Seven-five marks the singing

Claim (verbatim)

The Kakuichi-bon of the Heike monogatari — fixed from the dictation of the blind master-singer Kakuichi in 1371 — is a performance score in prose clothing. The heikyoku tradition delivers its sections in named melodic categories (kudoki, sanju, hiroi, and the rest) against plain spoken kotoba, and melody is what selects for the alternating seven-five syllabic phrasing that Japanese song demands; speech does not. If so, the music never left the letters: the sung-spoken partition of the work should be recoverable from prosodic texture alone, with no recourse to the notation. Prediction: computing the density of seven-five and five-seven syllable-phrase alternation per section of the Kakuichi-bon, sections carried by melodic categories in the notated heikyoku tradition will show at least twice the alternation density of spoken kotoba sections, and a two-class classifier on prosody alone will recover the sung-versus-spoken partition at 75% accuracy or better (primary clause: the twofold density ratio; the verdict follows it). Kill (not yet built): a machine-readable Kakuichi-bon with one performance-category label per section, buildable now by aligning the digitized Ryukoku University Kakuichi manuscript with the fushi tables printed in the Iwanami Nihon koten bungaku taikei edition of Heike monogatari.

Prediction clause (verbatim)

Prediction: computing the density of seven-five and five-seven syllable-phrase alternation per section of the Kakuichi-bon, sections carried by melodic categories in the notated heikyoku tradition will show at least twice the alternation density of spoken kotoba sections, and a two-class classifier on prosody alone will recover the sung-versus-spoken partition at 75% accuracy or better (primary clause: the twofold density ratio; the verdict follows it).

Kill-dataset (verbatim)

Kill (not yet built): a machine-readable Kakuichi-bon with one performance-category label per section, buildable now by aligning the digitized Ryukoku University Kakuichi manuscript with the fushi tables printed in the Iwanami Nihon koten bungaku taikei edition of Heike monogatari.

Novelty / leakage triage

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

The Kakuichi-bon's heikyoku melodic categories vs plain kotoba are documented in performance scholarship, but the specific claim that 7-5 prosody alone recovers the sung/spoken partition (>=2x alternation density; >=75% classifier accuracy) was not located; the labelled machine-readable text it needs is, by the conjecture's admission, not yet built.

• Iwanami NKBT Heike monogatari edition (fushi tables); heikyoku performance scholarship

Predictions

No prediction registered yet.