The handbook that everyone copied, and the tunes nobody did

Claim (verbatim)

The qin repertoire is one of the few premodern musics preserved as a large notated corpus rather than a bare name-list: roughly 150 handbooks (qinpu) survive, gathered in the Qinqu jicheng, and Zha Fuxi's concordance Cunjian guqin qupu jilan (存見古琴曲譜輯覽, 1958) indexes them, listing for each of some 650 distinct named pieces every handbook in which it appears. A corpus with an index is a corpus whose loss has a measurable shape. The claim is that qin-piece survival is radically long-tailed: the printed tradition has a small canonical spine - pieces first fixed in Zhu Quan's Shenqi mipu (神奇秘譜, 1425), the earliest surviving printed handbook - that were recopied into dozens of later handbooks, and a vast fringe of pieces attested in exactly one handbook and nowhere else, notated once and never taken up. The mechanism is the handbook economy: a compiler reprints the prestige core to legitimate his collection and adds a few local or personal pieces that die with him. So the attestation curve should look like a library's dark matter - a heavy head of a few dozen ubiquitous titles and a long tail of hapax pieces surviving on a single printing. Prediction restated: at least two in five distinct qin pieces are attested in exactly one handbook, while the most-copied pieces of the Shenqi mipu spine recur across dozens - a core-to-fringe spread of more than an order of magnitude.

Prediction clause (verbatim)

Prediction: tabulating, from Zha Fuxi's Cunjian guqin qupu jilan (存見古琴曲譜輯覽), the number of distinct handbooks in which each named qin piece appears, at least 40 percent of the distinct pieces will occur in exactly one handbook (hapax pieces), and the ten most-attested pieces will each appear in at least thirty handbooks - a core-to-fringe attestation ratio above 30 to 1. Primary clause: the hapax fraction is at least 40 percent. Disambiguation: pieces are counted as distinct by title after merging obvious variant titles (same piece, alternate name) per the concordance's own cross-references, and retunings or versions within a single handbook count once. Coverage guard: if the concordance indexes fewer than 400 distinct pieces, or fewer than 100 handbooks, the test voids.

Kill-dataset (verbatim)

Kill: Zha Fuxi, Cunjian guqin qupu jilan (存見古琴曲譜輯覽, 1958), which lists per-piece handbook attestations across the Qinqu jicheng corpus - build the attestation-frequency distribution over distinct pieces and read off the hapax fraction and the core-to-fringe ratio.

Novelty / leakage triage

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

Zha Fuxi's Jilan is itself the published instrument - a per-piece concordance of handbook attestations for the 3,360-odd surviving scores of some 650 pieces across the 150-odd handbooks - and the qin literature freely states the head of the distribution piecemeal (the Shenqi mipu spine endlessly recopied, famous pieces recurring in dozens of handbooks); but no located source states the hapax fraction or prints the attestation-frequency distribution, so the 40-percent primary clause is a genuine one-pass tabulation over Zha's tables that nobody appears to have run in print. Adjacent: concordance published, statistic un-run.

• Zha Fuxi, Cunjian guqin qupu jilan (Beijing: Renmin yinyue chubanshe, 1958)
• Qinqu jicheng (Zhonghua shuju), prefatory materials carrying Zha Fuxi's survey figures for handbooks, pieces and scores

Predictions

No prediction registered yet.