Bibliographies of the vanished

Claim (verbatim)

The Suda (c. 970s-980s, roughly 30,000 entries; ed. Ada Adler, 5 vols, 1928-1938; open and searchable as Suda On Line, whose translation was completed in 2014) closes its author entries with bibliographies: 'he wrote...' followed by titles, drawn largely from the lost epitome of Hesychius of Miletus' Onomatologon - a lost abridgement of a lost dictionary of authors, whose sixth-century lists were already partly obituary when the epitomator took them over. The entries credit the tragedians with plays by the dozen and enumerate the works of historians, sophists, grammarians and comic poets with the same notarial calm. That many of these titles are lost is notorious; what has never been cashed out as a resolvable number is the rate: what fraction of the works the Suda's bibliographies enumerate can still be read at all? The claim is that the Suda is, at title level, mostly a catalogue of nothing - a tenth-century index whose typical author entry points at zero surviving books - and that the encyclopedia which modern classics treats as a treasury is better described as the empire's largest register of graves, checkable entry by entry because Adler's numeration pins the entries and the TLG canon pins survival.

Prediction clause (verbatim)

Prediction: across all Suda entries for Greek poets and historians that enumerate three or more work-titles (entries pinned by Adler number via Suda On Line; homonyms follow SOL's headword division), fewer than one quarter of the enumerated titles will survive as independent, substantially complete texts in the TLG canon, and at least 40% of the sampled authors will have zero surviving enumerated titles (primary clause: the pooled title-survival share below 0.25; the verdict follows it); disambiguation: survival requires a continuous text outside quotation and epitome - fragments and testimonia count as lost; corrupt or duplicated titles are resolved per Adler's apparatus; coverage guard: void if the sample contains fewer than 50 qualifying author entries. Confidence: direction firm; exact shares MODERATE.

Kill-dataset (verbatim)

Kill: Suda On Line (Adler numeration) for the title-lists, joined to the TLG canon and Pinakes for per-title survival - a pooled survival-share census over enumerated titles.

Novelty / leakage triage

already answered in the literature

The census as pinned has not been run, but print fixes its answer far past the thresholds: the bibliographies' Hesychian source and obituary character are stated in Adler's prolegomena, and the survival arithmetic of the enumerated classes is canonical - seven plays apiece for Aeschylus and Sophocles against the dozens credited, Old Comedy reduced to one surviving author, and Strasburger's printed Truemmerfeld computation putting surviving Greek historiography at a few percent of the attested - so the pooled under-0.25 title-survival share and the zero-survivor author floor are guaranteed list-arithmetic on TrGF, FGrH and the TLG canon.

• A. Adler (ed.), Suidae Lexicon, 5 vols (Leipzig, 1928-1938), prolegomena on the Hesychian source of the author entries
• H. Strasburger, 'Umblick im Truemmerfeld der griechischen Geschichtsschreibung', in Historiographia antiqua: Commentationes Lovanienses in honorem W. Peremans (Leuven, 1977)
• B. Snell, R. Kannicht & S. Radt (eds.), Tragicorum Graecorum Fragmenta (Goettingen, 1971-2004)

Predictions

No prediction registered yet.