantine Manasses' Synopsis Chronike, the share written in POLITICAL VERSE (the fifteen-syllable decapentasyllable, the chronicle's own metre) exceeds the corpus-wide political-verse share by AT LEAST A FACTOR OF THREE; secondarily, the dodecasyllable loses its corpus-wide majority within that subset.
Resolution criteria: DATA: restore/parse the DBBE dump dbbe.sql (tables data.occurrence, data.document, data.document_contains, data.poem/data.original_poem, data.meter/data.poem_meter, data.document_title) to a scratch schema; FREEZE a sha256 of dbbe.sql in a manifest before computing. || POPULATION: the SUBSET = book-epigram OCCURRENCES whose carrying manuscript (document) also contains the Synopsis Chronike of Constantine Manasses (identify the Synopsis-Chronike-bearing documents via document_title / work linkage matching 'Synopsis Chronike'/'Manasses'; the build reports how many documents and how it matched). COMPARISON = the full DBBE corpus meter distribution over all occurrences with a non-null meter. || METRE CLASSES from the meter field: POLITICAL VERSE = the fifteen-syllable (decapentasyllable / 'political verse') metre; DODECASYLLABLE = the twelve-syllable metre; the build records the exact DBBE meter labels it maps to each. || PRIMARY STATISTIC E = (political-verse share among SUBSET occurrences) / (political-verse share corpus-wide). || CLAUSE PRECEDENCE, in order: (1) INCONCLUSIVE_BY_DESIGN if the dump does not load, OR the Synopsis Chronike cannot be identified among DBBE documents, OR fewer than 15 metre-bearing epigram occurrences fall in the subset (too few to compare shares), OR the meter field is null across the subset. (2) SUPPORTED if E >= 3.0. (3) KILLED if E < 3.0 (the threefold enrichment fails). || REPORT: subset occurrence count, political-verse share in subset vs corpus-wide, E, the full metre distribution within the subset, the dodecasyllable share within the subset (secondary/narrative: whether it drops below 50%), and the document-identification method. computed_at must postdate registered_at (rule 6).
Known priors disclosure: Held at registration: the conjecture's own reasoning (a scribe copying a political-verse chronicle tunes its paratextual book epigrams to the host metre, so the work's metre is 'catching' and should show in its paratextual halo) and general knowledge that DBBE records meter/occurrence/manuscript-contents and that its full relational dump is on disk (richer than the ingested ContainedText slice, whose epigram_text/meter is empty). NOT computed before registration: the subset size, the political-verse shares, the enrichment E, or the metre distribution. The threefold threshold and the dodecasyllable-majority-loss are the conjecture's own.
Per the registered criteria (packet dd4ff6c0). SUBSET = original_poem occurrences one hop of document_contains from the 11 genre-175 manuscripts (all metre-bearing, none nested). Metre map: POLITICAL VERSE = Decapentasyllable (15-syllable); DODECASYLLABLE = the 12-syllable metre. PRIMARY E = subset political-verse share / corpus political-verse share. Compute: Sonnet (compute.sql). INDEPENDENTLY RE-DERIVED FROM SCRATCH by the shepherd with fresh SQL — re-deriving the 11 documents from genre 175, and the subset + corpus distributions separately — reproducing every load-bearing number exactly. computed_at postdates registered_at (rule 6).
Dataset: The full DBBE relational dump (dbbe.sql, sha256 b6382757…, on disk since 2026-06-17; restored to a scratch schema, data.* integrity intact). Population: the book-epigram occurrences carried in the 11 DBBE manuscripts that contain Constantine Manasses' Synopsis Chronike, identified via DBBE's own content-genre marker (genre 175 = Historiographica × person 794 Manasses — his SOLE historiographical work; his only other content-genre leaf is 245 Erotica, his romance). vs the full-corpus metre distribution over all 12,097 metre-bearing DBBE occurrences.
computed 2026-07-16