📄 The Old English Poem "A Vampyre of the Fens": A Bibliographical Ghost

Summary

Olivares Merino investigates a phantom text: an Old English poem called "A Vampyre of the Fens", which several twentieth-century vampire books mention but which appears in no Old English corpus. He works through the bibliographies of these books to rebuild the chain of citations. Dudley Wright (1914), Montague Summers, Carol Senf (1988), Brian Frost (1989) and Matthew Bunson (1993) all lead back to an unsigned 1855 piece, "Vampyres", in Dickens's Household Words. That article is attributed to Edmund Ollier. Ollier only said there was an old poem about a fen-dwelling vampire. Later writers turned that description into a title and gave it a date. The author identifies the real source as Beowulf, with Grendel as the "vampire", because Grendel drinks a sleeping warrior's blood in lines 739–45. Grendel lacks the defining feature of a vampire, since he is not undead. The author uses reception theory (Jauss, Iser) to explain how nineteenth-century readers recast the monster through their own expectations. The study shows how secondary sources can manufacture a nonexistent work.

Key points

• Wright's Vampires and Vampirism (1914; reprinted 1991, p. 186) claims there is an Anglo-Saxon poem with this title. Old English scholars know no such poem exists.
• Ollier's 1855 Household Words article (vol. 11, pp. 39–43) described the poem only as being on that subject. Wright changed this into a title. Senf kept the original wording. Frost, who took the reference from Senf, still called it a poem by that name and was the first to date it to the early eleventh century. Bunson put the title in inverted commas and called it the first undead poem in the isles.
• The poem is Beowulf, and the "Vampyre of the Fens" is Grendel. He haunts fens and moors, attacks at night, drinks blood and eats flesh (ll. 739–45), and is beheaded. He is not a living-dead creature.
• According to the OED, the word "vampyre" first appears in English in 1734. Several websites merged that date with the phantom poem and produced the impossible claim of a 1734 Anglo-Saxon poem.
• Reception theory supplies the interpretive frame: meaning comes from the meeting of text and reader, so a reader can reshape a character so far that it detaches from its original work.

Interesting ideas

• I1. Tracing a claim back through each source's bibliography shows exactly where the wording changed, as when "on the subject of" became "with the title".
• I2. Describing what a text is about is a different relation from naming it. Confusing the two created a fake title.
• I3. Grendel matches many surface features of a vampire but fails its defining one (being undead). This shows the difference between resemblance and category membership.
• I4. An anachronism check exposes the error: a word first attested in 1734 cannot be the title of an eleventh-century poem.
• I5. Many websites repeat the same garbled sentence, so frequency gives no real support for the claim.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (learn): Keep a provenance chain for every silver label or teacher claim, so a wrong fact can be traced to the step where it was altered and repaired there.
• I2 (world): Store aboutness (topic) and naming (title/label) as separate relations in the world graph, so "a poem on X" never gets stored as a poem titled X.
• I3 (world): When placing entities in the ontology, mark some features as necessary and others as merely typical, so an entity with many typical traits is not accepted into a class it does not belong to.
• I4 (world): Check event order and date constraints when merging facts, and flag any combination in which something is named before its name existed.
• I5 (learn): Merge copies that come from one upstream source before counting agreement, so that only independent witnesses count toward consensus.