Antiquity is the deepest silence

Claim (verbatim)

The zero-stanza ghosts of Skaldatal are not scattered evenly down its roster; they should thin as the catalogue approaches its compilers' own age. Skaldatal is ordered by patron, and its patrons run from the legendary kings of the Danish and Swedish prehistory - Ragnarr lodbrok, the half-mythical Bragi - through the historical Norwegian dynasty to thirteenth-century Icelandic and Norwegian lords contemporary with the very men who wrote the sagas and the Snorra Edda that preserve skaldic verse. Because that verse survived by prose citation, and because the prose was written closest, in time and in interest, to the recent kings whose reigns the kings' sagas actually narrate, the survival of a skald's work should be a function of how near his patron stood to the age of the saga-writers. The court poets of the deep legendary past are names with no reachable verse; the poets of Haraldr hardradi or Hakon Hakonarson stand a real chance of quotation. Restated: rank the catalogue's patrons early to late, split the roster into thirds, and the zero-stanza fraction among skalds of the earliest patron-third will exceed that of the latest third by at least a factor of two. The catalogue's silence deepens the further back it reaches - loss is a gradient in time, and the gradient is legible because Skaldatal keeps its poets in chronological patron order.

Prediction clause (verbatim)

Prediction: ordering Skaldatal's skalds by the era of the patron each served and dividing the roster into an earliest, a middle, and a latest third, the zero-surviving-stanza fraction in the earliest third will exceed that in the latest third by at least a factor of two (primary clause: the >=2x early-over-late ratio of zero-stanza skalds; the verdict follows it). Survival is scored identically to the census test (one edited whole stanza in skaldic.org), patron-era is read from Skaldatal's own sequence and the standard regnal chronology, and the test voids for coverage if the patron-era of fewer than 100 catalogued skalds can be assigned.

Kill-dataset (verbatim)

Kill: the Skaldic Project database (skaldic.org) per-skald corpora, with each Skaldatal poet bucketed by patron-era from the catalogue's own ordering (Uppsala Edda DG 11 / Kringla apographs) and standard regnal dates - compare zero-stanza fractions across the earliest and latest patron-thirds.

Novelty / leakage triage

anticipated in the literature — this exact test has never been run

The direction is textbook - the skalds of the legendary Danish and Swedish section are names without verse, while the eleventh- to thirteenth-century court poets are quoted throughout the kings' sagas - and Fidjestol's census of the encomiastic corpus, organised by reign, is the nearest published instrument. But nobody has bucketed Skaldatal's own roster into patron-era thirds and compared zero-stanza fractions. The late third is not trivially safe either: Skaldatal's final sections name many thirteenth-century poets of Swedish kings and Norwegian and Icelandic magnates whose verse is also wholly lost, so the >=2x ratio is a real computation with an uncertain margin. Adjacent on un-run arithmetic over published rosters.

• B. Fidjestol, Det norrone fyrstediktet (Ovre Ervik, 1982)
• Edda Snorra Sturlusonar, vol. III (Copenhagen, 1880-87), Skaldatal
• K.E. Gade (ed.), Poetry from the Kings' Sagas 2, Skaldic Poetry of the Scandinavian Middle Ages II (Turnhout, 2009), introduction

Predictions

No prediction registered yet.