A language that only ever glossed

Claim (verbatim)

Old Kanembu is a language known almost entirely as handwriting between other lines. It survives as the archaic Kanuri-Kanembu idiom used for interlinear and marginal glossing (tarjama) of the Qur'an and core Arabic texts in Borno, documented in Dmitry Bondarev's corpus of Borno Qur'anic manuscripts — a running vernacular exegetical register, centuries deep, that was almost never used to write free-standing books. The mechanism is functional confinement: the vernacular was admitted to the page only in service to the sacred Arabic text, so it accreted a huge glossing corpus while producing almost no autonomous literature; and because a catalogue describes a Qur'an as an Arabic Qur'an, the Old Kanembu layer inside it is bibliographically invisible — a whole language's written remains hide inside items shelved under another language. This is distinct from the conservatism of the barnāwī script: the claim here is about a spoken language surviving as an apparatus, not about letterforms. Prediction: across catalogued Borno Qur'anic manuscripts, the overwhelming majority of Old Kanembu written material will occur as interlinear or marginal glossing rather than as independent text — over ninety percent of Old Kanembu word-tokens will be glossial — and standard catalogue records will tag these manuscripts by their Arabic base text, not by the presence of Old Kanembu, so the language is undercounted to near-invisibility in collection metadata (primary clause: over ninety percent of the language's attested tokens are glosses on Arabic rather than autonomous text; the verdict follows it). Kill: the corpus of Borno Qur'anic manuscripts documented by Dmitry Bondarev and collaborators (SOAS and the Hamburg Centre for the Study of Manuscript Cultures; the Old Kanembu annotated corpus and its EAP-digitized holdings); a token-level gloss-versus-text ratio is not yet built as a public query but can be computed from that annotated corpus.

Prediction clause (verbatim)

Prediction: across catalogued Borno Qur'anic manuscripts, the overwhelming majority of Old Kanembu written material will occur as interlinear or marginal glossing rather than as independent text — over ninety percent of Old Kanembu word-tokens will be glossial — and standard catalogue records will tag these manuscripts by their Arabic base text, not by the presence of Old Kanembu, so the language is undercounted to near-invisibility in collection metadata (primary clause: over ninety percent of the language's attested tokens are glosses on Arabic rather than autonomous text; the verdict follows it).

Kill-dataset (verbatim)

Kill: the corpus of Borno Qur'anic manuscripts documented by Dmitry Bondarev and collaborators (SOAS and the Hamburg Centre for the Study of Manuscript Cultures; the Old Kanembu annotated corpus and its EAP-digitized holdings); a token-level gloss-versus-text ratio is not yet built as a public query but can be computed from that annotated corpus.

Novelty / leakage triage

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

Bondarev's corpus work is exactly the scholarship that DESCRIBES Old Kanembu as an interlinear and marginal exegetical register (tarjama) which almost never produced free-standing books, and that flags its bibliographic invisibility under Arabic base-text catalogue records. But the primary clause pins a token-level proportion — over ninety percent of Old Kanembu word-tokens occurring as glosses — and the conjecture itself concedes that gloss-versus-text ratio 'is not yet built.' Bondarev characterises the phenomenon qualitatively; the ninety-percent count over the annotated corpus is un-run. This is precisely the discipline's named describe-not-count case.

• Dmitry Bondarev, 'The Language of the Glosses in the Bornu Quranic Manuscripts', Bulletin of the School of Oriental and African Studies 69/1 (2006)
• Meikal Mumin & Kees Versteegh (eds.), The Arabic Script in Africa: Studies in the Use of a Writing System (Brill, 2014)
• SOAS 'Borno and Old Kanembu Islamic Manuscripts' (OKIM) digital corpus / Hamburg CSMC annotated Old Kanembu corpus

Predictions

No prediction registered yet.