The rewrite buries its source

Claim (verbatim)

Carolingian hagiographers rewrote the Merovingian classics on principle - the old lives were rustic, the new courts wanted grammar - and the rewriting killed what it improved. Alcuin reworked Jonas's seventh-century Vita Vedastis and the primitive life of Richarius; at St Gall, Wetti and then Walahfrid Strabo successively rebuilt the Vita Galli, and the oldest version now survives only in fragments while Walahfrid's became the house standard; at Saint-Amand in 907 Hucbald composed a fresh Vita Rictrudis avowedly after the older records had perished. The mechanism is liturgical real estate: a saint's office needed one authoritative legenda, and when the new version took the lectern slot it took the copying demand with it - the superseded exemplar became a book nobody needed to reproduce. The BHL's version chains record these substitutions number by number, and the Bollandists' manuscript catalogues count the witnesses on each side. If replacement, not accumulation, was the rule, the geometry should be stark: rewritings rich in copies, originals reduced to a residue - a fragment, an odd survivor in a distant library - or to nothing but the rewriter's own admission that a source once existed. Rewriting is the rare loss mechanism that leaves a signed confession.

Prediction clause (verbatim)

Prediction: for the pinned panel of documented Carolingian-era rewritings of Merovingian-or-older lives - Vedastus (Jonas of Bobbio's vita, rewritten by Alcuin), Richarius (the primitive vita, rewritten by Alcuin), Gallus (the vetustissima, rewritten by Wetti and again by Walahfrid Strabo), Rictrudis (a lost original attested by Hucbald's rewriting of 907) - the rewritings' aggregate BHLms witness count will exceed the originals' aggregate by at least 5:1 (primary clause; the verdict follows it), and at least one original in three will survive only in fragments, in at most three witnesses, or not at all; disambiguation: a lost original counts only on explicit prologue or medieval testimony of its prior existence, never by inference; a pair voids if BHLms records fewer than five witnesses for the pair combined.

Kill-dataset (verbatim)

Kill: BHLms witness counts per BHL number along the version chains of the pinned saints, with original-versus-rewriting assignment taken from the standard editions (Krusch's MGH SRM volumes for the Merovingian texts; the editions of Alcuin's and Walahfrid's lives).

Novelty / leakage triage

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

Every pinned substitution is printed philology - Krusch's MGH volumes carry the fragmentary Vita Galli vetustissima beside Wetti and Walahfrid and the Vedastus and Richarius chains, Hucbald's 907 prologue confession is the textbook case, and the reecriture volumes treat replacement-kills-source as the Carolingian norm - and the second conjunct (one original in three reduced to fragments or nothing) is already guaranteed in print by Gallus and Rictrudis alone. The primary clause, though, is the aggregate 5:1 BHLms witness ratio across the quartet, and that version-chain witness count has not been run on either side.

• B. Krusch's editions in MGH SRM 3 (1896), 4 (1902), and 7 (1920): the Vedastus, Gallus (with the fragmentary vita vetustissima), and Richarius chains
• M. Goullet & M. Heinzelmann (eds.), La reecriture hagiographique dans l'Occident medieval (Beihefte der Francia 58, Ostfildern, 2003)
• J.M.H. Smith, 'The hagiography of Hucbald of Saint-Amand', Studi medievali 3rd ser. 35 (1994)

Predictions

No prediction registered yet.