The bureau translated; the books did not keep

Claim (verbatim)

The Jin dynasty gave the Jurchen two scripts — a large script promulgated in 1119 under Wanyan Xiyin and a small script added in 1138 — and a state apparatus to feed them: the Jin shi records an academy translating the Chinese classics, the histories, and the canons into Jurchen, with a translation examination and named translators. Almost none of that output survives. What remains of written Jurchen is a dozen-odd stone inscriptions (the 1185 victory stele among them), a scatter of seals and paiza, and — decisively — the Ming dynasty's teaching aids: the Bureau of Translators' Sino-Jurchen vocabulary and model documents, a foreign-language textbook that outlived the literature it sampled. The mechanism is a double filter: a conquest script has no diaspora to keep copying it once the successor Yuan and Ming chanceries drop it, and a state-only literacy leaves nothing in private hands to survive by accident; the epitaph and the enemy's phrasebook are what pass through. So the surviving Jurchen word-stock should be a vocabulary of officialdom and elementary instruction, not of the translated library the Jin shi describes. Prediction: lemmatising the complete attested Jurchen lexicon (inscriptions plus the Ming glossaries), the number of distinct Jurchen words recoverable will fall below three thousand — an order of magnitude under the working vocabulary any of the translated classics required — and over four-fifths of that lexicon will come from the Ming pedagogical glossaries rather than from Jin-era Jurchen documents (primary clause: the four-fifths dependence on the Ming glossaries; the verdict follows it). Kill: Gisaburo N. Kiyose, A Study of the Jurchen Language and Script (1977) and Daniel Kane, The Sino-Jurchen Vocabulary of the Bureau of Interpreters (1989), against the bibliographic notices in the Jin shi (Zhonghua Shuju edition).

Prediction clause (verbatim)

Prediction: lemmatising the complete attested Jurchen lexicon (inscriptions plus the Ming glossaries), the number of distinct Jurchen words recoverable will fall below three thousand — an order of magnitude under the working vocabulary any of the translated classics required — and over four-fifths of that lexicon will come from the Ming pedagogical glossaries rather than from Jin-era Jurchen documents (primary clause: the four-fifths dependence on the Ming glossaries; the verdict follows it).

Kill-dataset (verbatim)

Kill: Gisaburo N. Kiyose, A Study of the Jurchen Language and Script (1977) and Daniel Kane, The Sino-Jurchen Vocabulary of the Bureau of Interpreters (1989), against the bibliographic notices in the Jin shi (Zhonghua Shuju edition).

Novelty / leakage triage

already answered in the literature

The attested Jurchen lexicon is a closed pair of published corpora - a dozen-odd inscriptions, seals, and paiza against the Ming Bureau glossaries - and the field is built on the fact that the glossaries (Kane's edition of the Bureau vocabulary) supply the overwhelming bulk of known Jurchen words while the inscriptions add only a thin remainder. The four-fifths glossary-dependence the primary clause predicts is therefore guaranteed list-arithmetic on two published closed corpora, not an open question. Kiyose's study and Kane's edition already treat the glossaries as effectively the whole lexicon; only the exact fraction is left to compute.

• Daniel Kane, The Sino-Jurchen Vocabulary of the Bureau of Interpreters (Uralic and Altaic Series 153, Indiana University, 1989)
• Gisaburo N. Kiyose, A Study of the Jurchen Language and Script (Kyoto: Horitsubunka-sha, 1977)

Predictions

No prediction registered yet.