The one recording and the one catalogue barely name the same tunes

Status: Anticipated · untested

Status is derived only from the shepherd-authored triage/prediction data above -- community submissions and claims are a separate overlay and can never change it (see the participation panel below).

This is a proposed connection between two domains, generated by a language model. It is not an article and not evidence: it sits below the evidence/publication boundary. A quantitative prediction and a named kill-dataset are attached (when registered) so the claim stays falsifiable rather than merely evocative.

Claim (verbatim)

One manuscript is the sole notated window onto secular music from the Tang cultural horizon: Pelliot chinois 3808, a scroll from the Dunhuang cave library whose verso carries twenty-five short pieces in pipa (lute) tablature - dance-tunes and banquet pieces, the only surviving notation of the cosmopolitan entertainment repertoire that the written record otherwise knows only as titles. The great title-register of that repertoire is Cui Lingqin's Jiaofang ji (教坊記, c. 749), which lists the roughly 324 tune-names of the Tang Music Bureau's entertainment troupes. Here loss can be measured by intersection: two independent samples of the same lost repertoire - one that preserves sound (P.3808's 25 notated pieces) and one that preserves names (the Jiaofang ji's 324 titles) - should barely overlap, because each is a thin, differently-biased draw from a repertoire of thousands. If our one surviving recording and our one big catalogue name mostly different tunes, then the surviving fragments do not even reinforce one another: the notated window and the named window look past each other into the same darkness. Prediction restated: fewer than a quarter of P.3808's 25 pieces bear a title matching any of the Jiaofang ji's 324, so that even our best-attested Tang secular tunes are, for the most part, either heard-but-unnamed or named-but-unheard, rarely both.

Prediction clause (verbatim)

Prediction: matching the twenty-five piece-titles of the Dunhuang pipa manuscript (Pelliot chinois 3808, verso) against the roughly 324 tune-titles of the Jiaofang ji (教坊記), at most six of the 25 will have a securely-matching title in the Jiaofang ji (overlap under 25 percent). Primary clause: the title overlap is six or fewer. Disambiguation: a match requires identity of the tune-name proper (曲名), not mere shared genre words, and partially-legible or editorially-reconstructed P.3808 titles are excluded from both the numerator and the count of 25. Coverage guard: if fewer than 20 of the 25 P.3808 titles are legible enough to test, the test voids.

Kill-dataset (verbatim)

Kill: the published transcription of Pelliot chinois 3808 verso (the Dunhuang pipa-pu readings in Rao Zongyi and in Ye Dong's studies) for the 25 titles, against a standard edition of Cui Lingqin's Jiaofang ji (教坊記) tune-list - count exact tune-name matches.

Nobody has run this test. The kill-data is named above. If you can run it — or you know the paper that already settles it — claim the kill or submit the prior scholarship. Kills and prior scholarship are credited here, by name, as they come in.

Provenance

Run: Fresh agent generation · model: claude-fable-5

Fresh blind generation by claude-fable-5, 2026-07-17, Asian musical-transmission wave (music_liturgy section) against real corpora of notated and orally-fixed Asian music: qin tablature (Qinqu jicheng, Zha Fuxi's Cunjian guqin qupu jilan, Shenqi mipu 1425, the Jieshi diao Youlan), the Dunhuang pipa manuscript (Pelliot chinois 3808) vs the Jiaofang ji, gagaku/togaku (Picken-Marett Music from the Tang Court; Meiji senteifu) and the medieval Japanese tablature encyclopedias (Jinchi yoroku, Sango yoroku), Korea's aak and dangak (Akhak gwebeom 1493, Goryeosa Akji, Munmyo jeryeak), Vietnamese nha nhac, the Sanskrit theory chain (Natyasastra/Dattila/Matanga's Brhaddeshi/Sarngadeva's Sangitaratnakara via GRETIL-SARIT), Samavedic gana fixity (Wayne Howard; Staal's Nambudiri documentation), Tibetan dbyangs-yig (BDRC; Ellingson), Javanese gamelan notation-absence (kepatihan; Serat Centhini; Wedhapradangga), Uyghur On ikki muqam (Turdi Akhun), and Tamil Tevaram pan. Every kill names a real open corpus/edition and a countable operation (format/attestation censuses, title-matching, citation-overshoot, survival inversions, name-survival and disagreement rates) with thresholds far from 1 and coverage guards; 'Kill (not yet built)' flags items whose decisive dataset must still be assembled. HARD EXCLUSION honored: zero Latin/Western-liturgy items (owned by the concurrent music-liturgy wave). Disjoint from the 2026-07-08 w09 music/liturgy wave (Latin/Islamicate/Armenian chant) and from the East Asia ctext/w04 text-culture waves. Drops recorded in the run report: an Indian theory-chain named-lost-predecessor item steered away from Kohala (owned by breadth_india_w2 ord 6) and from the rasa-trio doxography (sanskrit_gretil ord 7); the Samavedic item kept to gana/stobha fixity, disjoint from the sakha-geography item (breadth_india ord 10); the gamelan item kept to notation-absence, disjoint from the Yogyakarta-1812 library item (breadth_seasia_w2 ord 5); Tevaram used as pan modal-loss, disjoint from the Tamil witness-count and Tamil icai-treatise items (breadth_india_w2 ord 17; breadth_india ord 17).

Novelty / leakage triage

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

Every serious study of the pipa scroll has checked its titles against the Tang registers - Hayashi's pioneering decipherments, Ren Bantang's Jiaofang ji jianding glossing each of the 324 tune-titles with cross-attestations, Chen Yingshi's synthesis - and the individual identities (Qingbeile, Xijiangyue, Yizhou, Shuiguzi) are printed facts; but the systematic 25-against-324 overlap count under this conjecture's exclusion rules is not stated in anything located. Factual caution: roughly ten of the 25 entries carry proper tune-names, the rest being generic tempo/genre labels (manquzi, jiquzi) or 'another piece' entries, so the six-or-fewer threshold is close to structurally assured and the legibility-keyed coverage guard misses the real denominator problem.

Hayashi Kenzo (trans. Pan Huaisu), Dunhuang pipapu de jiedu yanjiu (Shanghai: Shanghai yinyue chubanshe, 1957)
Ren Bantang, Jiaofang ji jianding (Beijing: Zhonghua shuju, 1962)
Chen Yingshi, Dunhuang yuepu jieyi bianzheng (Shanghai: Shanghai yinyue xueyuan chubanshe, 2005)

Predictions

No prediction registered yet.

Weigh in

No community feedback yet.

New here? Create an account first

Create an account or sign in and your feedback is tied to you — you can track it, get replies, and claim this conjecture so others know you’re working on it. Prefer not to? Just leave your take below as a guest — only the name you type is shown.

Add your take

Posted immediately (spam is removed). Community feedback is never an adjudicated verdict and never changes this conjecture's triage label or status above.