📄 Old English gārsecg — an Eke- Name?

Summary

Cassidy examines a puzzle in the Old English lexicon. Dictionaries give gārsecg two senses, "spear-man" and "ocean," but no surviving text uses it for a warrior. It appears only as a word for the sea. He reviews earlier explanations from Grein, Sweet, Walde-Pokorny/Holthausen, de Vries, and Bosworth-Toller. He rejects them as guesses or as links to Roman imagery such as Neptune's trident that tie the word to nothing Germanic. He then checks each part of the compound through its cognates, its Indo-European bases and its use in compounds. He proposes that "the Spearman" was an eke-name, an extra name, of Woden. On this view the ocean was personified as Woden, and the plain warrior sense passed to other words. Christian scribes removed the pagan reference, so only the sea meaning survived, and even that faded after the Old English period. A storm passage in Exodus serves as supporting evidence. The note shows how mythic context can explain a word that etymology alone cannot.

Key points

• King Alfred's Orosius equates gārsecg with Oceanus. The form occurs only in English, and the short form secg used for "ocean" points to gārsecg being the established term.
• A homonym secg "sword" (feminine, from IE *seq- "cut", also the source of seax and "sedge") is ruled out. A "spear-sword" makes no sense, and gārsecg is never feminine.
• Other words carried the plain "spear-man" sense: gārberend, gārwiga/-wigend, æscberend, æscwiga.
• The Woden link rests on several pieces of evidence:
  • his spear Gungnir;
  • the spear cast over the enemy as a sacrifice;
  • Coifi profaning an altar with a spear (Bede II.13);
  • the gallows allusion in Beowulf, lines 2444–9;
  • his role as husband of Nerthus, whose shrine Tacitus placed on an island in the Ocean;
  • his leading the Saxons across the sea.
• In Exodus, lines 487b–495, the raging sea is described with warrior vocabulary: mod, wedde, gyllende gryre, and an old sword.

Interesting ideas

• I1. A transparent compound can have a conventional meaning that its literal reading never actually shows in texts, so attested usage should outrank composition.
• I2. A candidate etymon is eliminated by two independent tests, grammatical gender and semantic plausibility, rather than by intuition.
• I3. When one sense is missing from the record, the author explains the gap by other words taking over that slot, which treats a missing sense as a fact to explain.
• I4. Personification is spotted when a non-human subject takes verbs and nouns normally reserved for human agents, such as anger, raging or yelling.
• I5. The author keeps the hypothesis openly unproven while showing how the evidence converges.

For Mova

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

• I1 (world): mark compounds and idioms whose meaning is not built from their parts, so the world builder uses the attested sense rather than the literal one.
• I2 (tokens): use agreement features such as gender and number as hard filters when choosing between homonymous lemma candidates, and record which test removed each candidate.
• I3 (learn): when silver data never shows a sense listed in the lexicon, check whether a competing word blocks it before trusting the dictionary sense.
• I4 (frames): flag animacy mismatches between a frame's agent role and its filler as possible metaphor or personification, and explain the flag by naming the clashing words.
• I5 (answer): attach a confidence level and the supporting chain to interpretive answers, keeping them separate from facts stated in the text.